CN118824043A - Intelligent parking guiding method, system and storage medium - Google Patents
Intelligent parking guiding method, system and storage medium Download PDFInfo
- Publication number
- CN118824043A CN118824043A CN202411122694.6A CN202411122694A CN118824043A CN 118824043 A CN118824043 A CN 118824043A CN 202411122694 A CN202411122694 A CN 202411122694A CN 118824043 A CN118824043 A CN 118824043A
- Authority
- CN
- China
- Prior art keywords
- data
- parking
- processing
- time
- carrying
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 89
- 238000003860 storage Methods 0.000 title claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 317
- 238000011156 evaluation Methods 0.000 claims abstract description 80
- 238000004458 analytical method Methods 0.000 claims abstract description 54
- 238000005457 optimization Methods 0.000 claims abstract description 41
- 238000013439 planning Methods 0.000 claims abstract description 23
- 238000006243 chemical reaction Methods 0.000 claims abstract description 17
- 238000012937 correction Methods 0.000 claims abstract description 17
- 238000003909 pattern recognition Methods 0.000 claims abstract description 16
- 238000012098 association analyses Methods 0.000 claims abstract description 14
- 238000012300 Sequence Analysis Methods 0.000 claims abstract description 13
- 238000009826 distribution Methods 0.000 claims description 67
- 238000004422 calculation algorithm Methods 0.000 claims description 56
- 238000013468 resource allocation Methods 0.000 claims description 39
- 230000006399 behavior Effects 0.000 claims description 38
- 238000004364 calculation method Methods 0.000 claims description 30
- 230000007613 environmental effect Effects 0.000 claims description 24
- 238000010219 correlation analysis Methods 0.000 claims description 21
- 238000000354 decomposition reaction Methods 0.000 claims description 20
- 238000007499 fusion processing Methods 0.000 claims description 19
- 230000004044 response Effects 0.000 claims description 19
- 230000011218 segmentation Effects 0.000 claims description 19
- 238000000605 extraction Methods 0.000 claims description 18
- 230000000737 periodic effect Effects 0.000 claims description 16
- 230000008451 emotion Effects 0.000 claims description 14
- 238000010276 construction Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 230000001932 seasonal effect Effects 0.000 claims description 9
- 239000012634 fragment Substances 0.000 claims description 8
- 230000004927 fusion Effects 0.000 claims description 7
- 239000002245 particle Substances 0.000 claims description 7
- 238000012731 temporal analysis Methods 0.000 claims description 7
- 238000000700 time series analysis Methods 0.000 claims description 7
- 230000007306 turnover Effects 0.000 claims description 7
- 238000007621 cluster analysis Methods 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 230000006978 adaptation Effects 0.000 claims description 4
- 238000013144 data compression Methods 0.000 claims description 4
- 238000004806 packaging method and process Methods 0.000 claims description 4
- 238000012913 prioritisation Methods 0.000 claims description 4
- 238000010845 search algorithm Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 4
- 238000005303 weighing Methods 0.000 claims description 4
- 238000009825 accumulation Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 35
- 230000006870 function Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 6
- 230000002093 peripheral effect Effects 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 206010063385 Intellectualisation Diseases 0.000 description 3
- 238000007906 compression Methods 0.000 description 3
- 230000006835 compression Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 238000007637 random forest analysis Methods 0.000 description 3
- 238000012163 sequencing technique Methods 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 230000000875 corresponding effect Effects 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000011157 data evaluation Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000012854 evaluation process Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Landscapes
- Traffic Control Systems (AREA)
Abstract
The application relates to the technical field of data processing, and discloses an intelligent parking guiding method, an intelligent parking guiding system and a storage medium. The method comprises the following steps: performing time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow prediction data and parking demand prediction data; carrying out multi-objective optimization processing on the parking demand prediction data to obtain objective decision data, and carrying out multi-channel information format conversion and priority ordering processing on the objective decision data to obtain real-time information data suitable for different release platforms; performing association analysis and path planning processing on the parking behavior data, the real-time traffic condition data and the parking lot occupancy data of the user to obtain candidate parking suggestion path data; and carrying out multidimensional cross analysis processing on the user feedback data and the occupancy rate data to obtain evaluation index data, and carrying out data correction on the candidate parking suggestion path data to obtain target parking suggestion path data. The intelligent parking guidance method and the intelligent parking guidance system improve timeliness and accuracy of intelligent parking guidance.
Description
Technical Field
The present application relates to the field of data processing technologies, and in particular, to an intelligent parking guidance method, system, and storage medium.
Background
With the acceleration of the urban process and the rapid increase of the keeping quantity of private cars, the problem of difficult parking is increasingly prominent, and the method becomes a great difficulty which puzzles the development of modern cities. To solve this problem, various types of parking guidance systems have been developed. Existing parking guidance systems typically employ fixed information dissemination devices, such as electronic display screens or mobile applications, to provide drivers with information about the location and free space in nearby parking lots. The system collects parking space occupation data of the parking lot in real time, transmits information to the central control center for processing and distributing, and helps drivers to quickly find available parking spaces.
However, the existing parking guidance system has a number of disadvantages. First, these systems often provide only static or semi-static parking information, and do not adequately account for real-time traffic conditions, user personal preferences, and dynamic changes in parking requirements. Second, existing systems lack deep analysis of user historical behavior and feedback, making it difficult to provide personalized parking advice. Moreover, these systems typically treat the parking lot as an isolated entity, ignoring interactions between parking behavior and factors such as ambient traffic flow, business activity, and the like. Finally, the existing system is low in efficiency in terms of data processing and information release, and is difficult to cope with the requirements of rapid processing and accurate pushing of large-scale real-time data.
Disclosure of Invention
The application provides an intelligent parking guiding method, an intelligent parking guiding system and a storage medium, which are used for improving timeliness and accuracy of intelligent parking guiding.
In a first aspect, the present application provides an intelligent parking guidance method, including: carrying out real-time acquisition and standardization processing on the multi-source traffic data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend; performing time sequence analysis and pattern recognition processing on the target traffic data set to obtain vehicle flow prediction data and parking demand prediction data; performing multi-objective optimization processing on the parking demand prediction data based on the traffic flow prediction data to obtain objective decision data, wherein the objective decision data comprises: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data; performing multi-channel information format conversion and priority ranking processing on the target decision data to obtain real-time information data applicable to different release platforms; collecting user parking behavior data, real-time traffic condition data and updated parking lot occupancy rate data in real time, and carrying out association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data and the parking lot occupancy rate data to obtain candidate parking suggestion path data comprising a target parking lot and a navigation route; performing multidimensional cross analysis processing on the user feedback data and the occupancy rate data acquired in real time to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency; and carrying out data correction on the candidate parking suggested path data through the evaluation index data to obtain target parking suggested path data, wherein the target parking suggested path data comprises the following components: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
With reference to the first aspect, in a first implementation manner of the first aspect of the present application, the multi-source traffic data includes road surface vehicle density data, vehicle passing data of an entrance and an exit of a parking lot, weather condition data, special event information, and historical parking data, and the real-time collecting and standardizing the multi-source traffic data to obtain a target traffic data set including traffic flow, availability of parking spaces, environmental factors, and historical trends includes: performing dynamic threshold segmentation processing on the road surface vehicle density data to obtain vehicle flow classification data reflecting traffic flow; performing difference value calculation processing on the vehicle passing data at the entrance and the exit of the parking lot to obtain real-time parking space occupancy rate data; carrying out semantic analysis processing on the weather condition data and the special event information to obtain environmental impact factor data; performing periodic decomposition processing on the historical parking data to obtain parking trend periodic data; performing space-time correlation analysis processing on the traffic flow classification data and the real-time parking space occupancy rate data to obtain parking demand hot spot distribution data; carrying out weighted fusion processing on the environmental impact factor data and the parking trend period data to obtain parking behavior prediction basic data; performing density clustering processing on the parking demand hot spot distribution data to obtain regional parking pressure evaluation data; performing time sequence interpolation processing on the parking behavior prediction basic data to obtain parking demand prediction data in a continuous time domain; performing self-adaptive weight distribution processing on the regional parking pressure evaluation data and the parking demand prediction data to obtain comprehensive parking index data; and carrying out multi-scale normalization processing on the comprehensive parking index data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend.
With reference to the first aspect, in a second implementation manner of the first aspect of the present application, the performing time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow prediction data and parking demand prediction data includes: performing time window segmentation processing on the target traffic data set to obtain a plurality of parking data fragments with time granularity, and performing Fourier transform processing on the parking data fragments with the time granularity to obtain parking periodic characteristic data; performing wavelet decomposition processing on the parking periodic characteristic data to obtain multi-scale parking trend data, and performing time sequence prediction processing on the multi-scale parking trend data through a long-short-term memory network algorithm to obtain initial parking demand prediction data; performing correlation analysis processing on the initial parking demand prediction data and real-time traffic flow data acquired in real time to obtain a flow-demand correlation factor, and performing dynamic weight distribution processing on the flow-demand correlation factor to obtain parking demand adjustment parameters; performing self-adaptive fusion processing on the initial parking demand prediction data and the parking demand adjustment parameters to obtain optimized parking demand prediction data, and performing trend extraction and seasonal decomposition processing on historical traffic flow data to obtain traffic flow basic characteristic data; and performing anomaly detection processing on the traffic flow basic characteristic data and the real-time traffic event data to obtain traffic flow fluctuation factors, and performing combined prediction processing on the traffic flow basic characteristic data and the traffic flow fluctuation factors to obtain traffic flow prediction data and parking demand prediction data.
With reference to the first aspect, in a third implementation manner of the first aspect of the present application, the performing, based on the traffic flow prediction data, multi-objective optimization processing on the parking demand prediction data to obtain objective decision data, where the objective decision data includes: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data, comprising: performing network flow model construction processing on the traffic flow prediction data to obtain a traffic flow network topology structure, and performing spatial clustering processing on the parking demand prediction data to obtain parking demand hot spot area data; performing matching degree calculation processing on the traffic flow network topological structure and the parking demand hot spot area data to obtain an initial traffic flow distribution scheme, and performing capacity constraint optimization processing on the initial traffic flow distribution scheme to obtain the traffic flow distribution scheme; performing differential analysis processing on the existing parking resource data and the parking demand prediction data to obtain parking resource gap data, and performing dynamic allocation processing on the parking resource gap data to obtain an initial parking resource allocation strategy; performing conflict detection processing on the initial parking resource allocation strategy to obtain the parking resource allocation strategy; carrying out fusion processing on the real-time road network state data and the traffic flow distribution scheme to obtain a multidimensional path evaluation index; carrying out path search processing on the multi-dimensional path evaluation indexes through a particle swarm optimization algorithm to obtain a candidate vehicle path set; and carrying out multi-objective weighing processing on the candidate vehicle path set to obtain the vehicle path recommended data, and integrating the traffic flow distribution scheme, the parking resource allocation strategy and the vehicle path recommended data into the objective decision data.
With reference to the first aspect, in a fourth implementation manner of the first aspect of the present application, the performing multi-channel information format conversion and prioritization on the target decision data to obtain real-time information data applicable to different publishing platforms includes: carrying out data compression processing on the traffic flow distribution scheme to obtain traffic flow summary data; carrying out space index construction processing on the parking resource allocation strategy to obtain parking space distribution data; performing path simplification processing on the vehicle path recommended data to obtain critical path node data; carrying out data fusion processing on the traffic flow summary data, the parking space distribution data and the critical path node data to obtain an information data packet; performing multi-level cache policy processing on the information data packet to obtain a layered storage information cache structure, and performing real-time update frequency analysis processing on the information cache structure to obtain information timeliness scoring data; the information timeliness scoring data is subjected to priority ranking processing to obtain an information release queue, and the data in the information release queue is subjected to multi-channel adaptation processing to obtain formatting information aiming at different release platforms; and carrying out secure encryption processing on the formatted information aiming at different release platforms to obtain the real-time information data applicable to the different release platforms.
