CN107121690B

CN107121690B - A kind of dwell point recognition methods and device based on parameter of doing more physical exercises

Info

Publication number: CN107121690B
Application number: CN201710209844.0A
Authority: CN
Inventors: 陈德军; 钟鸣; 熊壮
Original assignee: Huitong Landscape Engineering Consulting (wuhan) Co Ltd
Current assignee: Huitong Landscape Engineering Consulting (wuhan) Co Ltd
Priority date: 2017-03-31
Filing date: 2017-03-31
Publication date: 2019-05-28
Anticipated expiration: 2037-03-31
Also published as: CN107121690A

Abstract

The present invention relates to intelligent terminal technical field, a kind of dwell point recognition methods based on parameter of doing more physical exercises and device are provided.Method includes: when judging the align_type of current collection point for network positions type, to obtain collection point in preset duration t1；Count the accounting rate of network positions type collection point and positioning accurate angle value change degree in the collection point；If it is determined that the accounting rate of network positions type collection point is greater than default network accounting threshold value, and positioning accurate angle value change degree is less than default precision threshold, then judges current collection point for dwell point.The embodiment of the present invention is using the accounting of network positions collection point and to determine whether present terminal enters in the region with shielding GPS signal characteristic, GPS is switched to Network and positions this principle when being entered the room using user, using the last one the GPS available point entered the room as the physical location of dwell point, efficiently solve the problems, such as that accuracy is insufficient in the prior art.

Description

A kind of dwell point recognition methods and device based on parameter of doing more physical exercises

[technical field]

The present invention relates to intelligent terminal technical fields, more particularly to a kind of dwell point identification side based on parameter of doing more physical exercises Method and device.

[background technique]

Questionnaire survey, telephone interview be conventional survey resident trip main method, but research cost height, heavy workload, It is difficult that data summarization and long processing period, investigation content influenced by subjective consciousness, survey data is inaccurate etc., is intended to always The critical issue being resolved.

With the development of information technology, occurs track of going on a journey using Informationization Equipment record user in recent years, and carry out Dwell point knows method for distinguishing, for replacing traditional trip survey method.Since these methods overcome traditionally on paper and phone Access cost is high, and the problem that survey data Feasible degree leaves a question open, accuracy is not high has obtained the favor of researcher, has carried out big The in-depth study of amount.Document in recent years is concluded, the passive type trip that this kind of research method can be attributed to two quasi-representatives is adjusted Checking method: one kind is the trip track using dedicated GPS device or cellphone GPS function record resident, then to the data of acquisition The method for carrying out traffic start-stop point (ORIGIN DESTINATION, writes a Chinese character in simplified form are as follows: OD) investigation；Another kind of is using based on mobile phone bee The analysis of nest network positions data carries out user's trip and stops point analysis etc..

Outdoor dwell point recognition methods based on GPS, current main method have the exploration for relying on data acquisition experience Method and the clustering method for relying on data mining technology.Wherein heuristic approaches using it is more be rest point method, principle is with GPS Based on receiver still carries out record when stop motion speed is zero, using the walking speed of people as fiducial value, speed is determined Degree is less than the duration of walking speed, this duration is compared with time threshold T, is then judged at this if it is greater than T To stop.Such as: Zhang Bo " the GPS space-time trajectory data for traffic trip investigation is simplified and semantically enhancement is studied " Shanghai: China Eastern normal university's resource and environmental science institute, 2011 (subsequent to be referred to as " Zhang Bo (2011) ") they are setting rules are as follows: if 1. The difference of longitude or latitude between mobile target adjacent track is maintained in 0.00005 (about 15), and spatial position is maintained at half Within the scope of 50 meters of diameter；2. average speed is less than 0.6m/s；3. the state duration is greater than 120 seconds as dwell points.Also for example: Gong H,Chen C,Bialostozky E,et al.A GPS/GIS method for travel mode,detection in New York City[J].Computers,Environment and Urban Systems,2012,36(2):131- In 139 (subsequent to be referred to as " Gong H (2012) ") setting 200s, the distance between tracing point is less than 50m and is judged to stopping. The shortcomings that the method is due to being generated and fail to judge to the data comprising noise spot or shift point using the duration as threshold value.

Common clustering algorithm in the clustering method of data mining technology is relied on to have K- median method and DBSCAN.

K- median method, first, it is specified that least points m and cluster radius d is counted since first point in track in cluster The maximum distance in continuous m point between any two points is calculated, if the distance is less than d, a cluster is established, judges the middle position of the cluster The point is then added in cluster if it is less than d/2, otherwise terminates the cluster by the distance between next point outside point and cluster, final to establish Each cluster labeled as stop.The deficiency of algorithm is that the distance for using farthest two o'clock in cluster defines a cluster, vulnerable to making an uproar Sound shadow is rung.

The basic principle of DBSCAN algorithm is to choose cluster radius d, determines the tracing point threshold value n within the scope of cluster radius. Then it is successively calculated since the starting point of track centered on calculating point, with the track number in the range of the cluster radius d of selection Strong point number m.Calculated result is compared with threshold value, if calculated result m is less than threshold value n, which is considered as noise spot, otherwise A Density Cluster is just established centered on the point.The disadvantages of this method is that the shortage of data of GPS is affected to cluster result. Such as: Liu Chun traveler trip information semanteme excavates Shandong Technology Univ, 2015 (subsequent to be referred to as " Liu Chun (2015) ") fortune It is the characteristic of event using track data with DBSCAN algorithm, does not use the minimum track points MinPts conduct of close quarters Parameter, but replaced using shortest time MinTime, has calculated low velocity region, then according to it is actual electronically Scheme the building obtained and scenic spot, determines whether to stop using based on the algorithm of the time of coincidence.Inventor passes through actual test GPS signal is blocked or generates larger drift when the shortcomings that middle discovery method is most of stops indoors, so in practice simultaneously It will not fall in construction zone, cause the stop in indoors that can not detect, such as Fig. 1, solid stain is indoor positioning, only There is fewer parts point to fall in construction zone.

The typical recognition methods of dwell point based on cellular network location technology is then CELL-ID method (also known as Cell of Origin, COO), such as: user trip and park point identifying system research Beijing work of the Du Runqiang based on mobile phone track data Sparetime university is learned, 2014 (referred to as subsequent " Du Runqiang (2014) ") for judged by accident caused by irregular mobile phone location switching phenomenon into Go improvement, it is fixed to be easy to causeing switching and function or the adjacent cell of background geography information attribute analogous location merge Justice is functional areas, switches range when functional areas inscribe reaches time threshold when user mobile phone positions, i.e., regards entire functional areas Make the park point of user.Since CELL-ID positioning accuracy is at hundred meters or more, and combined functional areas are even more big, cause final The park point Information Granularity of acquisition is larger, can not get the thinner dwell point of granularity.

In addition, travel activity chain pattern Study of recognition [D] Dalian University Of Communications of the Wang Lei based on the track LBS, 2015 (after It is continuous to be referred to as " Wang Lei (2015) ") pass through the Euclidean distance average value that largely sampling calculates walking point and rest point, and with this Rest point and walking point are illustrated for threshold value, and the threshold value of rest point is 0.003234, and the threshold value of walking point is 0.03484.In reality Euclidean distance in the case of border, which is estimated, can divide in strict accordance with threshold value, since there is fluctuations for its data, will cause mistake Sentence or fail to judge, first is that the diversity of mode of transportation and the complexity of movement, reason two are that GPS data exists the reason of fluctuation Error and unstability.Although it has modified other modes of transportation by introducing the average smoothing method of sliding window to a certain extent The fluctuation problem of middle contingency, but since the Evaluation criterion of this method is overly dependent upon estimating for Euclidean distance, the measurement on GPS It fails to judge and misjudges caused by error and be not better to improve.