With reference to the first aspect, in a fifth implementation manner of the first aspect of the present application, the collecting, in real time, user parking behavior data, real-time traffic condition data, and parking lot occupancy data, and performing association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data, and the parking lot occupancy data, to obtain candidate parking suggestion path data including a target parking lot and a navigation route, includes: performing cluster analysis processing on the user parking behavior data to obtain user parking preference information, and performing road congestion degree calculation processing on the traffic condition data to obtain a traffic fluency index; carrying out difference value statistics on the parking lot occupancy rate data to obtain a dynamic parking space occupancy rate; performing correlation analysis processing on the user parking preference signal and the dynamic parking space occupancy rate to obtain a personalized parking lot recommendation list; threshold segmentation processing is carried out on the traffic fluency index to obtain a feasible driving road section set; carrying out path planning processing on the feasible driving road section set through a heuristic search algorithm to obtain a plurality of candidate navigation routes; carrying out geographic coordinate extraction processing on the parking lots in the personalized parking lot recommendation list to obtain a target parking lot coordinate set; performing matching degree calculation processing on the plurality of candidate navigation routes and the target parking lot coordinate set to obtain a route-parking lot matching scheme; and carrying out comprehensive scoring processing on the route-parking lot matching scheme to obtain an optimal parking suggested path, and carrying out data packaging processing on the optimal parking suggested path to obtain candidate parking suggested path data containing a target parking lot and a navigation route.
With reference to the first aspect, in a sixth implementation manner of the first aspect of the present application, the multidimensional cross analysis processing is performed on the user feedback data and the occupancy data acquired in real time to obtain evaluation index data, where the evaluation index data includes: user satisfaction and resource utilization efficiency, including: carrying out emotion analysis processing on the user feedback data to obtain initial scores of the parking experience satisfaction, and carrying out time attenuation processing on the initial scores of the parking experience satisfaction to obtain timeliness weighted satisfaction data; performing time sequence decomposition processing on the parking lot occupancy rate data to obtain parking lot utilization efficiency trend data, and performing correlation analysis processing on the timeliness weighted satisfaction degree data and the parking lot utilization efficiency trend data to obtain satisfaction degree-efficiency correlation indexes; carrying out distribution feature extraction processing on the parking time length data of the user to obtain parking time length mode data, and carrying out matching degree calculation processing on the parking time length mode data and the turnover rate data of the parking lot to obtain space-time utilization efficiency indexes; carrying out fusion processing on the satisfaction-efficiency association index and the space-time utilization efficiency index to obtain a comprehensive evaluation initial value, and carrying out statistical analysis processing on parking guidance response time data to obtain a response efficiency score; performing weighted average processing on the comprehensive evaluation initial value and the response efficiency score to obtain target evaluation index original data; normalizing and quantifying the target evaluation index original data to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency.
With reference to the first aspect, in a seventh implementation manner of the first aspect of the present application, the performing data correction on the candidate parking advice path data through the evaluation index data to obtain target parking advice path data, where the target parking advice path data includes: the target parking lot geographic coordinates, the target travel route coordinate sequence, the predicted travel time, the predicted parking cost, and the real-time road condition data include: performing threshold analysis processing on the user satisfaction in the evaluation index data to obtain a parking lot satisfaction sorting list; carrying out geographic information extraction processing on the target parking lot in the candidate parking suggestion path data to obtain a candidate parking lot coordinate set, and carrying out matching screening processing on the parking lot satisfaction ordered list and the candidate parking lot coordinate set to obtain target parking lot geographic coordinates; carrying out path smoothing processing on the navigation route in the candidate parking suggestion path data to obtain an initial driving route coordinate sequence, and carrying out optimization processing on the initial driving route coordinate sequence to obtain a target driving route coordinate sequence; carrying out road section segmentation processing on the target driving route coordinate sequence to obtain a road section unit set, and carrying out fusion analysis processing on the road section unit set and real-time traffic flow data to obtain predicted passing time of each road section; performing accumulation calculation processing on the estimated travel time of each road section to obtain the estimated travel time; performing time sequence prediction processing on the historical price data of the target parking lot to obtain predicted parking cost; semantically processing the collected road condition data to obtain the real-time road condition data, and integrating the geographic coordinates of the target parking lot, the coordinate sequence of the target driving route, the estimated driving time, the estimated parking cost and the real-time road condition data into the target parking suggestion path data.
In a second aspect, the present application provides an intelligent parking guidance system, comprising:
The processing module is used for carrying out real-time acquisition and standardization processing on the multi-source traffic data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend;
the recognition module is used for carrying out time sequence analysis and pattern recognition processing on the target traffic data set to obtain vehicle flow data and parking demand prediction data;
The optimizing module is used for carrying out multi-objective optimizing processing on the parking demand prediction data based on the traffic flow data to obtain objective decision data, wherein the objective decision data comprises: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data;
the conversion module is used for carrying out multi-channel information format conversion and priority ordering processing on the target decision data to obtain real-time information data suitable for different release platforms;
The planning module is used for collecting the parking behavior data of the user, the real-time traffic condition data and the parking lot occupancy rate data in real time, and carrying out association analysis and path planning processing on the parking behavior data of the user, the real-time traffic condition data and the parking lot occupancy rate data to obtain candidate parking suggestion path data comprising a target parking lot and a navigation route;
the analysis module is used for carrying out multidimensional cross analysis processing on the user feedback data and the occupancy rate data acquired in real time to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency;
The correction module is used for carrying out data correction on the candidate parking suggested path data through the evaluation index data to obtain target parking suggested path data, wherein the target parking suggested path data comprises: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
A third aspect of the present application provides a computer readable storage medium having instructions stored therein which, when run on a computer, cause the computer to perform the above-described intelligent parking guidance method.
According to the technical scheme provided by the application, the target traffic data set comprising traffic flow, parking space availability, environmental factors and historical trend is obtained by carrying out real-time acquisition and standardization processing on the multi-source traffic data, so that the timeliness of the data is improved, and the reliability of decision making is enhanced. And carrying out time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow data and parking demand prediction data, and predicting traffic conditions and parking demands in a period of time in the future so as to make reasonable resource allocation in advance. The method comprises the steps of carrying out multi-target optimization processing on parking demand prediction data based on traffic flow data to obtain target decision data comprising traffic flow distribution schemes, parking resource allocation strategies and vehicle path recommendation data, carrying out multi-channel information format conversion and priority ordering processing on the target decision data to obtain real-time information data suitable for different release platforms, and greatly improving information propagation efficiency and coverage, so that a user can obtain parking guide information through various devices and platforms. User behavior data, real-time traffic condition data and parking lot occupancy rate data are collected in real time, association analysis and path planning processing are carried out, candidate parking suggestion path data containing target parking lots and navigation routes are obtained, accurate response to personalized demands of users is achieved, real-time road conditions are considered, and recommendation accuracy and practicality are improved. The method comprises the steps of carrying out multidimensional cross analysis processing on user feedback data and occupancy rate data acquired in real time to obtain evaluation index data comprising user satisfaction and resource utilization efficiency, carrying out data correction on candidate parking advice path data through the evaluation index data to obtain target parking advice path data comprising target parking lot geographic coordinates, target driving route coordinate sequences, predicted driving time, predicted parking fees and real-time road condition data, and realizing full-flow intellectualization from data acquisition to end user recommendation through multi-step data processing and decision optimization, thereby greatly improving parking efficiency, reducing time and cost for searching parking spaces by users, and improving timeliness and accuracy of intelligent parking guidance.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained based on these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of an embodiment of a smart parking guidance method according to an embodiment of the present application;
Fig. 2 is a schematic view of an embodiment of the intelligent parking guidance system in accordance with an embodiment of the present application.
Detailed Description
The embodiment of the application provides an intelligent parking guiding method, an intelligent parking guiding system and a storage medium. The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.
For ease of understanding, a specific flow of an embodiment of the present application is described below with reference to fig. 1, where an embodiment of an intelligent parking guidance method according to an embodiment of the present application includes:
Step S101, carrying out real-time acquisition and standardization processing on multi-source traffic data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend;
It will be appreciated that the execution subject of the present application may be an intelligent parking guidance system, and may also be a terminal or a server, which is not limited herein. The embodiment of the application is described by taking a server as an execution main body as an example.
In particular, real-time traffic data is collected from a number of sources, including road sensors, cameras, on-board GPS devices, parking lot management systems, weather stations, and historian databases, among others. The first step in the normalization process is data cleansing to remove outliers and noise. For example, for traffic data, abnormal readings due to equipment failure may occur, which data needs to be statistically identified and culled. Next, the data from different sources are time synchronized and spatially aligned, ensuring that all data is analyzed under the same space-time framework. In processing traffic flow data, vehicle counts and speed measurements are used to quantify traffic conditions on a road. These raw data are subjected to aggregation and averaging processing, and converted into a vehicle flow per lane per hour and an average vehicle speed. The parking space availability data is calculated by monitoring the number of vehicles at the entrance and the exit of the parking lot in real time and combining the total parking space number. Environmental factor data including weather conditions, air quality index, etc., which require numerical and classification processing to facilitate subsequent analysis. By time series analysis of data over a period of time, a periodic pattern and a long-term trend are identified. The method relates to technologies such as seasonal decomposition, trend extraction and the like, and finally generates a data set capable of reflecting a history rule.
All of this processed data is ultimately integrated into the target traffic data set. The data set adopts a standardized format and comprises information of multiple dimensions such as traffic flow, parking space availability, environmental factors, historical trend and the like. Each record of the dataset contains fields such as a time stamp, geographic location, numerical values of various indicators, etc., to facilitate subsequent analysis and decision making processes.
For example, at a particular point in time, a record in the target traffic dataset may contain the following information: the real-time traffic flow of a main road is 1200 vehicles/hour, and the average speed is 35 km/hour; the available parking space rate of a large parking lot nearby is 15%; the current weather is light rain, and the visibility is 3 km; from historical data analysis, this time point is typically the peak traffic period of the workday, with 30% higher parking requirements than usual.
Step S102, performing time sequence analysis and pattern recognition processing on a target traffic data set to obtain vehicle flow prediction data and parking demand prediction data;
specifically, performing time sequence analysis and pattern recognition processing on the target traffic data set to obtain the predicted traffic flow data and the predicted parking demand data is a key step in the intelligent parking guidance method. This process involves first data preprocessing, including data cleaning, denoising, and normalization. The data is decomposed using time series analysis techniques to identify trends, seasonal and periodic patterns. Methods employed herein include moving average, exponential smoothing, and ARIMA (autoregressive integrated moving average) models, and the like.
For traffic flow prediction, time series analysis can capture intra-day, weekly, and seasonal changes in traffic flow. For example, patterns such as the morning and evening peaks of weekdays, the shopping peaks of weekends, etc. can be identified by this analysis. Meanwhile, considering the influence of special events (such as holidays and large activities) on traffic flow, the factors can be combined and integrated as external variables. The parking demand prediction combines time series analysis and pattern recognition technology. In addition to considering the time trend of historical parking data, it is also desirable to analyze the relevance of parking behavior to other factors (e.g., weather, peripheral activity, etc.). Machine learning algorithms, such as random forests or support vector machines, may be used herein to identify key features that affect parking requirements. The pattern recognition process can recognize repetitive patterns and anomalies from complex traffic data. For example, similar traffic patterns may be grouped by a clustering algorithm to find typical traffic condition types. These identified patterns are not only used for prediction, but also to help understand the intrinsic laws of traffic behavior.
In practical application, the latest real-time data is continuously used for updating and calibrating so as to ensure the accuracy of prediction. At the same time, the prediction results are accompanied by confidence intervals to represent the degree of uncertainty of the prediction.
For example: it is assumed that a prediction is made of the parking requirements of a certain business district. The parking data per hour of the past three months is collected, and the parking data comprises information such as occupancy rate of a parking lot, traffic flow of surrounding roads, weather conditions and the like. By time series analysis, a law of working days (such as that the parking demand is sharply increased at about 9 early and 6 late) and a mode of weekends (such as that the parking demand is highest between 2 pm and 5 pm on Saturday) are identified. Then, using a pattern recognition algorithm, factors such as weather, peripheral activity, etc. are correlated with the parking demand, e.g., it is found that the parking demand increases by 15% on average in rainy days. Finally, based on these analysis results, a future parking demand of 24 hours is predicted. For example, it is predicted that the parking demand at 3 pm on tomorrow Saturday will reach 85% of the parking lot capacity with a confidence interval of + -5%. This prediction takes into account the combination of the time pattern of the historical data, the forecasted weather conditions (sunny days), and the known ambient activity information (mall promotions). Such predictions provide accurate decision basis for the parking guidance system, helping to optimize parking resource allocation and reduce traffic congestion.
Step S103, based on the traffic flow prediction data, performing multi-objective optimization processing on the parking demand prediction data to obtain objective decision data, wherein the objective decision data comprises: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data;
specifically, in the intelligent parking guidance method, performing multi-objective optimization processing on the parking demand prediction data based on the traffic flow prediction data is a complex and critical step. This process aims at balancing traffic flow, parking requirements and resource utilization, ultimately generating target decision data comprising traffic flow allocation schemes, parking resource allocation policies and vehicle path recommendation data.
First, the vehicle flow rate prediction data and the parking demand prediction data are integrated and analyzed. The traffic flow prediction data reflects the expected traffic conditions of different road segments in the road network, and the parking demand prediction data indicates the expected use of the respective parking lots. By performing a correlation analysis on these two sets of data, potential traffic bottlenecks and parking pressure points can be identified. The multi-objective optimization process utilizes a variety of algorithms including genetic algorithms, particle swarm optimization, and multi-criteria decision analysis. These algorithms consider multiple objectives simultaneously, such as minimizing overall traffic congestion, maximizing parking resource utilization, minimizing the time the vehicle is looking for a parking space, etc. In the optimization process, each target is given a different weight to reflect its relative importance. When the traffic flow distribution scheme is generated, a dynamic traffic distribution model is adopted. This model takes into account the predicted traffic flow and attempts to evenly distribute traffic flow into the network to avoid overcrowding certain road segments. At the same time it also considers the distribution of parking demands to direct the vehicle towards as far as possible in those areas with sufficient parking space.
The formulation of a parking resource allocation strategy involves the optimal allocation of existing parking resources. This includes dynamically adjusting parking rates, encouraging drivers to use less utilized parking lots, or opening temporary parking areas during peak hours. Strategies also take into account the special needs of different types of vehicles (e.g., electric vehicles, disabled vehicles).
The generation of the vehicle path recommendation data comprehensively considers real-time traffic conditions, predicted traffic flow and available parking space information. The optimal route is calculated using a shortest path algorithm (such as Dijkstra algorithm), but the "optimal" herein includes not only distance but also expected travel time, congestion probability, and parking possibility.
For example: suppose that a business area expects a significant amount of shopper flow in the Saturday afternoon. The traffic flow predictions show that the traffic flow of the main road will reach 150% of weekdays, while the parking demand predictions indicate that the parking space demand will rise by 80%. The multi-objective optimization process first generates a traffic flow distribution scheme that directs portions of traffic flow to secondary roads to relieve arterial road pressure. Then, a parking resource allocation policy such as temporarily opening a parking lot of a nearby office building is formulated, and a preferential price is implemented for the parking lot far from the business center. Finally, generating a path recommendation for the personalized vehicle, taking the predicted traffic condition and the occupancy rate of the parking space into account. For example, for a vehicle intended for a business, the recommended route may avoid the primary entrance of the anticipated congestion, guide through the secondary road to a parking lot that is a little farther but has sufficient parking space, and provide a walking route from the parking lot to the destination.
Step S104, carrying out multi-channel information format conversion and priority ranking processing on the target decision data to obtain real-time information data suitable for different release platforms;
Specifically, the contents of traffic flow distribution schemes, parking resource allocation strategies, vehicle path recommendation data and the like are classified and sorted. And converting the format of the data according to the characteristics of different release platforms and the requirements of users. The format conversion process takes into account a variety of distribution platforms such as mobile applications, car navigation systems, roadside electronic displays, and the like. For mobile applications, the data is converted into a lightweight JSON format, which facilitates fast transmission and parsing. Vehicle navigation systems require more compact data formats, typically employing binary coding to reduce the amount of data. The roadside electronic display screen is required to convert information into compact text and graphic formats. The prioritization process is an important step to ensure that the most critical information is communicated to the user in a timely manner. The ordering algorithm considers factors such as timeliness, importance, relevance and the like of the information. For example, information of a severe traffic jam warning or a sudden closure of a large parking lot may be given higher priority. Meanwhile, the personalized recommendation information is also subjected to priority adjustment according to the position and the destination of the user.
Information compression and optimization is also involved in the data processing process. For dynamic information which needs to be frequently updated, such as real parking space occupancy rate, an incremental update strategy is adopted, only a changed part is transmitted, and the therby reduces the data transmission quantity. For relatively stable information, such as parking locations, a caching mechanism is employed to reduce duplicate transmissions. In practical applications, this process is dynamic and continuous. Along with the generation of new decision data, the information conversion and sequencing processes are continuously carried out, so that the user can be ensured to obtain the latest and most relevant information all the time. Meanwhile, the system can monitor the response and the user feedback of the user and dynamically adjust the information release strategy to optimize the user experience.
For example: assuming a busy business area, the target decision data includes traffic flow predictions, expected occupancy of each parking lot, and recommended routes for the next 2 hours. The data processing process first classifies these information: the traffic flow prediction is converted into road congestion level (1-5 level), the occupancy rate of the parking lot is converted into percentage, and the recommended route is converted into a series of GPS coordinate points. For mobile application users, the system generates a data packet in JSON format, which contains road congestion level within 500 meters of the current position of the user, occupancy prediction of the last 5 parking lots and an optimized route. This packet is about 10KB in size and can be transmitted in 1 second under a 3G network. For roadside electronic displays, the system generates a piece of concise information: "road congestion 2 km ahead (level 4), recommended right turn to parking lot a (current void rate 30%)". In the prioritization, if the user is detected to enter a severely congested area, the system immediately pushes a warning message, and the priority is higher than that of the conventional parking space update message.
Step S105, collecting user parking behavior data, real-time traffic condition data and updated parking lot occupancy rate data in real time, and carrying out association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data and the parking lot occupancy rate data to obtain candidate parking suggestion path data comprising a target parking lot and a navigation route;
It should be noted that, the parking behavior data of the user is collected through the mobile application and the vehicle-mounted device, and includes information such as historical parking positions, parking time periods, preferences and the like of the user. The real-time traffic condition data are from road surface sensors, traffic cameras and floating car data, and reflect the congestion degree and the traffic flow speed of the current road network. The occupancy rate data of the parking lots are obtained through a real-time monitoring system of each parking lot, and the current available parking space number of each parking lot is provided. After data acquisition, data cleaning and preprocessing are carried out to remove abnormal values and noise. Then, correlation analysis is performed on the three types of data. The association analysis uses a data mining technology, such as an association rule mining algorithm, to find out potential relations between the parking behaviors of the user and traffic conditions and occupancy of the parking lot. For example, a parking lot selection tendency of a user under a specific traffic condition, or a user parking behavior pattern of a high occupancy period is analyzed.
The path planning process is based on the correlation analysis results and the current real-time data. And generating a plurality of candidate paths by adopting an improved A-algorithm or Dijkstra algorithm in consideration of real-time traffic conditions, predicted running time and parking lot availability. Each path contains a recommended target parking lot and navigation routes to the parking lot. In the path generation process, the algorithm trades off several factors: travel time, distance of parking lot from destination, current and predicted availability of parking lot, user history preference, etc. By the multi-objective optimization method, a comprehensive score is calculated for each candidate path, and a plurality of paths with the highest score are selected as candidate parking suggestion path data. As new data is continuously coming in, the system will update the path suggestions in real time. For example, if an abrupt congestion of the original recommended route is detected, or the target parking lot is suddenly full, the system may quickly recalculate and provide a new recommended route.
For example: assume that a user is driving to a shopping mall of a city center. The system first analyzes the user's historical parking data and finds that the user tends to select a parking lot that is no more than 500 meters away from the destination. Real-time traffic data shows that the current congestion index of the main road leading to the city center is 0.8 (0-1 scale,1 is a serious congestion). Meanwhile, the parking lot occupancy data shows that the current occupancy of three parking lots A, B, C near the shopping mall is 85%, 70% and 55%, respectively. Through association analysis, when the main road congestion index exceeds 0.7, the probability of users selecting a route which is farther but smoother is increased by 30%. Based on this information, the path planning algorithm generates three candidate paths:
1. By the secondary road reaching parking lot B, the travel time was expected to be 18 minutes, and walking to the destination was carried out for 5 minutes.
2. By-passing parking lot C, the travel time is expected to be 22 minutes, walking to the destination for 8 minutes.
3. Waiting for the main road congestion to be relieved and then arriving at parking lot A, the total time is estimated to be 25 minutes, and walking to the destination is carried out for 2 minutes.
And comprehensively considering the driving time, the walking distance, the parking lot availability and the user preference, and finally providing the three paths as candidate parking suggestion path data to the user.
Step S106, multidimensional cross analysis processing is carried out on the user feedback data and the occupancy rate data acquired in real time to obtain evaluation index data, wherein the evaluation index data comprise: user satisfaction and resource utilization efficiency;
It should be noted that, the user feedback data includes information such as a score of the parking experience, comment text, use frequency, and the like, which are collected by the mobile application or the vehicle-mounted device. The occupancy rate data come from a real-time monitoring system of each parking lot to reflect the use condition of the parking lot. The data preprocessing stage includes data cleansing, outlier detection and format normalization. For user feedback data, text analysis and emotion analysis are required to convert unstructured comments into quantifiable indicators. The occupancy data then requires time series analysis to identify usage patterns and trends.
The method adopts various data mining and machine learning technologies, such as cluster analysis, association rule mining, regression analysis and the like, and explores the relationship between user feedback and the occupancy rate of the parking lot. For example, through cluster analysis, different types of user groups and parking preferences thereof can be identified; through association rule mining, associations between high satisfaction and specific parking lot characteristics (e.g., location, price, quality of service) can be found.
The calculation of the user satisfaction index comprehensively considers a plurality of factors including user scoring, comment emotion tendencies, use frequency and the like. The resource utilization efficiency index is mainly based on occupancy rate data of the parking lot, and factors such as average occupancy rate, peak period utilization condition, turnover rate and the like are considered. These two indices are integrated by weighted averaging or a more complex multi-objective evaluation model.