When inventor has found that user is in indoor stop, it is easy to appear GPS and network positions and does not stop mutually to switch, speed There is large change.The Euclidean distance figure of this sample is calculated with the Euclidean distance threshold formula of Wang Lei (2015), as shown in Fig. 2, From figure 2 it can be seen that this time really resting under Euclidean distance algorithm can be failed to judge due to the unstability of positioning.

This is the scene that a kind of typical indoor stop but track data show as movement, and this GPS is easily affected by environment Generate it is unstable be therefore the limitation of location data itself under the conditions of existing, needs reliable to move by other Parameter solves the problems, such as this.

From above-mentioned dwell point recognition methods it can be found that the trip survey method based on GPS technology and cellular network technologies All it is that data are obtained by digitizer, then analyzes the characteristics of obtaining data, thus reason out the trip of resident Mode and dwell point, it is clear that the correctness of its identification model will directly affect the accuracy of analysis result.Simultaneously as GPS believes Number only outdoor could obtain, therefore can only identify outdoor dwell point；And the recognition result granularity and error of mobile phone cellular information It is larger, it has difficulties to accurately identifying for dwell point of trip.Simultaneously for particular context such as GPS drift about, the above method there is also Wretched insufficiency also constitutes the practicality and significantly challenges.

[summary of the invention]

Technical problems to be solved of the embodiment of the present invention are to improve the accuracy of dwell point identification in the prior art.

The further technical problems to be solved of the embodiment of the present invention are to be directed to not only to have network type collection point, but also have Under the hybird environment of gPS class type collection point, how dwell point is effectively identified.The embodiment of the present invention uses following technical side Case:

In a first aspect, the embodiment of the invention also provides a kind of dwell point recognition methods based on parameter of doing more physical exercises, it is corresponding Each collection point possesses respective align_type, the judgement of carry out one acquisition dotted state of terminal periodic, which comprises

Obtain current collection point；

When determining the current collection point is network positions type, judge to obtain terminal institute before the current collection point Place's state；

If being judged as motion state, collection point in preset duration t1 is obtained；Wherein, collection point includes in preset duration t1 The collection point of GPS positioning type and the collection point of network positions type；Count network positions type collection point in the collection point Accounting rate and positioning accurate angle value change degree；It is accounted for if it is determined that the accounting rate of network positions type collection point is greater than default network Than threshold value, and positioning accurate angle value change degree is less than default precision threshold, then judges current collection point for dwell point；If being judged as Resting state, it is determined that the current acquisition point location coordinate is at a distance from previous dwell point, wherein distance is greater than pre-determined distance When, then modifying state is motion state, and otherwise hold mode is resting state；

When determining the current collection point is GPS positioning type, judge to obtain terminal institute before the current collection point Place's state；

If being judged as motion state, the set being made of in preset duration t2 GPS positioning type collection point is obtained, for Collection point is less than pre-set radius distance to cluster centre distance in the set, then belongs to the collection point in corresponding sub-clustering；System The GPS positioning type collection point number in the sub-clustering is counted relative to GPS positioning type collection point total in preset duration t2 Accounting, and arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point sequence；If described Accounting result is greater than default GPS accounting threshold value, and the default quantile in the GPS gathers point sequence is less than default GPS quartile Threshold value then judges current collection point for dwell point；If being judged as resting state, hold mode is resting state.

Optionally, the accounting rate and positioning accuracy for counting network positions type collection point in the collection point is being executed It is worth before change degree operation, the method also includes:

Collection point in preset duration t1 is successively divided at least two sections according to acquisition time, it is calculated to each section and adds The standard deviation of speed modulus value arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard difference sequence；

The default quantile in the standard difference sequence is taken, if the default quantile is less than preset standard differential bit threshold Value, then directly judge current collection point for dwell point；Otherwise network positions type acquisition in the statistics collection point is executed The accounting rate of point and the operation of positioning accurate angle value change degree and its subsequent content judge dwell point.

Optionally, before executing to set progress clustering operation, the method also includes:

Collection point in preset duration t2 is successively divided at least two sections according to acquisition time, it is calculated to each section and adds The standard deviation of speed modulus value arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard difference sequence；

The default quantile in the standard difference sequence is taken, if the default quantile is less than preset standard differential bit threshold Value, then directly judge current collection point for dwell point；Otherwise it executes and set progress clustering operation and its subsequent content is come Judge dwell point.

Second aspect, the embodiment of the invention provides a kind of dwell point recognition methods based on parameter of doing more physical exercises, collection points It is stored after being demarcated as GPS positioning type or network positions type, method includes:

When judging the align_type of current collection point for network positions type, collection point in preset duration t1 is obtained；Wherein, Collection point includes the collection point of GPS positioning type and/or the collection point of network positions type in preset duration t1；

Count the accounting rate of network positions type collection point and positioning accurate angle value change degree in the collection point；If it is determined that institute The accounting rate for stating network positions type collection point is greater than default network accounting threshold value, and positioning accurate angle value change degree is less than default Precision threshold then judges current collection point for dwell point.

Optionally, for being determined as one group of collection point of dwell point, take the current acquisition point location coordinate adjacent recently Correspondence have interference GPS positioning signal object place-centric be current collection point the position presented on map.

The third aspect, the embodiment of the invention provides a kind of dwell point recognition methods based on parameter of doing more physical exercises, collection points It is stored after being demarcated as GPS positioning type or network positions type, method includes:

When judging the align_type of current collection point for GPS positioning type, obtain in preset duration t2 by GPS positioning type The set that collection point is constituted carries out sub-clustering to the set；Wherein, clustering process includes:

Pre-set radius distance is less than for collection point in the set to cluster centre distance, then the collection point is belonged into phase It answers in sub-clustering；

The GPS positioning type collection point number in the sub-clustering is counted relative to GPS positioning class total in preset duration t2 The accounting of type collection point, and arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point sequence Column；

If the accounting result is greater than default GPS accounting threshold value, and the default quantile in the GPS gathers point sequence is small In default GPS quartile threshold value, then judge current collection point for dwell point.

Fourth aspect, it is described the embodiment of the invention also provides a kind of dwell point identification device based on parameter of doing more physical exercises Device includes that collection point obtains module, align_type judgment module, network positions type processing module, the processing of GPS positioning type Module, wherein collection point obtains module and connects the align_type judgment module, and the align_type judgment module is separately connected The network positions type processing module and GPS positioning type processing module, specific:

The collection point obtains module, for obtaining current collection point；

The align_type judgment module, for when determining the current collection point is network positions type, judgement to be obtained Take terminal status before the current collection point；

The network positions type processing module is used for, and when being judged as motion state, obtains acquisition in preset duration t1 Point；Wherein, collection point includes the collection point of GPS positioning type and the collection point of network positions type in preset duration t1；Statistics The accounting rate of network positions type collection point and positioning accurate angle value change degree in the collection point；If it is determined that the network positions class The accounting rate of type collection point is greater than default network accounting threshold value, and positioning accurate angle value change degree is less than default precision threshold, then Judge current collection point for dwell point；If being judged as resting state, it is determined that the current acquisition point location coordinate stops with previous The distance at stationary point, wherein when distance is greater than pre-determined distance, then modifying state is motion state, and otherwise hold mode is to stop shape State；

The align_type judgment module is also used to when determining the current collection point is GPS positioning type, and judgement obtains Take terminal status before the current collection point；

GPS positioning type processing module, for obtaining in preset duration t2 by GPS positioning when being judged as motion state The set that type collection point is constituted is less than pre-set radius distance for collection point in the set to cluster centre distance, then should Collection point belongs in corresponding sub-clustering；The GPS positioning type collection point number in the sub-clustering is counted relative to preset duration t2 The accounting of interior total GPS positioning type collection point, and arrange each collection point in the preset duration t2 from small to large according to speed, Obtain GPS gathers point sequence；If the accounting result is greater than default GPS accounting threshold value, and pre- in the GPS gathers point sequence If quantile is less than default GPS quartile threshold value, then judge current collection point for dwell point；If being judged as resting state, keep State is resting state.