The generation of the evaluation index data is dynamic and continuous. The system will update the index periodically to reflect the latest user feedback and resource utilization. Meanwhile, through time sequence analysis, the change trend of the index can be identified, and the prediction of future service quality and resource requirements is facilitated.
For example: a business area has three primary parking lots A, B, C. Over the past week, the system collected 500 pieces of user feedback and hourly occupancy data for these three parks. After preprocessing, the user feedback data is converted into a satisfaction score of 1-5 points and a emotion score in the range of-1 to 1. Occupancy data is normalized to a scale of 0-100%. The multidimensional cross analysis found that parking lot a had an average satisfaction score of 4.2 and an average emotion score of 0.6, but its average occupancy was only 60%. Parking lot B has a low score (3.5 points) with a emotion score of 0.1, but occupancy as high as 85%. The performance of the parking lot C is balanced, the score is 4.0, the emotion score is 0.4, and the occupancy rate is 75%.
Through correlation analysis, the system finds that high satisfaction is highly related to the convenience of the location and the quality of service of the parking lot, while high occupancy is closely related to price offers and peripheral business activities. Based on these findings, the system calculates an integrated user satisfaction index: a is 85 min, B is 70 min, and C is 80 min (full 100 min). The resource utilization efficiency index is: a is 65 minutes, B is 90 minutes, and C is 85 minutes. For example, for parking lot a, there is a need to enhance marketing strategies to increase occupancy; for B, improvement of service quality is required to improve user satisfaction; the behavior of C is relatively balanced but there is still room for optimization. Through the continuous data analysis and evaluation, the intelligent parking guidance method can continuously optimize the service, balance the user experience and the resource utilization efficiency, and further provide higher-quality parking guidance service.
Step S107, data correction is carried out on candidate parking suggested path data through evaluation index data to obtain target parking suggested path data, wherein the target parking suggested path data comprises: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
Specifically, the candidate parking advice paths are re-evaluated and ranked using the evaluation index data obtained previously, including the user satisfaction and the resource utilization efficiency. The data correction process adopts a multi-objective optimization algorithm, and comprehensively considers a plurality of factors such as user satisfaction, resource utilization efficiency, driving time, parking cost and the like. The algorithm calculates a composite score for each candidate path by assigning different weights to these factors. The allocation of weights is based on user preferences and system goals, such as perhaps more emphasis on user satisfaction or more emphasis on resource utilization efficiency during peak hours. In the correction process, candidate parking lots are screened first. Parking lots with lower assessment indicators may be rejected or prioritized. The remaining candidate paths are then optimized. This includes adjusting the travel route to avoid areas of congestion, or recommending a higher but slightly more satisfied parking lot. The optimization algorithm balances the increase in driving time and the improvement in parking experience, and finds the optimal balance point.
When generating the target parking advice path data, a plurality of data sources are integrated. The geographic coordinates of the target parking lot are obtained directly from the parking lot database. The target driving route coordinate sequence is generated through an improved path planning algorithm, and real-time road conditions and historical data are considered. The estimated travel time is calculated based on the current traffic conditions and the historical data and dynamically updated. The estimated parking cost is calculated from the parking lot price data and the estimated parking time period. The real-time road condition data come from the traffic monitoring system and the floating car data, and provide real-time traffic conditions of all road sections on the path. As new assessment index data is generated, the system will continually update and optimize the parking advice. This ensures timeliness and accuracy of the recommendation, being able to accommodate rapidly changing traffic and parking conditions.
For example: assuming a user is traveling to a commercial center business, the system originally generated three candidate routes, recommending parking lots A, B and C, respectively. The evaluation index data shows that the user satisfaction degree of A is 85 points, and the resource utilization efficiency is 70 points; b, the satisfaction degree is 75, and the resource utilization efficiency is 90; c satisfaction is 80 points, and resource utilization efficiency is 85 points (full points are 100 points). The system re-evaluates the three paths by means of a multi-objective optimization algorithm. Assuming that the algorithm gives weights of 0.6 and 0.4 to user satisfaction and resource utilization efficiency, respectively, the combined scores of the three parking lots are: 79 parts of A, 81 parts of B and 82 parts of C. Based on this result, C is selected as the target parking lot.
Next, detailed target parking advice path data is generated:
1. Geographic coordinates of the target parking lot: longitude 120.5 °, latitude 30.3 °;
2. Target travel route coordinate sequence: [ (120.4 °,30.2 °), (120.45 °,30.25 °), (120.5 °,30.3 °) ];
3. Estimated travel time: considering real-time road conditions, estimating to be 18 minutes;
4. The parking fee is estimated: based on the price of parking lot C (10 yuan per hour) and the expected residence time (2 hours), 20 yuan is estimated;
5. real-time road condition data: the main road on the path is clear, a section of the road near the destination is slightly congested, and the running time is expected to be increased by 2-3 minutes.
In the embodiment of the application, the target traffic data set comprising traffic flow, parking space availability, environmental factors and historical trend is obtained by carrying out real-time acquisition and standardization processing on the multi-source traffic data, so that the timeliness of the data is improved, and the reliability of decision making is enhanced. And carrying out time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow data and parking demand prediction data, and predicting traffic conditions and parking demands in a period of time in the future so as to make reasonable resource allocation in advance. The method comprises the steps of carrying out multi-target optimization processing on parking demand prediction data based on traffic flow data to obtain target decision data comprising traffic flow distribution schemes, parking resource allocation strategies and vehicle path recommendation data, carrying out multi-channel information format conversion and priority ordering processing on the target decision data to obtain real-time information data suitable for different release platforms, and greatly improving information propagation efficiency and coverage, so that a user can obtain parking guide information through various devices and platforms. User behavior data, real-time traffic condition data and parking lot occupancy rate data are collected in real time, association analysis and path planning processing are carried out, candidate parking suggestion path data containing target parking lots and navigation routes are obtained, accurate response to personalized demands of users is achieved, real-time road conditions are considered, and recommendation accuracy and practicality are improved. The method comprises the steps of carrying out multidimensional cross analysis processing on user feedback data and occupancy rate data acquired in real time to obtain evaluation index data comprising user satisfaction and resource utilization efficiency, carrying out data correction on candidate parking advice path data through the evaluation index data to obtain target parking advice path data comprising target parking lot geographic coordinates, target driving route coordinate sequences, predicted driving time, predicted parking fees and real-time road condition data, and realizing full-flow intellectualization from data acquisition to end user recommendation through multi-step data processing and decision optimization, thereby greatly improving parking efficiency, reducing time and cost for searching parking spaces by users, and improving timeliness and accuracy of intelligent parking guidance.
In a specific embodiment, the process of executing step S101 may specifically include the following steps:
(1) Carrying out dynamic threshold segmentation processing on the road surface vehicle density data to obtain vehicle flow classification data reflecting traffic flow;
(2) Performing difference value calculation processing on the vehicle passing data at the entrance and the exit of the parking lot to obtain real-time parking space occupancy rate data;
(3) Carrying out semantic analysis processing on the weather condition data and the special event information to obtain environmental impact factor data;
(4) Periodically decomposing the historical parking data to obtain parking trend periodic data;
(5) Carrying out space-time correlation analysis processing on the traffic flow classification data and the real-time parking space occupancy rate data to obtain parking demand hot spot distribution data;
(6) Carrying out weighted fusion processing on the environmental impact factor data and the parking trend period data to obtain parking behavior prediction basic data;
(7) Performing density clustering treatment on the parking demand hot spot distribution data to obtain regional parking pressure evaluation data;
(8) Performing time sequence interpolation processing on the parking behavior prediction basic data to obtain parking demand prediction data in a continuous time domain;
(9) Performing self-adaptive weight distribution processing on regional parking pressure evaluation data and parking demand prediction data to obtain comprehensive parking index data;
(10) And carrying out multi-scale normalization processing on the comprehensive parking index data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend.
Specifically, dynamic threshold segmentation processing is performed on the road surface vehicle density data. This step uses a variant of the Otsu method to dynamically adjust the segmentation threshold according to real-time traffic conditions. The calculation formula is as follows:
wherein T is the optimal threshold value, Is the inter-class variance. The obtained traffic flow classification data can more accurately reflect the current traffic condition, for example, the traffic flow is divided into low, medium and high levels.
For vehicle passing data at the entrance and exit of the parking lot, a simple difference calculation is used:
Wherein, O r is the occupancy, N in is the number of vehicles entering, N out is the number of vehicles leaving, and N total is the total number of vehicles. The semantic analysis of weather conditions and special event information adopts a TF-IDF algorithm:
TF-IDF(t，d，D)＝TF(t，d)×IDF(t，D)；
where t is the term, D is the document, and D is the document set. The environmental impact factor data thus processed can quantify the extent of impact of different events.
Periodic decomposition of historical parking data uses a Fast Fourier Transform (FFT):
X k is frequency domain data, X n is time domain data, and N is the number of data points. The thus obtained parking trend period data helps to predict future parking demand patterns.
The space-time correlation analysis of the traffic classification data and the real-time parking space occupancy rate data adopts a Kriging interpolation method:
Z *(x0) is the value of the predicted point, lambda i is the weight, and Z (x i) is the value of the known point. This step generates parking demand hotspot distribution data.
The weighted fusion of the environmental impact factor data and the parking trend period data uses a random forest algorithm, wherein the construction of the decision tree is based on a CART algorithm:
Gini (D) is the base index and C k is the sample subset of the kth class.
DBSCAN density clustering processing is carried out on the parking demand hot spot distribution data:
N∈(p)＝{q∈D|dist(p，q)≤∈}；
n ∈ (p) is the epsilon neighborhood of p points. This step yields regional parking pressure assessment data.
The parking behavior prediction basic data is processed through cubic spline interpolation:
Si(x)＝ai(x-xi)3+bi(x-xi)2+ci(x-xi)+di;
S i (x) is an interpolation function of the i-th section.
X is an argument representing time; is the x coordinate of the ith known data point; a i,bi,ci,di is the coefficient to be determined, which can be solved by known data points and continuity conditions.
The regional parking pressure evaluation data and the parking demand prediction data are subjected to self-adaptive weight distribution, and a gradient descent method is used:
w j is the weight, α is the learning rate, and J (w) is the cost function.
Finally, carrying out Min-Max normalization processing on the comprehensive parking index data:
x norm is the normalized value, x is the original value, and x min and x max are the minimum and maximum values, respectively.
For example: a commercial area has a road vehicle density of 80 vehicles per kilometer at 2 pm on Saturday. Using dynamic threshold segmentation, the threshold is calculated to be 60 vehicles/km, and is therefore determined to be a "high traffic" level. The entrance of the parking lot A passes through 200 vehicles, the exit passes through 150 vehicles, the total parking space is 300, and the occupancy rate is calculated to be (200-150)/300 approximately equal to 16.7%. The weather forecast 'rain' is processed by TF-IDF to obtain weight of 0.6. The FFT decomposition shows the peak of the parking demand period during this period. The cubic spline interpolation generates a heat point diagram, which shows that the center of the business area has the highest demand. Random forests predict a 20% increase in parking demand during this period. DBSCAN clusters show commercial center parking pressures up to 90% saturation. Cubic spline interpolation predicts a continuous rise in demand for the next 3 hours. The gradient descent method yields the integrated parking index 85 (full 100). Finally, min-Max is normalized to the range of 0-1 to form a final target traffic data set.
In a specific embodiment, the process of executing step S102 may specifically include the following steps:
(1) Performing time window segmentation processing on the target traffic data set to obtain a plurality of parking data fragments with time granularity, and performing Fourier transform processing on the parking data fragments with the time granularity to obtain parking periodic characteristic data;
(2) Performing wavelet decomposition processing on the parking periodic characteristic data to obtain multi-scale parking trend data, and performing time sequence prediction processing on the multi-scale parking trend data through a long-short-term memory network algorithm to obtain initial parking demand prediction data;
(3) Performing correlation analysis processing on the initial parking demand prediction data and real-time traffic flow data acquired in real time to obtain a flow-demand correlation factor, and performing dynamic weight distribution processing on the flow-demand correlation factor to obtain parking demand adjustment parameters;
(4) Carrying out self-adaptive fusion processing on the initial parking demand prediction data and the parking demand adjustment parameters to obtain optimized parking demand prediction data, and carrying out trend extraction and seasonal decomposition processing on historical traffic flow data to obtain traffic flow basic characteristic data;
(5) And performing anomaly detection processing on the traffic flow basic characteristic data and the real-time traffic event data to obtain traffic flow fluctuation factors, and performing combined prediction processing on the traffic flow basic characteristic data and the traffic flow fluctuation factors to obtain traffic flow prediction data and parking demand prediction data.
Specifically, the target traffic data set is subjected to time window segmentation processing to obtain a plurality of parking data fragments with time granularity. This step uses a sliding window technique to segment continuous time series data into data segments of different lengths, such as hours, days and weeks. And carrying out Fourier transform processing on the data fragments to obtain parking periodic characteristic data. The formula of the fourier transform is as follows:
where X (k) is frequency domain data, X (n) is time domain data, M is the number of data points, and z is the frequency index.
And carrying out wavelet decomposition processing on the parking periodic characteristic data to obtain multi-scale parking trend data. Wavelet decomposition enables capturing features of data on different scales. And then, performing time sequence prediction processing on the multi-scale parking trend data through a long-short term memory network (LSTM) algorithm to obtain initial parking demand prediction data. One of the core formulas of LSTM is:
ft＝σ(Wf·[ht-1,xt]+bf)；
Wherein f t is the output of the forgetting gate, σ is the sigmoid function, W f is the weight matrix, h t-1 is the hidden state at the last moment, x t is the current input, and b f is the bias term.
And carrying out correlation analysis processing on the initial parking demand prediction data and the real-time traffic flow data acquired in real time to obtain a flow-demand correlation factor. This step uses pearson correlation coefficients:
where r is the correlation coefficient, x i and y i are observations of two variables respectively, AndAre their average values. The following steps comprise dynamic weight distribution, self-adaptive fusion, trend extraction, seasonal decomposition, anomaly detection, combination prediction and the like, and the steps comprehensively utilize various statistical and machine learning technologies to finally obtain optimized traffic flow prediction data and parking demand prediction data.
For example: an intelligent parking guidance system in a business area collects hourly parking data over the past year. First, the data is partitioned into three time scale windows of 1 hour, 24 hours, and 7 days. Fourier transforming the data for the 24 hour window found a significant peak at frequency k=1 (corresponding to a 24 hour period) with an amplitude of 500, indicating a significant daily periodicity. After wavelet decomposition, trends in parking demand are captured on different scales, such as weekday and weekend differences. After the LSTM model is trained, the parking requirement of the next 24 hours is predicted, and the initial prediction result is 180 average vehicles per hour. And carrying out correlation analysis on the prediction result and real-time traffic flow data (such as 2000 vehicles per hour on a main road), and calculating a correlation coefficient r=0.85 to show strong correlation. Based on the correlation, the prediction result is adjusted, and finally the optimized parking demand prediction is obtained as an average 195 vehicles per hour.
In a specific embodiment, the process of executing step S103 may specifically include the following steps:
(1) Carrying out network flow model construction processing on the traffic flow prediction data to obtain a traffic flow network topological structure, and carrying out spatial clustering processing on the parking demand prediction data to obtain parking demand hot spot area data;
(2) Matching degree calculation processing is carried out on the traffic flow network topological structure and the parking demand hot spot area data to obtain an initial traffic flow distribution scheme, capacity constraint optimization processing is carried out on the initial traffic flow distribution scheme to obtain a traffic flow distribution scheme;
(3) Performing differential analysis processing on the existing parking resource data and the parking demand prediction data to obtain parking resource gap data, and performing dynamic allocation processing on the parking resource gap data to obtain an initial parking resource allocation strategy;
(4) Conflict detection processing is carried out on the initial parking resource allocation strategy, and the parking resource allocation strategy is obtained;
(5) Carrying out fusion processing on the real-time road network state data and the traffic flow distribution scheme to obtain a multidimensional path evaluation index;
(6) Carrying out path search processing on the multi-dimensional path evaluation indexes through a particle swarm optimization algorithm to obtain a candidate vehicle path set;
(7) And carrying out multi-objective weighing processing on the candidate vehicle path set to obtain vehicle path recommendation data, and integrating the traffic flow distribution scheme, the parking resource allocation strategy and the vehicle path recommendation data into objective decision data.
Specifically, the traffic flow prediction data is subjected to network flow model construction processing to obtain a traffic flow network topology structure. The road network is abstracted into a directed graph, wherein nodes represent intersections, edges represent road segments, and weights of the edges represent predicted traffic flow. And meanwhile, spatial clustering processing is carried out on the parking demand prediction data to obtain parking demand hot spot area data. Here DBSCAN (density-based spatial clustering algorithm) is employed to identify areas of high parking demand.
And carrying out matching degree calculation processing on the traffic flow network topological structure and the parking demand hot spot area data to obtain an initial traffic flow distribution scheme. The matching calculation takes into account road capacity, predicted traffic flow and peripheral parking requirements, and uses a weighted matching algorithm to distribute traffic flow. And then, carrying out capacity constraint optimization processing on the initial traffic flow distribution scheme to obtain a final traffic flow distribution scheme. This step uses the Frank-Wolfe algorithm to minimize the overall travel time under conditions that meet the road capacity constraints. For allocation of parking resources, firstly, difference analysis processing is carried out on existing parking resource data and parking demand prediction data to obtain parking resource gap data. The difference analysis adopts simple supply and demand comparison to identify the areas with insufficient supply of the parking spaces. And then, dynamically distributing the parking resource gap data to obtain an initial parking resource allocation strategy. This step uses a greedy algorithm to preferentially allocate resources to the area with the largest gap. And then, carrying out conflict detection processing on the initial parking resource allocation strategy to obtain a final parking resource allocation strategy. The conflict detection mainly considers the problems of repeated allocation and non-uniform allocation of resources, and solves the conflict through iterative adjustment.
In order to generate vehicle path recommended data, the real-time road network state data and the traffic flow distribution scheme are fused to obtain a multidimensional path evaluation index. These include expected travel time, path length, degree of congestion, parking convenience, and the like. And then, carrying out path search processing on the multi-dimensional path evaluation indexes through a particle swarm optimization algorithm to obtain a candidate vehicle path set. The particle swarm optimization algorithm simulates swarm intelligence, and can find an approximate optimal solution in a complex path space. And finally, performing multi-objective weighing processing on the candidate vehicle path set to obtain vehicle path recommendation data. The multi-objective balance uses the pareto optimal principle to balance factors such as driving time, distance, parking convenience and the like. And finally, integrating the traffic flow distribution scheme, the parking resource allocation strategy and the vehicle path recommendation data into target decision data, and providing comprehensive decision support for the intelligent parking guidance system.
For example: a commercial district in a city faces serious traffic congestion and difficult parking problems in the afternoon of Saturday. The area is abstracted into a directed graph containing 50 nodes (intersections) and 80 edges (road segments) through network flow model construction. The DBSCAN clustering algorithm identifies 3 major parking demand hotspots, located near shopping centers, dining streets, and entertainment venues, respectively. The initial traffic flow distribution scheme shows that the traffic flow of the main road exceeds 120% of the capacity, and after the optimization by the Frank-Wolfe algorithm, 20% of the traffic flow is distributed to the parallel secondary main road, so that the traffic flow of the main road is reduced to 105% of the capacity. The parking resource analysis finds that the gap of the shopping center area is maximum and reaches 200 parking spaces. The dynamic allocation strategy decides to open 150 temporary parking spaces in the nearby office building, and adds 50 roadside temporary parking spaces within the range of 500 meters. The particle swarm optimization algorithm generates 3-5 candidate paths for each vehicle, and the driving time, the distance and the parking convenience are considered. For example, for a vehicle destined for a shopping mall, the algorithm recommends a route around the secondary road, while increasing the distance by 10%, it is expected that 15 minutes of travel time will be saved and 80% of the probability of finding a parking space will be found.
In a specific embodiment, the process of executing step S104 may specifically include the following steps:
(1) Carrying out data compression processing on the traffic flow distribution scheme to obtain traffic flow summary data;
(2) Carrying out space index construction processing on a parking resource allocation strategy to obtain parking space distribution data;
(3) Performing path simplification processing on the vehicle path recommended data to obtain critical path node data;
(4) Carrying out data fusion processing on the traffic flow summary data, the parking space distribution data and the critical path node data to obtain an information data packet;
(5) Performing multi-level cache policy processing on the information data packet to obtain a layered storage information cache structure, and performing real-time update frequency analysis processing on the information cache structure to obtain information timeliness scoring data;
(6) The information timeliness grading data is subjected to priority ordering treatment to obtain an information release queue, and multi-channel adaptation treatment is carried out on the data in the information release queue to obtain formatting information aiming at different release platforms;
(7) And carrying out secure encryption processing on the formatted information aiming at different release platforms to obtain real-time information data applicable to the different release platforms.
Specifically, the traffic flow distribution scheme is subjected to data compression processing to obtain traffic flow summary data. This step employs a Discrete Cosine Transform (DCT) compression algorithm to convert the original traffic flow distribution data into a frequency domain representation and preserve the dominant frequency components, thereby greatly reducing the amount of data but preserving the critical information. And carrying out space index construction processing on the parking resource allocation strategy to obtain parking space distribution data. The R tree index structure is used for organizing the parking space position information in the two-dimensional space into a hierarchical tree structure, so that quick query and retrieval are facilitated. And carrying out path simplification processing on the vehicle path recommended data to obtain the critical path node data. This step applies the Douglas-Peucker algorithm to simplify the path by progressively deleting non-critical points while maintaining the overall shape and critical turning points of the path.
And carrying out data fusion processing on the traffic flow summary data, the parking space distribution data and the critical path node data to obtain an information data packet. The fusion process adopts a multi-level data structure, and organizes different types of data into a uniform format, so that the subsequent processing and transmission are facilitated. And carrying out multi-level cache policy processing on the information data packet to obtain a hierarchically stored information cache structure. This step uses an LRU (least recently used) caching algorithm to split the data into hot spot data that is stored in the fast-access cache and cold data that is stored in the slower storage tier. And carrying out real-time updating frequency analysis processing on the information cache structure to obtain information timeliness scoring data. The sliding window technology is adopted, the access frequency of each piece of information in the latest time window is calculated, and the timeliness score is given according to the frequency. And then, carrying out priority ranking treatment on the information timeliness scoring data to obtain an information release queue. The sequencing adopts a heap sequencing algorithm to ensure the preferential release of the information with high timeliness.
And carrying out multi-channel adaptation processing on the data in the information release queue to obtain formatting information aiming at different release platforms. This step uses the adapter schema to customize the data format and presentation for each distribution platform (e.g., mobile APP, vehicle device, roadside display, etc.). And finally, carrying out secure encryption processing on the formatted information aiming at different release platforms to obtain real-time information data applicable to the different release platforms. The encryption process adopts an AES (advanced encryption Standard) algorithm, so that the security of the data in the transmission process is ensured.