In order to overcome GPS sensor unavailable indoors and the case where outdoor is blocked.It is fixed that invention introduces mixing Position technology, while being carried out to solve the feature disunity of GPS positioning data and network positions data under hybrid positioning technology Particular analysis and feature extraction, have been finally introducing the unreliability that acceleration transducer compensates for location data, integrated use The outdoor and indoor dwell point identification model of above-mentioned condition building, shows dwell point proposed in this paper by examples comparative and actual measurement Recognizer accuracy with higher and practicability preferably solve existing for current trip survey dwell point recognition methods Problem is worth with importantly theory and actual application.

[Detailed description of the invention]

In order to illustrate the embodiment of the utility model or the technical proposal in the existing technology more clearly, below will be to embodiment Or attached drawing needed to be used in the description of the prior art is briefly described, it should be apparent that, the accompanying drawings in the following description is only It is some embodiments of the utility model, for those of ordinary skill in the art, in the premise not made the creative labor Under, it can also be obtained according to these attached drawings other attached drawings.

Anchor point distributed effect figure when Fig. 1 is a kind of indoor stop provided in an embodiment of the present invention；

Fig. 2 is typical indoor stop provided in an embodiment of the present invention but track data shows as exercise data Euclidean distance Relational graph；

Fig. 3 is a kind of network side dwell point recognition methods process based on parameter of doing more physical exercises provided in an embodiment of the present invention Figure；

Fig. 4 is a kind of network side dwell point identification improved method flow chart provided in an embodiment of the present invention；

Fig. 5 is a kind of side GPS dwell point recognition methods flow chart based on parameter of doing more physical exercises provided in an embodiment of the present invention；

Fig. 6 is a kind of side GPS dwell point identification improved method flow chart provided in an embodiment of the present invention；

Fig. 7 is another dwell point recognition methods flow chart based on parameter of doing more physical exercises provided in an embodiment of the present invention；

Fig. 8 is another dwell point recognition methods flow chart based on parameter of doing more physical exercises provided in an embodiment of the present invention；

Fig. 9 is a kind of data acquisition interface effect picture provided in an embodiment of the present invention；

Figure 10 is some experimental data display diagram provided in an embodiment of the present invention；

Window sample point acceleration sampled value curve graph when Figure 11 is provided in an embodiment of the present invention static；

Each section of acceleration standard dygoram of window when Figure 12 is provided in an embodiment of the present invention static；

Window sample point acceleration sampled value curve graph when Figure 13 is movement provided in an embodiment of the present invention；

Each section of acceleration standard dygoram of window when Figure 14 is movement provided in an embodiment of the present invention；

Window sample point precision sampled value curve graph when Figure 15 is provided in an embodiment of the present invention static；

Window sample point precision sampled value curve graph when Figure 16 is provided in an embodiment of the present invention mobile.

[specific embodiment]

In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and It is not used in the restriction present invention.

In the description of the present invention, term "inner", "outside", " longitudinal direction ", " transverse direction ", "upper", "lower", "top", "bottom" etc. refer to The orientation or positional relationship shown be based on the orientation or positional relationship shown in the drawings, be merely for convenience of description the present invention rather than It is required that the present invention must be constructed and operated in a specific orientation, therefore it is not construed as limitation of the present invention.

In addition, as long as technical characteristic involved in the various embodiments of the present invention described below is each other not Constituting conflict can be combined with each other.

Embodiment 1:

The embodiment of the present invention 1 provides a kind of dwell point recognition methods based on parameter of doing more physical exercises, and collection point is demarcated as It is stored after GPS positioning type or network positions type, as shown in Figure 3, which comprises

In step 201, it when judging the align_type of current collection point for network positions type, obtains in preset duration t1 Collection point；Wherein, collection point includes the collection point of GPS positioning type and/or the acquisition of network positions type in preset duration t1 Point.

Wherein, current collection point can be the anchor point collected in real time by terminal；It is also possible in post analysis mistake Cheng Zhong, for picking out the anchor point for analyzing, determining whether it is dwell point in history collection point.

Wherein, preset duration t1 can be 1-5 minutes, be also possible to 10s or 30s, the setting of parameter mainly considers The cycle duration of the judgement of carry out one acquisition dotted state of the terminal periodic of the frequency acquisition and setting of collection point, and also Take into account the sum (too high to will affect terminal operation efficiency) of collection point in preset duration t1.In general control preset duration t1 Collection point quantity is 100 or so.Wherein, for the different periods, the embodiment of the invention provides two kinds of analytical models:

A kind of for segmented, i.e. front and back is twice in analytic process, in corresponding respective preset duration t1 between collection point The collection point not being overlapped, precondition are the duration that preset duration t1 is less than the period.The segmented can have The reduction calculation amount of effect, and occasion higher for frequency acquisition can effectively utilize the advantage of frequency acquisition itself.

Another kind is sliding window mode, and extreme case is exactly that cycle duration is 1s, i.e., is continuously analyzed collection point. Its disadvantage is: will increase calculation amount, still, advantage i.e. woods sensitivity are higher, also, the feelings short enough in preset duration t1 Under condition, its use occasion range can be largely improved.

In step 202, the accounting rate and positioning accurate angle value for counting network positions type collection point in the collection point become Change degree.

Wherein, for interfering stronger region in GPS signal, two kinds of align_type acquisitions in preset duration t1 will occur Point exists simultaneously situation, and the interference of its GPS signal can be wherein analyzed according to the accounting rate of network positions type collection point is It is no sufficiently strong, to estimate whether user enters dwell regions.The positioning accurate angle value is obtained from base station side.

In step 203, however, it is determined that the accounting rate of network positions type collection point is greater than default network accounting threshold value, And positioning accurate angle value change degree is less than default precision threshold, then judges current collection point for dwell point.

Wherein, presetting network accounting threshold value can be set as 0.7-0.9, and the present embodiment uses 0.8, the default precision threshold Value can be 10-40, related according to the points for including in preset duration t1, the default precision threshold meeting accordingly if the points the more What is set is larger.

The embodiment of the present invention proposes a kind of state judging method for being directed to network positions type collection point, it is suitable for Terminal is provided simultaneously under the conditions of GPS positioning and network positions, and the embodiment of the present invention is exactly analyze GPS positioning in the terminal excellent First grade is higher than network positions, i.e. network positions then can be abandoned or be neglected to terminal in the case where that can collect GPS positioning Collection point.Therefore, the embodiment of the present invention is using the accounting of network positions collection point and to determine whether entrance has present terminal In the region for shielding GPS signal characteristic, GPS is switched to Network and positions this principle when being entered the room using user, will enter It is insufficient to efficiently solve accuracy in the prior art for the last one the indoor physical location of GPS available point as dwell point The problem of.Such as the region with shielding GPS signal characteristic can be in building, in tunnel etc., wherein building type is this Inventive embodiments more concerned with object.In addition, the accuracy in order to guarantee network positions point data, it is contemplated that CELL- The neighboring community developed in ID method is to an influence for judgement is stopped, and therefore, is further considering net in embodiments of the present invention On the basis of the accounting factor of network positioning acquisition point, positioning accurate angle value factor is introduced, to guarantee the location data of network positions point It is believable.