For example: intelligent parking guidance systems in urban commercial areas process large amounts of real-time data during the Saturday afternoon. The original traffic flow distribution scheme comprises 5000 data points, 500 main frequency components are reserved after DCT compression, and the data volume is reduced by 90% but 95% of information is reserved. The area has 200 parking lots, and the query time is reduced from linear O (n) to logarithmic O (log n) by R-tree index construction. One recommended path originally comprises 100 GPS coordinate points, 20 key nodes are reserved after the traffic track thinning (Douglas-Peucker) algorithm is simplified, and the path length error is less than 2%.
The size of the information packet after data fusion is 2MB, and 500KB of hot spot data is stored in a memory and the rest is stored in a solid state disk through multi-level cache policy processing. The real-time updating frequency analysis shows that the access frequency of the parking space information is 3 times that of the traffic flow information, so that higher priority is obtained in the information release queue. For a mobile APP user, a data packet in a JSON format is generated, and the size of the data packet is 50KB; whereas for roadside displays, simplified text format data is generated, only 10KB. All data are encrypted by AES-256 bits before transmission, so that information security is ensured.
In a specific embodiment, the process of executing step S105 may specifically include the following steps:
(1) Performing cluster analysis processing on the user parking behavior data to obtain user parking preference information, and performing road congestion degree calculation processing on traffic condition data to obtain a traffic fluency index;
(2) Carrying out difference value statistics on the occupancy rate data of the parking lot to obtain the occupancy rate of the dynamic parking space;
(3) Carrying out correlation analysis processing on the user parking preference signals and the occupancy rate of the dynamic parking spaces to obtain a personalized parking lot recommendation list;
(4) Threshold segmentation processing is carried out on the traffic fluency index to obtain a feasible driving road section set;
(5) Carrying out path planning processing on the feasible driving road section set through a heuristic search algorithm to obtain a plurality of candidate navigation routes;
(6) Carrying out geographic coordinate extraction processing on the parking lots in the personalized parking lot recommendation list to obtain a target parking lot coordinate set;
(7) Performing matching degree calculation processing on a plurality of candidate navigation routes and a target parking lot coordinate set to obtain a route-parking lot matching scheme;
(8) And carrying out comprehensive scoring processing on the route-parking lot matching scheme to obtain an optimal parking suggested path, and carrying out data packaging processing on the optimal parking suggested path to obtain candidate parking suggested path data containing the target parking lot and the navigation route.
Specifically, cluster analysis processing is carried out on the user parking behavior data to obtain user parking preference information. This step employs a K-means clustering algorithm to divide users into different groups based on their historical parking selections (e.g., distance destination, price, type of parking lot, etc.). The objective function of the K-means algorithm can be expressed as:
Wherein J is an objective function, h is a cluster number, n is a data point number, For the ith data point in the jth class, c j is the cluster center of the jth class.
And (5) calculating and processing the road congestion degree of the traffic condition data to obtain a traffic fluency index. The ratio of the vehicle average speed to the road design speed is used here to represent:
F i is the fluency index, V a is the actual average speed, and V d is the road design speed.
And carrying out difference value statistics processing on the occupancy rate data of the parking lot to obtain the occupancy rate of the dynamic parking space. And then, carrying out correlation analysis processing on the parking preference information of the user and the occupancy rate of the dynamic parking space to obtain a personalized parking lot recommendation list. This step uses a weighted scoring system that considers both user preferences and current parking lot conditions. And (3) carrying out threshold segmentation processing on the traffic fluency index to obtain a feasible driving road section set, and generally considering road sections with fluency indexes higher than 70% as feasible driving road sections.
And (3) carrying out path planning processing on the feasible driving road section set through a heuristic search algorithm (such as an A-algorithm) to obtain a plurality of candidate navigation routes. The evaluation function of the a-algorithm is:
f(n)＝g(n)+h(n)；
f (n) is the estimated total cost of node n, g (n) is the actual cost from the start point to n, and h (n) is the estimated cost from n to the end point.
And carrying out geographic coordinate extraction processing on the parking lots in the personalized parking lot recommendation list to obtain a target parking lot coordinate set. And then, carrying out matching degree calculation processing on the plurality of candidate navigation routes and the target parking lot coordinate set to obtain a route-parking lot matching scheme. The matching degree calculation considers factors such as the distance between the route end point and the parking lot, the expected driving time and the like.
And finally, comprehensively scoring the route-parking lot matching scheme to obtain an optimal parking suggested route, and carrying out data packaging processing on the optimal parking suggested route to obtain candidate parking suggested route data containing the target parking lot and the navigation route.
For example: a user prepares to go to a business for shopping during the afternoon of Saturday. The system analyzes the historical parking data of the user and finds out a parking lot with moderate price within 500 meters of the destination. The K-means algorithm (k=3) is used to classify users belonging to a "convenience priority" group. The real-time traffic data show that the average speed of the main road A is 40km/h, the design speed is 60km/h, the calculated fluency index FA=66.7% is lower than the 70% threshold value, and the fluency index FA=66.7% is excluded from the feasible driving road sections. The smoothness index fb=85% of the secondary trunk B is incorporated into the feasible section.
3 Parking lots P1, P2 and P3 near the commercial area are detected, and the current occupancy rates are 85%, 60% and 40% respectively. Taking the user preference and the current occupation condition into consideration, the system generates a personalized parking lot recommendation list for the user: [ P2, P3, P1]. 3 candidate routes are planned on the feasible road segments using an a-algorithm, wherein the f (n) value of the route R1 reaching P2 is the lowest, which is the optimal choice. After the comprehensive scoring, the final recommended user selects parking lot P2, arriving via route R1. The recommended route is expected to travel for 15 minutes, the parking lot is 450 meters from the destination, and the parking cost is 10 yuan/hour. The system encapsulates this information into a data packet containing the coordinates (longitude: 120.5 deg., latitude: 30.3 deg.) of the target parking lot P2 and the key node coordinate sequence of the navigation route R1, as final candidate parking advice path data, to be provided to the user.
In a specific embodiment, the process of executing step S106 may specifically include the following steps:
(1) Carrying out emotion analysis processing on the user feedback data to obtain initial scores of the parking experience satisfaction, and carrying out time attenuation processing on the initial scores of the parking experience satisfaction to obtain time-lapse weighted satisfaction data;
(2) Performing time sequence decomposition processing on the occupancy rate data of the parking lot to obtain utilization efficiency trend data of the parking lot, and performing correlation analysis processing on the weighted satisfaction data of time efficiency and the utilization efficiency trend data of the parking lot to obtain satisfaction-efficiency correlation indexes;
(3) Carrying out distribution feature extraction processing on the parking time length data of the user to obtain parking time length mode data, and carrying out matching degree calculation processing on the parking time length mode data and the turnover rate data of the parking lot to obtain space-time utilization efficiency indexes;
(4) Carrying out fusion processing on satisfaction-efficiency association indexes and space-time utilization efficiency indexes to obtain comprehensive evaluation initial values, and carrying out statistical analysis processing on parking guidance response time data to obtain response efficiency scores;
(5) Performing weighted average processing on the comprehensive evaluation initial value and the response efficiency score to obtain target evaluation index original data;
(6) Normalizing and quantifying the original data of the target evaluation index to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency.
In particular, the evaluation process of the intelligent parking guidance method involves a number of complex data processing steps aimed at comprehensively evaluating system performance and user experience. Firstly, emotion analysis processing is carried out on the user feedback data, and initial scores of parking experience satisfaction are obtained. This step converts the text comments into numerical scores using natural language processing techniques. The core formula of emotion analysis is:
wherein S is emotion score, w i is weight of the ith word, p i is emotion polarity value of the ith word, and n is word number in the comment. And carrying out time attenuation treatment on the initial score of the satisfaction degree of the parking experience to obtain time-efficiency weighted satisfaction degree data. The time decay function adopts an exponential decay model:
W(t)＝S·e-λt
w (t) is the weighted satisfaction after time t, λ is the decay coefficient, and t is the time interval.
And carrying out time sequence decomposition processing on the occupancy rate data of the parking lot to obtain utilization efficiency trend data of the parking lot. Using the STL (Seasonal and Trend decomposition using Loess) method, the time series was decomposed into trend, seasonal and residual components. And then, carrying out correlation analysis processing on the timeliness weighted satisfaction data and the parking lot utilization efficiency trend data to obtain a satisfaction-efficiency correlation index. Here pearson correlation coefficient calculation is used.
And carrying out distributed feature extraction processing on the parking time length data of the user to obtain parking time length mode data. And identifying the distribution characteristics of the parking time length by using a nuclear density estimation method. And then, carrying out matching degree calculation processing on the parking time length mode data and the parking lot turnover rate data to obtain space-time utilization efficiency indexes. The matching degree calculation considers the matching degree of the parking time length distribution and the design turnover rate of the parking lot. And carrying out fusion processing on the satisfaction-efficiency association index and the space-time utilization efficiency index to obtain a comprehensive evaluation initial value. The fusion process employs a weighted summation approach. And meanwhile, carrying out statistical analysis processing on the parking guidance response time data to obtain a response efficiency score. The Cumulative Distribution Function (CDF) is used here to evaluate the performance of the response time. And carrying out weighted average processing on the comprehensive evaluation initial value and the response efficiency score to obtain evaluation index original data, and then carrying out normalization and quantization processing on the evaluation index original data to obtain final evaluation index data, wherein the final evaluation index data comprises user satisfaction and resource utilization efficiency. Normalization uses a Min-Max method to map data to the [0,1] interval.
For example: a commercial parking lot has collected 1000 user feedback over the past month. By emotion analysis, the system calculated an average satisfaction initial score of 4.2 (5 points full). The scoring weight was reduced to about 70% a week ago using a time decay function (λ=0.05). After the occupancy rate data of the parking lot is decomposed by STL, the utilization rate of the weekend is shown to be 20% higher than that of the weekday. The satisfaction-efficiency correlation index calculation was 0.75, indicating a strong positive correlation.
The analysis of the parking time data shows that the average parking time of the parking lot is 2.5 hours, and the average parking time is matched with the designed turnover rate (6 times a day), and the space-time utilization efficiency index is 0.85. The integrated evaluation initial value was calculated to be 0.80. The average response time of the parking guidance system was 3 seconds, 90% of the requests were responded to within 5 seconds, and the response efficiency score was 0.88. Finally, the original data of the evaluation index is normalized to obtain a user satisfaction index of 0.85 and a resource utilization efficiency index of 0.82.
In a specific embodiment, the process of executing step S107 may specifically include the following steps:
(1) Threshold analysis processing is carried out on the user satisfaction in the evaluation index data, and a parking lot satisfaction sorting list is obtained;
(2) Geographic information extraction processing is carried out on the target parking lot in the candidate parking suggestion path data to obtain a candidate parking lot coordinate set, and matching screening processing is carried out on the parking lot satisfaction ordered list and the candidate parking lot coordinate set to obtain geographic coordinates of the target parking lot;
(3) Carrying out path smoothing processing on the navigation route in the candidate parking suggestion path data to obtain an initial driving route coordinate sequence, and carrying out optimization processing on the initial driving route coordinate sequence to obtain a target driving route coordinate sequence;
(4) Carrying out road section segmentation processing on the target driving route coordinate sequence to obtain a road section unit set, and carrying out fusion analysis processing on the road section unit set and real-time traffic flow data to obtain predicted passing time of each road section;
(5) The estimated travel time of each road section is accumulated and calculated to obtain estimated travel time;
(6) Performing time sequence prediction processing on historical price data of a target parking lot to obtain predicted parking cost;
(7) Semantically processing the collected road condition data to obtain real-time road condition data, and integrating the geographic coordinates of the target parking lot, the target driving route coordinate sequence, the predicted driving time, the predicted parking cost and the real-time road condition data into target parking suggestion path data.
Specifically, threshold analysis processing is performed on the user satisfaction in the evaluation index data, and a parking lot satisfaction sorting list is obtained. Parking lots with satisfaction scores higher than 75% quantiles are screened out by using a quantile method, and are ranked from high to low according to scores. And then, carrying out geographic information extraction processing on the target parking lot in the candidate parking suggestion path data to obtain a candidate parking lot coordinate set. The parking lot address is converted to latitude and longitude coordinates using a geocoding technique. And then, carrying out matching screening processing on the parking lot satisfaction degree ordered list and the candidate parking lot coordinate set to obtain the geographic coordinates of the target parking lot. The matching process considers the dual factors of satisfaction rank and geographic location, selecting the optimal parking lot.
And carrying out path smoothing processing on the navigation route in the candidate parking suggested path data to obtain an initial driving route coordinate sequence. This step uses the Douglas-Peucker algorithm to simplify the redundant points in the path while preserving the critical turning points. And then, carrying out optimization processing on the initial travel route coordinate sequence to obtain a target travel route coordinate sequence. The optimization process considers the actual conditions of the road, such as a single-way road, turning limitation and the like, and uses an A-algorithm to conduct path re-planning. And carrying out road section segmentation processing on the target driving route coordinate sequence to obtain a road section unit set. The segmentation divides the whole route into a plurality of independent road segments based on the road intersection and the major inflection point. And then, carrying out fusion analysis processing on the road segment unit set and the real-time traffic flow data to obtain the estimated passing time of each road segment. This step uses the historical traffic data and real-time traffic information to calculate the estimated transit time for each road segment by a weighted average method. And carrying out accumulation calculation processing on the estimated travel time of each road section to obtain the total estimated travel time.
And carrying out time sequence prediction processing on the historical price data of the target parking lot to obtain the predicted parking cost. An ARIMA (autoregressive integral moving average) model is adopted, and parking cost in a certain future time is predicted by taking time, date, seasonality and other factors into consideration. And simultaneously, semantically processing the collected road condition data to obtain real-time road condition data. The semantic processing converts the digitized road condition information into word descriptions which are easy to understand by users, such as 'smooth', 'slight congestion', and the like. And integrating the geographic coordinates of the target parking lot, the target travel route coordinate sequence, the estimated travel time, the estimated parking cost and the real-time road condition data into target parking advice path data. This integration creates a structured data object containing all the necessary information for subsequent presentation and use.
For example: a user searches for a parking space near a commercial area using the intelligent parking guidance system during the afternoon of Saturday. Firstly, user satisfaction data of 10 peripheral parking lots are analyzed, a satisfaction threshold is set to be 4.2 minutes (5 minutes), and 5 high-satisfaction parking lots are screened out. After the geographic information extraction, the coordinates of the 5 parking lots are mapped onto a two-dimensional plane. The match screening process considers the satisfaction rank and distance from the destination, and finally selects parking lot a with coordinates (39.9153,116.4038) as the target parking lot.
The initial navigation route comprises 100 coordinate points, and 25 key points are reserved after the initial navigation route is smoothed by a Douglas-Peucker algorithm. The algorithm optimization considers the actual road condition, such as single-way road limitation, and generates a final target driving route coordinate sequence. The route is divided into 8 road segment units, and the estimated transit time of each road segment is calculated by combining the real-time traffic flow data. For example, the first segment is expected to take 3 minutes, the second segment 5 minutes, and so on. The cumulative calculation gives a total estimated travel time of 22 minutes. The ARIMA model analyzes the price data of parking lot A for 3 months, and predicts a parking fee of 15 yuan/hour in the afternoon. The real-time road condition data is semantically processed and displayed, wherein 70% of road sections in the route are smooth, 20% of road sections are slightly congested, and 10% of road sections are moderately congested. Finally, the generated target parking advice path data includes: coordinates (39.9153,116.4038) of the target parking lot A, a driving route coordinate sequence consisting of 25 key points, estimated driving time of 22 minutes, estimated parking cost of 15 yuan/hour and detailed real-time road condition information.
The method for guiding intelligent parking in the embodiment of the present application is described above, and the system for guiding intelligent parking in the embodiment of the present application is described below, referring to fig. 2, an embodiment of the system for guiding intelligent parking in the embodiment of the present application includes:
The processing module 201 is configured to perform real-time collection and standardization processing on the multi-source traffic data to obtain a target traffic data set including traffic flow, availability of parking spaces, environmental factors and historical trends;
the identifying module 202 is configured to perform time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow data and parking demand prediction data;
the optimizing module 203 is configured to perform multi-objective optimization processing on the parking requirement prediction data based on the traffic flow data to obtain objective decision data, where the objective decision data includes: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data;
The conversion module 204 is configured to perform multi-channel information format conversion and priority ranking processing on the target decision data to obtain real-time information data applicable to different release platforms;
The planning module 205 is configured to collect user parking behavior data, real-time traffic condition data, and parking lot occupancy data in real time, and perform association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data, and the parking lot occupancy data to obtain candidate parking suggestion path data including a target parking lot and a navigation route;
The analysis module 206 is configured to perform multidimensional cross analysis processing on the user feedback data and the occupancy data acquired in real time, so as to obtain evaluation index data, where the evaluation index data includes: user satisfaction and resource utilization efficiency;
a correction module 207, configured to perform data correction on the candidate parking advice path data according to the evaluation index data, to obtain target parking advice path data, where the target parking advice path data includes: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
Through the cooperation of the components, the multi-source traffic data is acquired and standardized in real time, so that a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trends is obtained, the timeliness of the data is improved, and the reliability of decision making is enhanced. And carrying out time sequence analysis and pattern recognition processing on the target traffic data set to obtain traffic flow data and parking demand prediction data, and predicting traffic conditions and parking demands in a period of time in the future so as to make reasonable resource allocation in advance. The method comprises the steps of carrying out multi-target optimization processing on parking demand prediction data based on traffic flow data to obtain target decision data comprising traffic flow distribution schemes, parking resource allocation strategies and vehicle path recommendation data, carrying out multi-channel information format conversion and priority ordering processing on the target decision data to obtain real-time information data suitable for different release platforms, and greatly improving information propagation efficiency and coverage, so that a user can obtain parking guide information through various devices and platforms. User behavior data, real-time traffic condition data and parking lot occupancy rate data are collected in real time, association analysis and path planning processing are carried out, candidate parking suggestion path data containing target parking lots and navigation routes are obtained, accurate response to personalized demands of users is achieved, real-time road conditions are considered, and recommendation accuracy and practicality are improved. The method comprises the steps of carrying out multidimensional cross analysis processing on user feedback data and occupancy rate data acquired in real time to obtain evaluation index data comprising user satisfaction and resource utilization efficiency, carrying out data correction on candidate parking advice path data through the evaluation index data to obtain target parking advice path data comprising target parking lot geographic coordinates, target driving route coordinate sequences, predicted driving time, predicted parking fees and real-time road condition data, and realizing full-flow intellectualization from data acquisition to end user recommendation through multi-step data processing and decision optimization, thereby greatly improving parking efficiency, reducing time and cost for searching parking spaces by users, and improving timeliness and accuracy of intelligent parking guidance.
The present application also provides a computer readable storage medium, which may be a non-volatile computer readable storage medium, and may also be a volatile computer readable storage medium, in which instructions are stored which, when executed on a computer, cause the computer to perform the steps of the intelligent parking guidance method.
It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, systems and units may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims (10)
1. An intelligent parking guidance method, characterized in that the intelligent parking guidance method comprises:
Carrying out real-time acquisition and standardization processing on the multi-source traffic data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend;
performing time sequence analysis and pattern recognition processing on the target traffic data set to obtain vehicle flow prediction data and parking demand prediction data;
performing multi-objective optimization processing on the parking demand prediction data based on the traffic flow prediction data to obtain objective decision data, wherein the objective decision data comprises: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data;
Performing multi-channel information format conversion and priority ranking processing on the target decision data to obtain real-time information data applicable to different release platforms;
collecting user parking behavior data, real-time traffic condition data and updated parking lot occupancy rate data in real time, and carrying out association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data and the parking lot occupancy rate data to obtain candidate parking suggestion path data comprising a target parking lot and a navigation route;
Performing multidimensional cross analysis processing on the user feedback data and the occupancy rate data acquired in real time to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency;
And carrying out data correction on the candidate parking suggested path data through the evaluation index data to obtain target parking suggested path data, wherein the target parking suggested path data comprises the following components: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
2. The intelligent parking guidance method of claim 1, wherein the multi-source traffic data includes road surface vehicle density data, vehicle passing data at the entrance and exit of the parking lot, weather condition data, special event information, and historical parking data, and the real-time collection and standardization of the multi-source traffic data results in a target traffic data set including traffic flow, availability of parking spaces, environmental factors, and historical trends, comprising:
performing dynamic threshold segmentation processing on the road surface vehicle density data to obtain vehicle flow classification data reflecting traffic flow;
Performing difference value calculation processing on the vehicle passing data at the entrance and the exit of the parking lot to obtain real-time parking space occupancy rate data;
Carrying out semantic analysis processing on the weather condition data and the special event information to obtain environmental impact factor data;
performing periodic decomposition processing on the historical parking data to obtain parking trend periodic data;
Performing space-time correlation analysis processing on the traffic flow classification data and the real-time parking space occupancy rate data to obtain parking demand hot spot distribution data;
Carrying out weighted fusion processing on the environmental impact factor data and the parking trend period data to obtain parking behavior prediction basic data;
Performing density clustering processing on the parking demand hot spot distribution data to obtain regional parking pressure evaluation data;
performing time sequence interpolation processing on the parking behavior prediction basic data to obtain parking demand prediction data in a continuous time domain;
Performing self-adaptive weight distribution processing on the regional parking pressure evaluation data and the parking demand prediction data to obtain comprehensive parking index data;
And carrying out multi-scale normalization processing on the comprehensive parking index data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend.
3. The intelligent parking guidance method of claim 1, wherein the performing time-series analysis and pattern recognition on the target traffic data set to obtain traffic flow prediction data and parking demand prediction data comprises:
Performing time window segmentation processing on the target traffic data set to obtain a plurality of parking data fragments with time granularity, and performing Fourier transform processing on the parking data fragments with the time granularity to obtain parking periodic characteristic data;
Performing wavelet decomposition processing on the parking periodic characteristic data to obtain multi-scale parking trend data, and performing time sequence prediction processing on the multi-scale parking trend data through a long-short-term memory network algorithm to obtain initial parking demand prediction data;
Performing correlation analysis processing on the initial parking demand prediction data and real-time traffic flow data acquired in real time to obtain a flow-demand correlation factor, and performing dynamic weight distribution processing on the flow-demand correlation factor to obtain parking demand adjustment parameters;
Performing self-adaptive fusion processing on the initial parking demand prediction data and the parking demand adjustment parameters to obtain optimized parking demand prediction data, and performing trend extraction and seasonal decomposition processing on historical traffic flow data to obtain traffic flow basic characteristic data;
And performing anomaly detection processing on the traffic flow basic characteristic data and the real-time traffic event data to obtain traffic flow fluctuation factors, and performing combined prediction processing on the traffic flow basic characteristic data and the traffic flow fluctuation factors to obtain traffic flow prediction data and parking demand prediction data.