The embodiment of the invention provides in dwell point judgement, collection point is the processing in a judgement of network positions type Improvement project in process, and accordingly in the process flow, however, it is determined that current collection point is dwell point, then can be according to being The dwell point is recorded in dwell point Maintenance Table by the operational requirements of system；Alternatively, will be deemed as determining for the collection point of dwell point Position information marks in map.If judging in step 203, current collection point, can be according to point of systemic presupposition for non-dwell point Analyse the period, wait next round analyze and determine process, if still encounter collection point be network positions type when, repeat step 201- step 203.

It is in a process flow of network positions type, in addition to can directly pass through step 201- step for collection point Outside 203 judgement to realize dwell point, the embodiment of the invention also provides a kind of expansion schemes, i.e., in step 201- step 203 There is provided the judgement processing of acceleration in a terminal before, the judgement treatment process of the acceleration can be with if smoothly The implementation procedure of step 201- step 203 directly is skipped, and show that current collection point is the judgement of dwell point, as shown in figure 4, institute Extended method is stated specifically to execute are as follows:

In step 301, the collection point in preset duration t1 is successively divided at least two sections according to acquisition time, to every One section of standard deviation for calculating its acceleration modulus value, arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard deviation sequence Column.

In step 302, the default quantile in the standard difference sequence is taken, if the default quantile is less than pre- bidding Quasi- differential bit threshold value then directly judges current collection point for dwell point；It is no to then follow the steps 201- step 203 to judge to stop Point.

Wherein, preset standard differential bit threshold value uses 0.057 in Android platform, uses 0.013 in iOS platform.It is above-mentioned Two parameter values are only the parameter value used when test of the embodiment of the present invention, and corresponding parametric values may be used also in the actual implementation process To do corresponding adjustment according to the walking acceleration of user.

Wherein, the treatment process of this step 301- step 302 be to be arranged in step 201- step 203 before It is performed, if step 301- step 302, which is judged as non-dwell point just, can execute step 201- step 203.The reason is that terminal When under resting state, it is also possible to the activity with significantly acceleration change is done, such as: user's carried terminal has been done indoors Oxygen movement, or be that user holds smart phone and plays game, it or is that user rewards in order to obtain and goes to complete some APP Given task of walking, completion means are exactly ceaselessly to get rid of smart phone etc. with hand indoors.For above-mentioned several feelings Condition, step 301- step 302 can all judge incorrectly, and need step 201- step 203 just at this time to make up.But if do not go out Existing above-mentioned several special circumstances, then step 301- step 302 just can obtain correct conclusion, to greatly simplify entire dwell point Judge process cycle, under the premise of guaranteeing result correctness, improves the computational efficiency of terminal.

For the application scenarios of the above-mentioned dwell point position that terminal is presented on map, the embodiment of the present invention also proposed one The preferred display strategy of kind takes the current acquisition point location coordinate adjacent most for being determined as one group of collection point of dwell point The position shape that there is close correspondence the place-centric of the object of interference GPS positioning signal to present on map for current collection point State.

The display mode avoids bounce of the dwell point in screen is shown, in addition, similar processing mode can also answer With on the record of dwell point, to simplify the data management of dwell point.

The realization of the embodiment of the present invention is illustrated from programming language angle, wherein using the analytical model of sliding window,

T_p={ p_j|p_i.t_i-p_j.t_j< Δ t, j ∈ [1, | T_p|]},|T_p| it is set sizes；Wherein, Δ t is that the present invention is real Preset duration t1 described in example is applied, | T_p| meaning be current collection point i before all acquisitions belonged in the Δ t time Point.Wherein, p_iAnd t_iBetween symbol " " be ownership symbol in data structure class, i.e. t_iFor structure class p_iIncluded in it is all One in more attributes.

Secondly, defining the point set of network align_type in window are as follows:

T_p.D={ p_i|p_j.type_j==1, p_j∈T_p, wherein p_j.type_j=1 is network positions, p_j.type_j=0 is GPS positioning.

Then, accuracy value change degree T is defined_p.c, whereinEssence Angle value change degree is the summation of the absolute value of the accuracy value difference in the time mouthful between adjacent two o'clock, is reacted on certain procedures The degree of accuracy value variation.

Finally, dwell point expression formula are as follows:

S_T={ T_i||T_i.D | > 0.8 | T_i|∩T_i.c 15 < }, i.e. T_iInterior align_type is that the accounting of network is greater than 80% (for the default network accounting threshold value in the embodiment of the present invention), and it (is of the invention real that accuracy value change degree, which is less than given threshold 15, Apply the default precision threshold in example), it can be judged as dwell point.

Based on the stops detection of Network positioning, rule of thumb rule is determined, core concept is according to interior Depending on positional parameter variation foundation environmental change, and without GPS signal (in most cases).In Zhao F, Ghorpade A, Pereira F C,et al.Stop detection in smartphone-based travel surveys[J] .Transportation it proposes to be in multiple base stations letter when building in Research Procedia, 2015,11:218-22 When the overlapping region of number tower, indoor dwell point causes to occur more due to the switching (hand-over) of this multiple base station signal tower A dwell point for deviateing building, which introduce the signatures match users of signal tower, and place often to be gone to solve the problems, such as this, thinks Think it is that when user is in these places, backstage algorithm will record the GSM signal towers that can be detected all at this time.These and some Often go the associated GSM signal tower signature (GSM signature) in place.When a new dwell point undetermined generates, algorithm Can check mobile phone at this time whether use GSM sign in some, if successful match, dwell point will be moved to associated Often go to place.This method can be effectively reduced the drift in dwell point.But there is also defects for the method, that is, if with The actual dwell point in family does not go to place often, can be also matched to by force and often go to place, to obtain the result of mistake.This implementation This method that is itd is proposed of example is that GPS is switched to Network and positions this principle when being entered the room using user, will enter room The last one the interior physical location of GPS available point as dwell point, efficiently solves the problems, such as this.

Embodiment 2:

In embodiment 1 from collection point be network positions type when, a corresponding process flow is illustrated.And the present invention is real Apply example then and be from collection point be GPS positioning type when, illustrate respective handling process.As shown in figure 5, the embodiment of the present invention proposes A kind of dwell point recognition methods based on parameter of doing more physical exercises, collection point are demarcated as GPS positioning type or network positions type After store, method includes:

In step 401, it when judging the align_type of current collection point for GPS positioning type, obtains in preset duration t2 The set being made of GPS positioning type collection point.

Wherein, preset duration t2 can be 1-5 minutes, be also possible to 10s or 30s, the setting of parameter mainly considers The cycle duration of the judgement of carry out one acquisition dotted state of the terminal periodic of the frequency acquisition and setting of collection point, and also Take into account the sum (too high to will affect terminal operation efficiency) of collection point in preset duration t1.In general control preset duration t1 Collection point quantity is 100 or so.Wherein, for the different periods, the embodiment of the invention provides two kinds of analytical models:

In step 402, pre-set radius distance is less than for collection point in the set to cluster centre distance, then adopted this Collection point belongs in corresponding sub-clustering.

Wherein, pre-determined distance usually can be 5-10M, can also be currently located architectural object in environment according to terminal Size dynamic generation, such as: for improve precision, pre-determined distance can be set to occupied area in neighboring buildings object Minimum one corresponding size.Wherein, the center of cluster generallys use current collection point.

In step 403, the GPS positioning type collection point number in the sub-clustering is counted relative to total in preset duration t2 GPS positioning type collection point accounting, and arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point sequence.

Wherein, the speed of collection point is usually to be calculated by GPS satellite and fed back to installation by the terminal of GPS module.For For optional scheme, the speed of above-mentioned collection point can also be calculated by terminal itself according to the location information got. Above two mode belongs to the implementation of the embodiment of the present invention.

In step 404, if the accounting result is greater than default GPS accounting threshold value, and in the GPS gathers point sequence Default quantile is less than default GPS quartile threshold value, then judges current collection point for dwell point.

Wherein, the default GPS accounting threshold value can be set as 0.8, appropriate can be increased according to the precision of judgement Or reduce the default GPS accounting threshold value；The default GPS quartile threshold value is carried out generally according to the minimum walking speed of user Setting.

The embodiment of the present invention proposes a kind of state judging method for being directed to network positions type collection point, it is suitable for Terminal is provided simultaneously under the conditions of GPS positioning and network positions, and the embodiment of the present invention is exactly analyze GPS positioning in the terminal excellent First grade is higher than network positions, i.e. network positions then can be abandoned or be neglected to terminal in the case where that can collect GPS positioning Collection point.Therefore, the embodiment of the present invention is exactly and exists analyzing existing GPS positioning data elegant problem may occur, because This, the default quantile increased newly in the comparison GPS gathers point sequence according to acquisition spot speed arrangement for GPS positioning data is big Small method, eliminates noise spot, ensure that the accuracy of last judging result.

It is in a process flow of network positions type, in addition to can directly pass through step 401- step for collection point Outside 404 judgement to realize dwell point, the embodiment of the invention also provides a kind of expansion schemes, i.e., in step 401- step 404 There is provided the judgement processing of acceleration in a terminal before, the judgement treatment process of the acceleration can be with if smoothly The implementation procedure of step 401- step 404 directly is skipped, and show that current collection point is the judgement of dwell point, as shown in fig. 6, institute Extended method is stated specifically to execute are as follows:

In step 501, the collection point in preset duration t2 is successively divided at least two sections according to acquisition time, to every One section of standard deviation for calculating its acceleration modulus value, arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard deviation sequence Column.

In step 502, the default quantile in the standard difference sequence is taken, if the default quantile is less than pre- bidding Quasi- differential bit threshold value then directly judges current collection point for dwell point；It is no to then follow the steps 402- step 404 to judge to stop Point.

Wherein, the treatment process of this step 501- step 502 be to be arranged in step 401- step 404 before It is performed, if step 501- step 502, which is judged as non-dwell point just, can execute step 401- step 404.The reason is that terminal When under resting state, it is also possible to the activity with significantly acceleration change is done, such as: user's carried terminal has been done indoors Oxygen movement, or be that user holds smart phone and plays game, it or is that user rewards in order to obtain and goes to complete some APP Given task of walking, completion means are exactly ceaselessly to get rid of smart phone etc. with hand indoors.For above-mentioned several feelings Condition, step 501- step 502 can all judge incorrectly, and need step 401- step 404 just at this time to make up.But if do not go out Existing above-mentioned several special circumstances, then step 501- step 502 just can obtain correct conclusion, to greatly simplify entire dwell point Judge process cycle, under the premise of guaranteeing result correctness, improves the computational efficiency of terminal.

In embodiments of the present invention, data structure designed in binding test test illustrates how to complete.Firstly, setting i Track for current point sequence, window indicates are as follows: T_p={ p_j|p_i.t_i-p_j.t_j< Δ t, j ∈ [1, Tp], Tp are set sizes； Here pj indicates the point before current point pi, in the Δ t time.Wherein, Δ t is preset duration described in the embodiment of the present invention t2。

Secondly, calculating T_pIt is interior with p_iFor the point set in the field Δ d of center:

T_p.C={ p_i|distance(p_i, p_j) < Δ d, p_j∈T_p, T_pIt .C is set sizes；Here distance (p_i, p_j) indicating the distance between two o'clock, Tp.C indicates that in set Tp be the center of circle using △ d as the collection of the point in the region of radius using pi It closes.

Definition set T_p.B={ p_i.v|v_i-1≤v_i≤v_i+1, i ∈ [1, | T_p|] it is T_pVelocity amplitude ascending order set, T_pIt .B is set sizes.

GPS dwell point expression formula isS_TTable Show stop point set, | Ti.C | > 0.9* | Ti | indicate that in Tp be the center of circle using △ d as the quantity of the point in the region of radius using pi Percentage is greater than 90% (the default GPS accounting threshold value i.e. in the embodiment of the present invention) of entire sample size；P_0.8*|Tp.B|.v table Show the 80th percentile (the default quantile i.e. in the embodiment of the present invention, and v of set Tp.B_minI.e. in the embodiment of the present invention Default GPS quartile threshold value).

The representative meaning of expression formula is, if meeting most of points (90%) in given radius, and big The speed of most points (80%) is less than minimum walking speed, that is, meets the judgement of resting state.

T_iIt is interior with p_iIt is greater than T for the point quantity in the field Δ d of center_i90 percent, and 80 percent point of speed Value is less than the minimum threshold v of walking_min, can be judged as dwell point.

Embodiment 3:

Providing a kind of dwell point recognition methods (network positions type based on parameter of doing more physical exercises described in embodiment 1 Side) and a kind of dwell point recognition methods (GPS positioning type side) based on parameter of doing more physical exercises as described in example 2 after, in this hair It also proposed a kind of dwell point recognition methods based on parameter of doing more physical exercises in bright embodiment, the embodiment of the present invention is by 1 He of embodiment Method in embodiment 2 is integrated into a complete model and is illustrated.In embodiments of the present invention, collection point can store In terminal (such as: the smart machines such as mobile phone, bracelet, plate), there may also be on server.Corresponding each collection point possesses respectively Align_type, the judgement of carry out one acquisition dotted state of terminal periodic, as described in Figure 7, the method in model includes:

In step 601, current collection point is obtained.

In embodiments of the present invention, current collection point is the anchor point collected in real time by terminal in real time.

In step 602, when determining the current collection point is network positions type, judge to obtain the current acquisition Terminal status before point.

In step 603, if being judged as motion state, collection point in preset duration t1 is obtained；Wherein, preset duration t1 Interior collection point includes the collection point of GPS positioning type and the collection point of network positions type；It is fixed to count network in the collection point The accounting rate and positioning accurate angle value change degree of position type collection point；If it is determined that the accounting rate of network positions type collection point is big In default network accounting threshold value, and positioning accurate angle value change degree is less than default precision threshold, then judges current collection point to stop Stationary point；If being judged as resting state, it is determined that the current acquisition point location coordinate is at a distance from previous dwell point, wherein away from When from being greater than pre-determined distance, then modifying state is motion state, and otherwise hold mode is resting state.And terminate sentencing for dwell point Disconnected process (as shown in Figure 7).

In step 603, can also be before executing step 202-203, first carry out step 301-302 passes through acceleration Judge whether it is the process of dwell point.

In step 604, when determining the current collection point is GPS positioning type, judge to obtain the current acquisition Terminal status before point.

In step 605, it if being judged as motion state, obtains in preset duration t2 by GPS positioning type collection point structure At set, for collection point in the set to cluster centre distance be less than pre-set radius distance, then the collection point is belonged to In corresponding sub-clustering；The GPS positioning type collection point number in the sub-clustering is counted relative to GPS positioning total in preset duration t2 The accounting of type collection point, and arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point Sequence；If the accounting result is greater than default GPS accounting threshold value, and the default quantile in the GPS gathers point sequence is less than Default GPS quartile threshold value, then judge current collection point for dwell point；If being judged as resting state, hold mode is to stop shape State.And terminate the deterministic process (as shown in Figure 7) of dwell point.

Wherein, in step 605, passing through for step 501-502 can also first be carried out before executing step 402-404 Acceleration judges whether it is the process of dwell point.

Respective processing means in summation of embodiment of the present invention embodiment 1 and embodiment 2, propose a kind of complex environment Under dwell point judgment method, propose a kind of effective solution.Compare various processing sides in the prior art Formula, the embodiment of the present invention have had both the beneficial effect of embodiment 1 and embodiment 2, and details are not described herein.

Corresponding embodiment 1, the embodiment of the present invention is before executing step 603, and there is also a kind of preferred spread steps, i.e., Dwell point judgement is carried out using the data that acceleration transducer detects, if the not successful calculating process for executing step 603 again. The elaboration of step 301- step 302, particular content are as follows in similar embodiment 1:

Corresponding embodiment 1, the embodiment of the present invention is before executing step 605, and there is also a kind of preferred spread steps, i.e., Dwell point judgement is carried out using the data that acceleration transducer detects, if the not successful calculating process for executing step 605 again. The elaboration of step 501- step 502, particular content are as follows in similar embodiment 1:

According to the acceleration that each collection point in preset duration t2 is configured, arrange from small to large it is that each collection point is configured plus Standard deviation of the speed in corresponding preset duration t2, obtains standard difference sequence；

The default quantile in the standard difference sequence is taken, if the default quantile is less than preset standard differential bit threshold Value, then directly judge current collection point for stationary point；Otherwise it executes and clustering operation and its subsequent content is carried out to the set to sentence Disconnected dwell point.

In of the invention embodiment, it can be configured for ease of calculation for each collection point by positioning with statistics, terminal The parameter that one or more forms in coordinate, timestamp, speed, positioning accuracy, acceleration and align_type.Such as embodiment 1 With the collection point p given in data structure specific in embodiment 2_iThe timestamp expression way p having_i.t_i。

It in embodiments of the present invention, simultaneously include positioning coordinate, timestamp, speed, positioning accuracy, acceleration with collection point For align_type, by taking the above-mentioned expansion scheme realized based on acceleration transducer detection as an example, wherein in preset duration t1 Collection point its be presented as that data structure is as follows:

Time window T_p={ p_i|p_i=(l_i, t_i, alt_i, v_i, Accu_i, a_i, s_i, type_i), i ∈ I }；Wherein, l_iIt represents a little Latitude and longitude coordinates (i.e. above-mentioned positioning coordinate), t_iTimestamp, alt_iRepresent height above sea level, v_iRepresent the speed of collection point, Accu_iGeneration The horizontal position precision of table positioning, a_iAcceleration modulus value, s_iState (including resting state and motion state), type_iTo position class Type, I are the points for forming track.By time window T_pIt is divided into n sections, every section of collection point sample size for being included is that m (is drawn here The purpose divided is mainly from the mass motion state for carrying out analysis window from angle of statistics, for example time window is 5 minutes, to each Minute is used as one section, totally 5 sections, individually carries out the calculating of acceleration standard deviation to each section, shows as if having more than 3 sections in 5 sections It is static, can determine that entire time window is static), then there is T_p={ s_i|s_i=(p_1+(i-1)*m, p_i*m), i ∈ [1, n] }.

Calculate the acceleration standard deviation of every orbit segment: Definition set T_p.A={ s_i.a_std|s_i-1.a_std≤s_i.a_std≤s_i+1.a_std, i ∈ [1, n] } and it is sorted n from small to large The standard deviation of section sample, calculates T_p.A third quartile q₃=s_0.75*n.a_std, then the expression formula of rest point is S_T={ T_i| T_i.A.q₃< 0.057 }.

It within the scope of entire time window, is segmented as unit of 1 minute, calculates each section the standard of its acceleration modulus value Difference, and it is compared with preset standard differential bit threshold value, Android platform uses 0.057, iOS platform 0.013 here, Once standard deviation is less than this preset standard differential bit threshold value, this section is judged as static segment, to all sections of progress in window Calculate, then static segment counted, judge if static segment number occupies the majority at this time as static dwell point, here we All sections of acceleration standard deviation is observed, with third quartile q3 compared with preset standard differential bit threshold value, if it is counted Value is less than preset standard differential bit threshold value, then the most of sections of the window can confidence for static.

Embodiment 4:

Illustrating that form describes how the embodiment of the present invention realizes improved stop by above method step and principle After point recognition methods, the embodiment of the present invention is then to have more detailed parameter attribute in conjunction with what is used in pseudocode and specific test Embodiment needs to do related data structures to model as described in Example 3 in embodiments of the present invention predefining, in the present invention Terminal is embodied in smart phone in embodiment, then first has to the foundation for carrying out mobile phone positioning track data model.Specific packet It includes:

The original location data of mobile phone is made of a geographical coordinate point and a timestamp plus other parameters, each A point is unique on space-time.Due to being according to use using network positions and using GPS module positioning under mixed positioning mode The surrounding enviroment condition at family and change, and the track data parameter under both of which has certain difference, in conjunction with the two similarities and differences And it is aided with acceleration information, it defines mobile phone and does more physical exercises supplemental characteristic model.

Define track, T_p={ p_i|p_i=(l_i, t_i, alt_i, v_i, Accu_i, a_i, s_i, type_i), i ∈ I }, wherein l_iIt represents The latitude and longitude coordinates (i.e. above-mentioned positioning coordinate) of point, t_iTimestamp, alt_iRepresent height above sea level, v_iRepresent the speed of collection point, Accu_i Represent the horizontal position precision of positioning, a_iAcceleration modulus value, s_iState (including resting state and motion state), type_iFor positioning Type, I are the points for forming track.

Secondly, to do one of screening for the data of acquisition, avoid obviously not being inconsistent logical acquisition point data mixed Enter the calculating and analytic process in embodiment 1-3.The process is also referred to as initial data cleaning.It specifically includes:

In initial data wash phase, for GPS positioning point, its error be mainly derived from intensive building area, Indoor, subterranean zone etc., the signal strength of GPS is very faint and has certain noise, even without GPS signal, simultaneously Angle between when satellite transit and it also will affect positioning accuracy.Rule of thumb, the anchor point of mistake can generate mistake simultaneously Altitude value, by comparing Ti.alt and local mean sea level value.If much deviateing, give up current point.According to current point and The distance between upper point calculates average speed V=Distance (l divided by the difference of timestamp_i,l_i-1If) much big In limit velocity v, also filter out.In addition, rule of thumb learn, the positioning accuracy of GPS or AGPS generally within 30m, for Point of the Accu much larger than 30 can also be filtered processing.Meanwhile in practical application, sometimes there is also precision less than 30 But the point of really drift, when such as mobile phone close to window, when being in indoor and outdoor critical zone.At this moment, pass through lesser acceleration value Carry out height and assertorically filter out the biggish point of these distance changes, such as sleep is had a meal when especially user is in resting state.

For network positions point, due to most of overall data be it is inaccurate, can only be filtered according to Accu, Accu > > 50, we can filter out.It is noted here that the filtering of the data for network positions point, the embodiment of the present invention are intended merely to use The beauty at family track interface does not allow it to show on map, there is no them are really deleted, in next dwell point Them can be used in algorithm detection.It then is exactly that dwell point detection is carried out using embodiment 1-3 the method, specific implementation is such as Under:

Sliding window mode is used in embodiments of the present invention, and the datum mark of window is current point, and the size of window is current point All the points to before away from current point in the △ t time, △ t are the time threshold that the dwell point that step 704 marks in Fig. 8 determines. Simultaneously global parameter is set to keep current state, wherein currentState be active user be in state (stop or Move), lastStopTime is the finish time of a upper dwell point；TraceList is the set tracelist=of all the points {T_i, 1≤i≤S, gpsList are the set tracelist={ T of GPS point_i|T_i} .type=0 1≤i≤S.StopList is Stop point set, the i.e. output of algorithm.

The algorithm flow of entire dwell point is as follows, and specific flow chart is as shown in Figure 8, comprising:

In step 701, cold start-up, parameter initialization, traceList is all point sets, and gpsList is GPS data Point set, state are state, and lastStopTime is the last end time stopped.

In a step 702, current point is put into traceList, judges current align_type, if it is GPS, execute step 703, step 706 is executed if it is network positions.

In step 703, current point is put into gpsList, judges current state state, if it is stop, judgement is current Whether point with last dwell point distance is greater than △ d, if so, current state by stop- > move, otherwise updates The laststoptime time；If current state is move, step 704 is executed.

In step 704, acceleration static detection is run to traceList, if it is determined that it is static, step 709 is executed, It is no to then follow the steps 705, i.e. correlation technique step in embodiment 2.

In step 705, to detection is stopped outside the gpsList operation room GPS, if it is determined that stopping, step 709 is executed, Otherwise step 702 is repeated.

In step 706, judge current state state, if it is stop, update the laststoptime time；If worked as Preceding state is move, executes step 707.

In step 707, acceleration static detection is run to traceList, if it is determined that it is static, step 709 is executed, It is no to then follow the steps 708, i.e. correlation technique step in embodiment 1.

In step 708, to detection is stopped in the traceList operation room Network, if it is determined that stopping, step is executed Rapid 709, otherwise repeatedly step 702.

In step 709, State updates the laststoptime time by move- > stop, and newly-increased dwell point arrives StopList repeats step 702.

Herein in conjunction with the hybrid positioning technology and acceleration sensor module under smart phone, a kind of real-time base is proposed Dwell point generation method under mobile phone does more physical exercises parameter, this method take full advantage of the advantage of GPS positioning and network positions, together When the unreliability of single location data is compensated in conjunction with acceleration transducer, the building of integrated use above-mentioned condition is outdoor and indoor Dwell point identification model, and threshold classification has been carried out to kinematic parameter feature when moving and when static, it is finally flat in Android This method is realized on platform, demonstrates its feasibility, is preferably solved existing for current trip survey dwell point recognition methods Problem.

Embodiment 5:

In order to verify the model and method of this paper, we have developed the masses based on Android platform to go out in research process Row data collection system.The interface software APP is as shown in figure 9, entire realize nearly hundred volunteers of process collection, total 280,000 rows Track data adds up to 361 hours, and Figure 10 is that the part of data shows schematic diagram.

Figure 10 is the data tested when user stops indoors, it can be found that when user is in indoor stop, GPS and network Positioning does not stop mutually to switch, and angle (bearing) is not GPS data for 0 in figure.It is this time true due to the unstability of positioning Real resting under Euclidean distance algorithm can be failed to judge.But pass through calculated result such as Figure 11-figure of acceleration static detection method 12, it can be seen that the detection of dwell point is successful.

The acceleration sampled value and windowed segments standard deviation of Figure 12 sample thus, horizontal axis are sample point sequence, sampling interval For 5s, the longitudinal axis is acceleration modulus value, it can be seen that static brief acceleration value is generally 0.1 hereinafter, from sample segmentation (every 10 Point is one section) after standard dygoram, all sections of standard deviation is respectively less than threshold value 0.057.Through the embodiment of the present invention in 1-3 Dwell point algorithm, we can be determined that the sample is resting state, it can be seen that acceleration parameter as reliable parameter, The unstability that the recognition result generated when GPS positioning drift can be made up, accurately identifies the true dwell point of user.

Simultaneously by comparison motion state brief acceleration data, there is significant characteristics differences for the two.Figure 13 is walking When acceleration sampled value, Figure 14 is the standard dygoram after sample segmentation, by comparison diagram 11, it can be found that when walking Acceleration range is in 1-5；By comparison diagram 12, it can be found that standard deviation generally fluctuates between 0.5 to 2.It is real through the invention The dwell point algorithm in a 1-3 is applied, we can be determined that the sample is motion state.

As can be seen that can effectively be known when user is in resting state by the acceleration static detection method of this paper It does not come out.

Dwell point unrecognized for acceleration rate threshold, our used location datas are to determine whether dwell point.Such as figure 15 for indoor stop when, GPS signal is blocked completely, and uses the accuracy value curve under network positions environment.Wherein 100 sample The accuracy value change degree of point sequence is respectively 12 and 4, respectively corresponds the left side Figure 15 and the right schematic diagram；Compared to moving condition, stop Surrounding environment change is smaller when staying, and the precision of positioning differs very little.When mobile, since ambient enviroment constantly changes, the essence of positioning Degree also changes therewith.Precision curve graph when Figure 16 is mobile, wherein the accuracy value change degree of 100 sample points is respectively 565 Hes 188, respectively correspond the left side Figure 16 and the right schematic diagram (the accuracy value change degree refer to the cumulative of the difference of adjacent two o'clock and), this In accuracy value change degree refer to adjacent two o'clock difference cumulative and.It can be seen that accuracy value change degree when stop and change when movement Change degree has biggish difference.

Therefore reasonable accuracy value change degree threshold value can preferably identify stop and movement, by 58 volunteers Data analysis in 361 hours, 5 minutes accuracy value change degree average out to 15 when indoor stop can reach preferable threshold value.

Prototype system provides five basic functions altogether, the track data including recording user in real time, track it is smooth Filtering processing, the case history record that dwell point is shown in the form of log of living, individual subscriber center.

Embodiment 6:

The embodiment of the invention also provides a kind of dwell point identification device based on parameter of doing more physical exercises, the device is for executing Embodiment 3 and method as described in example 4, described device include that obtain module, align_type judgment module, network fixed for collection point Position type processing module, GPS positioning type processing module, wherein collection point obtains the module connection align_type and judges mould Block, the align_type judgment module are separately connected the network positions type processing module and GPS positioning type processing module, It is specific:

The collection point obtains module, for obtaining current collection point；

It is worth noting that the contents such as information exchange, implementation procedure between module, unit in above-mentioned apparatus, due to It is based on same design with processing method embodiment of the invention, for details, please refer to the description in the embodiment of the method for the present invention, Details are not described herein again.

Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of embodiment is can to lead to Program is crossed to instruct relevant hardware and complete, which can be stored in a computer readable storage medium, storage medium It may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), disk or CD etc..

The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.

Claims

1. a kind of dwell point recognition methods based on parameter of doing more physical exercises, which is characterized in that it is respective fixed that corresponding each collection point possesses Position type, the judgement of carry out one acquisition dotted state of terminal periodic, which comprises

Obtain current collection point；

When determining the current collection point is network positions type, judge to obtain shape locating for terminal before the current collection point State；

If being judged as motion state, collection point in preset duration t1 is obtained；Wherein, collection point includes GPS in preset duration t1 The collection point of align_type and the collection point of network positions type；Count accounting for for network positions type collection point in the collection point Ratio and positioning accurate angle value change degree；If it is determined that the accounting rate of network positions type collection point is greater than default network accounting threshold Value, and positioning accurate angle value change degree is less than default precision threshold, then judges current collection point for dwell point；If being judged as stop State, it is determined that the current acquisition point location coordinate is at a distance from previous dwell point, wherein when distance is greater than pre-determined distance, Then modifying state is motion state, and otherwise hold mode is resting state；

When determining the current collection point is GPS positioning type, judge to obtain shape locating for terminal before the current collection point State；

If being judged as motion state, the set being made of in preset duration t2 GPS positioning type collection point is obtained, for described Collection point is less than pre-set radius distance to cluster centre distance in set, then belongs to the collection point in corresponding sub-clustering；Statistics institute Accounting of the GPS positioning type collection point number relative to GPS positioning type collection point total in preset duration t2 in sub-clustering is stated, And arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point sequence；If the accounting knot Fruit is greater than default GPS accounting threshold value, and the default quantile in the GPS gathers point sequence is less than default GPS quartile threshold value, then Judge current collection point for dwell point；If being judged as resting state, hold mode is resting state.

2. dwell point recognition methods according to claim 1, which is characterized in that in executing the statistics collection point Before accounting rate and positioning accurate angle value the change degree operation of network positions type collection point, the method also includes:

Collection point in preset duration t1 is successively divided at least two sections according to acquisition time, calculates its acceleration to each section The standard deviation of modulus value arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard difference sequence；

The default quantile in the standard difference sequence is taken, if the default quantile is less than preset standard differential bit threshold value, Directly judge current collection point for dwell point；Otherwise accounting for for network positions type collection point in the statistics collection point is executed The operation of ratio and positioning accurate angle value change degree and its subsequent content judge dwell point.

3. dwell point recognition methods according to claim 1, which is characterized in that carry out sub-clustering behaviour to the set executing Before work, the method also includes:

Collection point in preset duration t2 is successively divided at least two sections according to acquisition time, calculates its acceleration to each section The standard deviation of modulus value arranges the standard deviation of each section of acceleration modulus value from small to large, obtains standard difference sequence；

The default quantile in the standard difference sequence is taken, if the default quantile is less than preset standard differential bit threshold value, Directly judge current collection point for dwell point；Otherwise it executes and clustering operation and its subsequent content is carried out to the set to judge to stop Stationary point.

4. dwell point recognition methods according to claim 1 to 3, which is characterized in that each collection point configuration is by positioning The parameter that one or more forms in coordinate, timestamp, speed, positioning accuracy, acceleration and align_type.

5. a kind of dwell point recognition methods based on parameter of doing more physical exercises, which is characterized in that collection point is demarcated as GPS positioning type Or stored after network positions type, method includes:

When judging the align_type of current collection point for network positions type, collection point in preset duration t1 is obtained；Wherein, it presets Collection point includes the collection point of GPS positioning type and/or the collection point of network positions type in duration t1；

Count the accounting rate of network positions type collection point and positioning accurate angle value change degree in the collection point；If it is determined that the net The accounting rate of network align_type collection point is greater than default network accounting threshold value, and positioning accurate angle value change degree is less than default precision Threshold value then judges current collection point for dwell point.

6. dwell point recognition methods according to claim 5, which is characterized in that in executing the statistics collection point Before accounting rate and positioning accurate angle value the change degree operation of network positions type collection point, the method also includes:

According to the acceleration that each collection point in preset duration t1 is configured, the acceleration that each collection point is configured is arranged from small to large The standard deviation in corresponding preset duration t1 is spent, standard difference sequence is obtained；

7. dwell point recognition methods according to claim 5, which is characterized in that for being determined as one group of acquisition of dwell point Point takes the adjacent nearest correspondence of the current acquisition point location coordinate to have the place-centric of object of interference GPS positioning signal For the position of current collection point presented on map.

8. a kind of dwell point recognition methods based on parameter of doing more physical exercises, which is characterized in that collection point is demarcated as GPS positioning type Or stored after network positions type, method includes:

When judging the align_type of current collection point for GPS positioning type, obtains and acquired in preset duration t2 by GPS positioning type The set that point is constituted carries out sub-clustering to the set；Wherein, clustering process includes:

Pre-set radius distance is less than for collection point in the set to cluster centre distance, then the collection point is belonged to corresponding point In cluster；

The GPS positioning type collection point number counted in the sub-clustering is adopted relative to GPS positioning type total in preset duration t2 Collect the accounting of point, and arrange each collection point in the preset duration t2 from small to large according to speed, obtains GPS gathers point sequence；

If the accounting result is greater than default GPS accounting threshold value, and the default quantile in the GPS gathers point sequence is less than in advance If GPS quartile threshold value then judges current collection point for dwell point.

9. dwell point recognition methods according to claim 8, which is characterized in that carry out sub-clustering behaviour to the set executing Before work, the method also includes:

10. a kind of dwell point identification device based on parameter of doing more physical exercises, which is characterized in that described device includes that collection point obtains mould Block, align_type judgment module, network positions type processing module, GPS positioning type processing module, wherein collection point obtains Module connects the align_type judgment module, and the align_type judgment module is separately connected the network positions type processing Module and GPS positioning type processing module, specific:

The collection point obtains module, for obtaining current collection point；

The align_type judgment module obtains institute for judging when determining the current collection point is network positions type State terminal status before current collection point；

The network positions type processing module is used for, and when being judged as motion state, obtains collection point in preset duration t1；Its In, collection point includes the collection point of GPS positioning type and the collection point of network positions type in preset duration t1；It is adopted described in statistics The accounting rate and positioning accurate angle value change degree of network positions type collection point in collection point；If it is determined that the network positions type acquisition The accounting rate of point is greater than default network accounting threshold value, and positioning accurate angle value change degree is less than default precision threshold, then judgement is worked as Preceding collection point is dwell point；If being judged as resting state, it is determined that the current acquisition point location coordinate and previous dwell point Distance, wherein when distance is greater than pre-determined distance, then modifying state is motion state, and otherwise hold mode is resting state；

The align_type judgment module is also used to when determining the current collection point is GPS positioning type, judges to obtain institute State terminal status before current collection point；

GPS positioning type processing module, for obtaining in preset duration t2 by GPS positioning type when being judged as motion state The set that collection point is constituted is less than pre-set radius distance for collection point in the set to cluster centre distance, then by the acquisition Point belongs in corresponding sub-clustering；The GPS positioning type collection point number in the sub-clustering is counted relative to total in preset duration t2 GPS positioning type collection point accounting, and arrange each collection point in the preset duration t2 from small to large according to speed, obtain GPS gathers point sequence；If the accounting result is greater than default GPS accounting threshold value, and default point in the GPS gathers point sequence Digit is less than default GPS quartile threshold value, then judges current collection point for dwell point；If being judged as resting state, hold mode For resting state.