4. The intelligent parking guidance method of claim 1, wherein the performing multi-objective optimization processing on the parking demand prediction data based on the traffic prediction data to obtain objective decision data, wherein the objective decision data includes: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data, comprising:
Performing network flow model construction processing on the traffic flow prediction data to obtain a traffic flow network topology structure, and performing spatial clustering processing on the parking demand prediction data to obtain parking demand hot spot area data;
performing matching degree calculation processing on the traffic flow network topological structure and the parking demand hot spot area data to obtain an initial traffic flow distribution scheme, and performing capacity constraint optimization processing on the initial traffic flow distribution scheme to obtain the traffic flow distribution scheme;
Performing differential analysis processing on the existing parking resource data and the parking demand prediction data to obtain parking resource gap data, and performing dynamic allocation processing on the parking resource gap data to obtain an initial parking resource allocation strategy;
performing conflict detection processing on the initial parking resource allocation strategy to obtain the parking resource allocation strategy;
carrying out fusion processing on the real-time road network state data and the traffic flow distribution scheme to obtain a multidimensional path evaluation index;
Carrying out path search processing on the multi-dimensional path evaluation indexes through a particle swarm optimization algorithm to obtain a candidate vehicle path set;
And carrying out multi-objective weighing processing on the candidate vehicle path set to obtain the vehicle path recommended data, and integrating the traffic flow distribution scheme, the parking resource allocation strategy and the vehicle path recommended data into the objective decision data.
5. The intelligent parking guidance method of claim 4, wherein the performing multi-channel information format conversion and prioritization on the target decision data to obtain real-time information data applicable to different distribution platforms comprises:
carrying out data compression processing on the traffic flow distribution scheme to obtain traffic flow summary data;
carrying out space index construction processing on the parking resource allocation strategy to obtain parking space distribution data;
performing path simplification processing on the vehicle path recommended data to obtain critical path node data;
Carrying out data fusion processing on the traffic flow summary data, the parking space distribution data and the critical path node data to obtain an information data packet;
Performing multi-level cache policy processing on the information data packet to obtain a layered storage information cache structure, and performing real-time update frequency analysis processing on the information cache structure to obtain information timeliness scoring data;
The information timeliness scoring data is subjected to priority ranking processing to obtain an information release queue, and the data in the information release queue is subjected to multi-channel adaptation processing to obtain formatting information aiming at different release platforms;
And carrying out secure encryption processing on the formatted information aiming at different release platforms to obtain the real-time information data applicable to the different release platforms.
6. The intelligent parking guidance method of claim 1, wherein the collecting the user parking behavior data, the real-time traffic condition data, and the parking lot occupancy data in real time, and performing association analysis and path planning processing on the user parking behavior data, the real-time traffic condition data, and the parking lot occupancy data to obtain candidate parking advice path data including a target parking lot and a navigation route, comprises:
Performing cluster analysis processing on the user parking behavior data to obtain user parking preference information, and performing road congestion degree calculation processing on the traffic condition data to obtain a traffic fluency index;
carrying out difference value statistics on the parking lot occupancy rate data to obtain a dynamic parking space occupancy rate;
Performing correlation analysis processing on the user parking preference signal and the dynamic parking space occupancy rate to obtain a personalized parking lot recommendation list;
threshold segmentation processing is carried out on the traffic fluency index to obtain a feasible driving road section set;
Carrying out path planning processing on the feasible driving road section set through a heuristic search algorithm to obtain a plurality of candidate navigation routes;
Carrying out geographic coordinate extraction processing on the parking lots in the personalized parking lot recommendation list to obtain a target parking lot coordinate set;
Performing matching degree calculation processing on the plurality of candidate navigation routes and the target parking lot coordinate set to obtain a route-parking lot matching scheme;
And carrying out comprehensive scoring processing on the route-parking lot matching scheme to obtain an optimal parking suggested path, and carrying out data packaging processing on the optimal parking suggested path to obtain candidate parking suggested path data containing a target parking lot and a navigation route.
7. The intelligent parking guidance method of claim 6, wherein the multi-dimensional cross analysis is performed on the user feedback data and the occupancy data acquired in real time to obtain evaluation index data, wherein the evaluation index data includes: user satisfaction and resource utilization efficiency, including:
carrying out emotion analysis processing on the user feedback data to obtain initial scores of the parking experience satisfaction, and carrying out time attenuation processing on the initial scores of the parking experience satisfaction to obtain timeliness weighted satisfaction data;
Performing time sequence decomposition processing on the parking lot occupancy rate data to obtain parking lot utilization efficiency trend data, and performing correlation analysis processing on the timeliness weighted satisfaction degree data and the parking lot utilization efficiency trend data to obtain satisfaction degree-efficiency correlation indexes;
Carrying out distribution feature extraction processing on the parking time length data of the user to obtain parking time length mode data, and carrying out matching degree calculation processing on the parking time length mode data and the turnover rate data of the parking lot to obtain space-time utilization efficiency indexes;
Carrying out fusion processing on the satisfaction-efficiency association index and the space-time utilization efficiency index to obtain a comprehensive evaluation initial value, and carrying out statistical analysis processing on parking guidance response time data to obtain a response efficiency score;
Performing weighted average processing on the comprehensive evaluation initial value and the response efficiency score to obtain target evaluation index original data;
normalizing and quantifying the target evaluation index original data to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency.
8. The intelligent parking guidance method of claim 7, wherein the data correction is performed on the candidate parking advice path data by the evaluation index data to obtain target parking advice path data, wherein the target parking advice path data includes: the target parking lot geographic coordinates, the target travel route coordinate sequence, the predicted travel time, the predicted parking cost, and the real-time road condition data include:
Performing threshold analysis processing on the user satisfaction in the evaluation index data to obtain a parking lot satisfaction sorting list;
carrying out geographic information extraction processing on the target parking lot in the candidate parking suggestion path data to obtain a candidate parking lot coordinate set, and carrying out matching screening processing on the parking lot satisfaction ordered list and the candidate parking lot coordinate set to obtain target parking lot geographic coordinates;
Carrying out path smoothing processing on the navigation route in the candidate parking suggestion path data to obtain an initial driving route coordinate sequence, and carrying out optimization processing on the initial driving route coordinate sequence to obtain a target driving route coordinate sequence;
Carrying out road section segmentation processing on the target driving route coordinate sequence to obtain a road section unit set, and carrying out fusion analysis processing on the road section unit set and real-time traffic flow data to obtain predicted passing time of each road section;
performing accumulation calculation processing on the estimated travel time of each road section to obtain the estimated travel time;
performing time sequence prediction processing on the historical price data of the target parking lot to obtain predicted parking cost;
Semantically processing the collected road condition data to obtain the real-time road condition data, and integrating the geographic coordinates of the target parking lot, the coordinate sequence of the target driving route, the estimated driving time, the estimated parking cost and the real-time road condition data into the target parking suggestion path data.
9. An intelligent parking guidance system for performing the intelligent parking guidance method of any one of claims 1-7, the intelligent parking guidance system comprising:
The processing module is used for carrying out real-time acquisition and standardization processing on the multi-source traffic data to obtain a target traffic data set containing traffic flow, parking space availability, environmental factors and historical trend;
the recognition module is used for carrying out time sequence analysis and pattern recognition processing on the target traffic data set to obtain vehicle flow data and parking demand prediction data;
The optimizing module is used for carrying out multi-objective optimizing processing on the parking demand prediction data based on the traffic flow data to obtain objective decision data, wherein the objective decision data comprises: traffic flow allocation scheme, parking resource allocation strategy and vehicle path recommendation data;
the conversion module is used for carrying out multi-channel information format conversion and priority ordering processing on the target decision data to obtain real-time information data suitable for different release platforms;
The planning module is used for collecting the parking behavior data of the user, the real-time traffic condition data and the parking lot occupancy rate data in real time, and carrying out association analysis and path planning processing on the parking behavior data of the user, the real-time traffic condition data and the parking lot occupancy rate data to obtain candidate parking suggestion path data comprising a target parking lot and a navigation route;
the analysis module is used for carrying out multidimensional cross analysis processing on the user feedback data and the occupancy rate data acquired in real time to obtain evaluation index data, wherein the evaluation index data comprises: user satisfaction and resource utilization efficiency;
The correction module is used for carrying out data correction on the candidate parking suggested path data through the evaluation index data to obtain target parking suggested path data, wherein the target parking suggested path data comprises: the system comprises a target parking lot geographic coordinate, a target travel route coordinate sequence, a predicted travel time, a predicted parking cost and real-time road condition data.
10. A computer readable storage medium having instructions stored thereon, which when executed by a processor, implement the intelligent parking guidance method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411122694.6A CN118824043A (en) | 2024-08-15 | 2024-08-15 | Intelligent parking guiding method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411122694.6A CN118824043A (en) | 2024-08-15 | 2024-08-15 | Intelligent parking guiding method, system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118824043A true CN118824043A (en) | 2024-10-22 |
Family
ID=93069672
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411122694.6A Pending CN118824043A (en) | 2024-08-15 | 2024-08-15 | Intelligent parking guiding method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118824043A (en) |
-
2024
- 2024-08-15 CN CN202411122694.6A patent/CN118824043A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yao et al. | Short‐term traffic speed prediction for an urban corridor | |
| Liu et al. | Intelligent bus routing with heterogeneous human mobility patterns | |
| Zhang et al. | A framework for passengers demand prediction and recommendation | |
| Nocera et al. | Assessing carbon emissions from road transport through traffic flow estimators | |
| RU2406158C2 (en) | Methods of predicting destinations from partial trajectories employing open- and closed-world modeling methods | |
| WO2016124118A1 (en) | Order processing method and system | |
| Lyu et al. | Research on the big data of traditional taxi and online car-hailing: A systematic review | |
| Ionita et al. | Where to park? predicting free parking spots in unmonitored city areas | |
| Liu et al. | Exploiting heterogeneous human mobility patterns for intelligent bus routing | |
| CN111582559B (en) | Arrival time estimation method and device | |
| CN115790636B (en) | Unmanned retail vehicle cruise path planning method and device based on big data | |
| Huang et al. | Investigation of bimodal macroscopic fundamental diagrams in large-scale urban networks: empirical study with GPS data for Shenzhen city | |
| Duddu et al. | Principle of demographic gravitation to estimate annual average daily traffic: Comparison of statistical and neural network models | |
| Deng et al. | Heterogenous Trip Distance‐Based Route Choice Behavior Analysis Using Real‐World Large‐Scale Taxi Trajectory Data | |
| Saha et al. | Deep learning approach for predictive analytics to support diversion during freeway incidents | |
| Yun et al. | Minimum sampling size of floating cars for urban link travel time distribution estimation | |
| Cerqueira et al. | On how to incorporate public sources of situational context in descriptive and predictive models of traffic data | |
| Lee et al. | Taxi vacancy duration: a regression analysis | |
| Nicoletta et al. | Bayesian spatio-temporal modelling and prediction of areal demands for ambulance services | |
| Liu et al. | Short-term online taxi-hailing demand prediction based on the multimode traffic data in metro station areas | |
| Yu et al. | Combining travel behavior in metro passenger flow prediction: A smart explainable Stacking-Catboost algorithm | |
| Chen et al. | A Data-driven Region Generation Framework for Spatiotemporal Transportation Service Management | |
| Bing et al. | Integrating semantic zoning information with the prediction of road link speed based on taxi GPS data | |
| Phiboonbanakit et al. | Analyzing Bangkok city taxi ride: reforming fares for profit sustainability using big data driven model | |
| Liang et al. | Modeling taxi cruising time based on multi-source data: a case study in Shanghai |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |