CN114257988B - Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles - Google Patents
Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles Download PDFInfo
- Publication number
- CN114257988B CN114257988B CN202111384794.2A CN202111384794A CN114257988B CN 114257988 B CN114257988 B CN 114257988B CN 202111384794 A CN202111384794 A CN 202111384794A CN 114257988 B CN114257988 B CN 114257988B
- Authority
- CN
- China
- Prior art keywords
- dnn
- vehicles
- link
- resource allocation
- expression
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 73
- 238000013468 resource allocation Methods 0.000 title claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 28
- 238000001228 spectrum Methods 0.000 claims abstract description 25
- 238000004364 calculation method Methods 0.000 claims abstract description 22
- 238000009826 distribution Methods 0.000 claims abstract description 9
- 230000014509 gene expression Effects 0.000 claims description 32
- 230000009977 dual effect Effects 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 16
- 230000006855 networking Effects 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 11
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 9
- 230000010354 integration Effects 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 238000013144 data compression Methods 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 2
- 238000011426 transformation method Methods 0.000 claims description 2
- 238000012512 characterization method Methods 0.000 claims 1
- 238000012795 verification Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000013374 right angle light scattering Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000013500 data storage Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/46—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for vehicle-to-vehicle communication [V2V]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/04—Wireless resource allocation
- H04W72/044—Wireless resource allocation based on the type of the allocated resource
- H04W72/0453—Resources in frequency domain, e.g. a carrier in FDMA
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
- H04W72/542—Allocation or scheduling criteria for wireless resources based on quality criteria using measured or perceived quality
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention belongs to the technical field of Internet of vehicles, and discloses a resource allocation method and a system for imperfect CSI in an ultra-reliable Internet of vehicles, wherein the resource allocation method for imperfect CSI in the ultra-reliable Internet of vehicles comprises the following steps: correspondingly heterogeneous service requirements by utilizing V2I and V2V links; converting the V2V reliability constraint and the V2I throughput into a mode capable of being calculated by a robust method; training a parallel DNN architecture by using a feedback control-based method to obtain power control and local calculation allocation decisions; and (5) pushing out an optimal spectrum allocation decision based on the Hungary algorithm. The invention can be used for solving the problems of vehicle transmitting power and local calculation proportion distribution in the high-speed moving vehicle environment. The method is suitable for a high-speed mobile network with imperfect CSI, and overcomes the difficulty brought by the fact that the accurate channel state cannot be acquired; and resource allocation decision is carried out based on the trained DNN, so that the real-time requirement of the Internet of vehicles is met compared with the traditional iterative algorithm.
Description
Technical Field
The invention belongs to the technical field of Internet of vehicles, and particularly relates to a resource allocation method and system for imperfect CSI in ultra-reliable Internet of vehicles.
Background
At present, an intelligent traffic system has great potential in improving traffic efficiency, road safety, providing rich driving entertainment and the like. Vehicle-to-everything (V2X) communication is one of the indispensable technologies in intelligent transportation systems, which enables automatic driving through vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) links. In particular, non-security related services, such as entertainment services and high definition map information, involve real-time data interactions, requiring access to the internet or remote servers via high capacity V2I connections. In contrast, reliability is a key indicator of safety critical applications, such as sharing Collaboration Awareness Messages (CAM) and distributed environment notification messages (denom) between surrounding vehicles over V2V links. In addition, V2I wireless links are subject to significant stress when large amounts of data need to be uploaded due to the limited V2I communication resources. Fortunately, as vehicle technology evolves, vehicles are equipped with certain computing devices. Thus, the vehicle local computing unit may compress portions of the data, which further reduces latency of V2I communications. Thus, one reasonable resource allocation scheme is to jointly optimize radio resources and locally calculated proportions of the vehicle.
Unlike conventional communication networks, allocation of internet of vehicles resources presents special challenges. First, the high mobility characteristics of the vehicle cause rapid changes in the wireless Channel State (CSI). Thus, it is impractical to track channel variations over such a short time scale. Secondly, in order to achieve high spectral efficiency, V2V and V2V links multiplex the spectrum, thereby introducing a complex interference problem. On the other hand, the resource allocation scheme in the V2X network needs to have low complexity and low latency characteristics. Thus, conventional iterative algorithms with high complexity are no longer suitable, considering that iterations require a non-negligible delay.
Through the above analysis, the problems and defects existing in the prior art are as follows:
(1) The high mobility characteristics of existing vehicles result in rapid changes in the wireless Channel State (CSI) and are therefore impractical to track channel changes on such short time scales.
(2) In order to achieve high spectrum efficiency, V2V and V2I links multiplex spectrum, thereby introducing complex interference problems, the resource allocation scheme in the V2X network needs to have low complexity and low latency characteristics.
(3) In the allocation of internet of vehicles resources, the conventional iterative algorithm with high complexity is no longer applicable because of the non-negligible delay required for the iteration.
In summary, the existing technical problems are: most of the existing internet of vehicles resource allocation methods are based on accurate network state information, and cannot guarantee the high-reliability transmission requirement of high-speed moving vehicles.
The difficulty of solving the technical problems is as follows: the technical problem is solved by establishing a network model capable of describing heterogeneous service requirements of the high-speed mobile Internet of vehicles and processing uncertainty of network information states. In addition, V2V and V2I links are multiplexed, resulting in complex interference problems, and conventional iteration-based algorithms cannot be used for low latency demand networking. Therefore, how to establish a communication mode of internet of vehicles that can characterize heterogeneous service requirements and propose a resource allocation algorithm with low computational complexity becomes the biggest difficulty in solving the above technical problems.
The meaning of solving the problems and the defects is as follows: the method solves the rapid problem, can overcome the difficulty that the channel state cannot be accurately acquired due to the high-speed movement of the vehicle, and designs an algorithm with low computational complexity, thereby better completing the resource allocation of the Internet of vehicles.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a resource allocation method and a system for imperfect CSI in ultra-reliable Internet of vehicles.
The invention is realized in such a way that the resource allocation method facing to imperfect CSI in the ultra-reliable Internet of vehicles comprises the following steps:
Firstly, utilizing the corresponding heterogeneous service requirements of V2I and V2V links; then, converting the V2V reliability constraint and the V2I throughput into a mode capable of being calculated by a robust method; secondly, training a parallel DNN architecture by using a feedback control-based method to obtain power control and local calculation allocation decisions; and finally, the optimal spectrum allocation decision is deduced based on the Hungary algorithm.
Further, the imperfect CSI-oriented resource allocation method in the ultra-reliable Internet of vehicles comprises the following steps:
firstly, modeling, namely constructing a vehicle networking model of diversified services; establishing a model suitable for heterogeneous services of the Internet of vehicles: V2I (Low latency) V2V (high reliability)
Step two, converting the V2V reliability constraint into a computable form by using a feasible domain conversion method, and obtaining a computable expression of the V2I throughput by using a robust SINR correction method; the method is used for processing the situation that the channel state caused by high-speed movement in the Internet of vehicles cannot be obtained.
Designing a parallel DNN network structure and constructing a loss function of DNN; a DNN structure based on learning is proposed that is suitable for the project.
Training DNN parameters; and training DNN network parameters according to the proposed feedback control method.
Step five, DNN reasoning, namely acquiring the transmitting power of V2I and V2V according to the trained DNN And locally calculating a ratio lambda m; based on the trained DNN, DNN reasoning is performed.
And step six, obtaining an optimal spectrum matching decision x m,k according to the Hungary algorithm. The difficulty brought by integer programming is solved.
Further, in the first step, the building the internet of vehicles model of the diversified services includes:
(1) Description of imperfect channels;
(2) Characterizing V2V reliability constraints;
(3) And establishing a V2I time delay expression.
Further, in the second step, the method of using feasible domain transformation transforms the V2V reliability constraint into a calculable form, and obtains a calculable expression of the V2I throughput by using a robust SINR correction method, including:
(1) The V2V reliability constraint is converted into a computable form by an integral-based feasible domain conversion method;
(2) The V2I expression is converted into a pattern that can be calculated using a robust SINR correction method.
Further, in the third step, the designing a parallel DNN network structure, and constructing a loss function of DNN, including:
(1) Taking into account the V2I, V V transmit power The local calculation proportion lambda m of the vehicle is adopted to design a parallel DNN framework capable of simultaneously deciding a plurality of variables;
(2) According to the Lagrangian theorem, a dual variable is introduced, a Lagrangian expression is derived and used as a loss function of the training DNN.
Further, in the fourth step, the training DNN parameters include:
(1) Initializing DNN parameters and dual variables;
(2) And updating the dual variables and the neural network parameters by adopting a small-batch gradient method.
Further, in step six, the obtaining, according to the hungarian algorithm, the optimal spectrum matching decision x m,k includes:
(1) Obtaining the most probable solution for each possible V2I-V2V pair;
(2) And sequencing each possible solution to push out an optimal spectrum allocation decision.
Another object of the present invention is to provide a system for allocating imperfect CSI-oriented resources in an ultra-reliable internet of vehicles, to which the method for allocating imperfect CSI-oriented resources in an ultra-reliable internet of vehicles is applied, the system for allocating imperfect CSI-oriented resources in an ultra-reliable internet of vehicles comprising:
the vehicle networking model building module is used for building a vehicle networking model of diversified services;
The calculable expression acquisition module is used for converting the V2V reliability constraint into a calculable form by using a feasible domain conversion method and obtaining a calculable expression of the V2I throughput by using a robust SINR correction method;
The DNN network structure construction module is used for designing a parallel DNN network structure and constructing a loss function of DNN;
The DNN parameter training module is used for training DNN parameters;
A DNN reasoning module for acquiring the transmitting power of V2I and V2V according to the trained DNN And locally calculating a ratio lambda m;
the spectrum matching decision acquisition module is used for acquiring an optimal spectrum matching decision x m,k according to the Hungary algorithm.
It is a further object of the present invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:
Firstly, utilizing the corresponding heterogeneous service requirements of V2I and V2V links; then, converting the V2V reliability constraint and the V2I throughput into a mode capable of being calculated by a robust method; secondly, training a parallel DNN architecture by using a feedback control-based method to obtain power control and local calculation allocation decisions; and finally, the optimal spectrum allocation decision is deduced based on the Hungary algorithm.
The invention further aims to provide an information data processing terminal which is used for realizing the imperfect CSI-oriented resource distribution system in the ultra-reliable Internet of vehicles.
By combining all the technical schemes, the invention has the advantages and positive effects that: the invention provides a resource allocation method for imperfect CSI in ultra-reliable Internet of vehicles, in particular relates to an imperfect CSI system communication technology, and is a resource allocation method based on learning. The invention can be used for solving the problems of vehicle transmitting power and local calculation proportion distribution in the high-speed moving vehicle environment. The method is suitable for the high-speed mobile network with imperfect CSI, and overcomes the difficulty brought by the fact that the accurate channel state cannot be acquired; and resource allocation decision is carried out based on the trained DNN, so that the real-time requirement of the Internet of vehicles is met compared with the traditional iterative algorithm.
Conventional internet of vehicles resource allocation algorithms assume that the BS can obtain accurate CSI, which is impractical in a high-speed moving internet of vehicles scenario. The invention provides an imperfect CSI-based internet of vehicles resource allocation algorithm, which converts uncertain constraints into computable expressions through an integration and robust method. In addition, a resource allocation algorithm with low computational complexity based on learning is provided for solving the interference problem with high computational complexity.
The basic idea of the invention is to minimize the time delay of the V2I link on the premise of ensuring the reliability of the V2V link, and simultaneously, the vehicle calculates part perception data; firstly, establishing a car networking model corresponding to heterogeneous service, and secondly, overcoming the difficulty in expressing V2V reliability and V2I throughput caused by imperfect CSI by using a robust method; then designing a parallel DNN architecture for simultaneously outputting a plurality of resource allocation decisions; then, an unsupervised DNN training method is designed, optimal DNN parameters are obtained, V2I, V V transmitting power is obtained, and a vehicle local calculation proportion distribution decision is made; and finally, based on the Hungary algorithm, obtaining an optimal spectrum allocation decision.
Compared with the prior art, the invention has the following advantages:
Firstly, the invention overcomes the difficulty of V2V reliability and V2I throughput closed expression caused by the fact that perfect CSI cannot be accurately obtained in the prior art through a feasible domain conversion method and a robust SINR correction method. In addition, the invention only depends on the CSI of a large time scale, so that the invention has the advantage of small network overhead.
Secondly, the invention designs a parallel DNN architecture, one is a transmission power control unit, and the other is a local calculation proportion distribution unit. The invention has the advantage of deciding a plurality of resource allocation decisions at one time.
Thirdly, the invention ensures the reliability of the DNN network through an unsupervised DNN training method based on feedback control. The invention only needs to use the trained DNN reasoning to obtain the optimal decision, and compared with the traditional iterative algorithm, the invention has the characteristic of low time delay, and meets the real-time requirement of the Internet of vehicles.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a method for allocating resources for imperfect CSI in an ultra-reliable internet of vehicles according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a resource allocation method for imperfect CSI in an ultra-reliable internet of vehicles according to an embodiment of the present invention.
FIG. 3 is a block diagram of a resource allocation system for imperfect CSI in an ultra-reliable Internet of vehicles according to an embodiment of the present invention;
in the figure: 1. the vehicle networking model building module; 2. a computable expression acquisition module; 3. a DNN network structure construction module; 4. a DNN parameter training module; 5. a DNN reasoning module; 6. and the frequency spectrum matching decision acquisition module.
Fig. 4 is a diagram of an internet of vehicles system provided by an embodiment of the present invention.
Fig. 5 is a parallel DNN architecture diagram provided by an embodiment of the present invention.
Fig. 6 is a schematic diagram of average data upload delay when the maximum V2V transmit power is increased according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Aiming at the problems in the prior art, the invention provides a resource allocation method and a system for imperfect channel information state (CSI) in ultra-reliable Internet of vehicles, and the invention is described in detail below with reference to the accompanying drawings.
For the existing resource allocation algorithm, most of the resource allocation algorithm is based on accurate CSI, and the resource allocation decision cannot be guaranteed to meet the network quality of service (QoS) requirement. The invention overcomes the difficulty that imperfect CSI can not display expression based on an integral and robust method from the aspect that the actual large time scale is slowly changed and the small time scale is not available. And a parallel DNN algorithm based on feedback control is provided, so that the calculation complexity of the algorithm is effectively reduced, and the method is suitable for a high-speed mobile Internet of vehicles environment.
As shown in fig. 1, the method for allocating resources for imperfect CSI in the ultra-reliable internet of vehicles according to the embodiment of the present invention includes the following steps:
S101, modeling, namely constructing a vehicle networking model of diversified services;
S102, converting the V2V reliability constraint into a computable form by using a feasible domain conversion method, and obtaining a computable expression of the V2I throughput by using a robust SINR correction method;
S103, designing a parallel Deep Neural Network (DNN) network structure, and constructing a loss function of DNN;
S104, training DNN parameters;
S105, DNN reasoning, namely acquiring the transmitting power of V2I and V2V and the local calculation proportion according to the trained DNN;
And S106, acquiring an optimal spectrum matching decision according to the Hungary algorithm.
The schematic diagram of the resource allocation method for imperfect CSI in the ultra-reliable Internet of vehicles provided by the embodiment of the invention is shown in fig. 2.
As shown in fig. 3, the resource allocation system for imperfect CSI in the ultra-reliable internet of vehicles provided by the embodiment of the present invention includes:
The vehicle networking model building module 1 is used for building a vehicle networking model of diversified services;
the calculable expression acquisition module 2 is used for converting the V2V reliability constraint into a calculable form by using a feasible domain conversion method and obtaining a calculable expression of the V2I throughput by using a robust SINR correction method;
the DNN network structure construction module 3 is used for designing a parallel DNN network structure and constructing a loss function of DNN;
the DNN parameter training module 4 is used for training DNN parameters;
a DNN reasoning module 5 for obtaining the transmission power of V2I and V2V according to the trained DNN And locally calculating a ratio lambda m;
The spectrum matching decision obtaining module 6 is configured to obtain an optimal spectrum matching decision x m,k according to the hungarian algorithm.
The Internet of vehicles system diagram provided by the embodiment of the invention is shown in fig. 4.
The technical scheme of the invention is further described below with reference to specific embodiments.
Step one, constructing a vehicle networking model of diversified services, which comprises the following steps:
The invention assumes that one V2V multiplexes only one V2I link, and that one V2I link can only be shared with one V2V link. Therefore, the power and the local calculation proportion of each possible V2I-V2V pair can be solved, and then the optimal spectrum allocation decision can be obtained according to the Hungary algorithm.
(1) Description of imperfect channel.
The CSI defining an m-thV2I link may be denoted as h m,B＝gm,Bαm,B, where g m,B denotes the small time scale part of the CSI and a m,B is the large time scale parameter. Because a fast moving vehicle cannot acquire accurate CSI, we assume that the small time scale parameter g m,B satisfies the exponential distribution of unit mean. Meanwhile, the channel state of the V2V link, the interference of the m-thV2I link to the k-thV2V link, and the interference of the k-th V2V link to the m-thV2I link are defined identically and are respectively expressed as h k,hm,k,hk,B.
(2) The V2V reliability constraint is characterized.
For safety-related traffic, the V2V reliability constraint is expressed in the form of a probability,Wherein the method comprises the steps of Representing the probability of a communication outage, 1-p 0 represents the reliability requirement.
(3) And establishing a V2I time delay expression.
The SINR of an m-thV2I user can be expressed asTherefore, the V2I capacity can be expressed asFor non-safety related business, partial original data can be locally compressed to reduce small data volume, thereby reducing transmission burden and time delay. D m represents the size of the data, λ m represents the local calculation ratio, f l represents the vehicle local CPU calculation frequency, and β represents the local calculated original data compression ratio. The local computation delay may be expressed asThe delay of directly transmitting the original data is/>The uploading time delay after the local data compression isThe delay of the m-thV2I link is T local+Tcompressed when T local＞Traw, otherwise T raw+Tcompressed. Through the above analysis, the delay of the m-thV2I link can be expressed as
In addition, the V2I and V2V links have maximum transmit power constraints, denoted as
2. And converting the V2V reliability constraint into a computable form by using a feasible domain conversion method, and obtaining a computable expression of the V2I throughput by using a robust SINR correction method.
(1) The integration-based feasible-area transformation method transforms the V2V reliability constraint into a computable form.
Since g k and g m,k satisfy independent unit index distributions, the reliability constraint can be translated by integration, namely:
(2) The V2I expression is converted into a pattern that can be calculated using a robust SINR correction method.
Unlike the V2V reliability constraint, the log function in the V2I expression makes integration impossible to obtain a closed form. The present invention proposes a robust method.
The introduction of the auxiliary variable γ c, the V2I expression can be obtained by solving the following problem:
Rm＝max E[log2(1+γc)]
Wherein, Epsilon represents the outage probability allowed for the V2I link. The computable expression for deriving V2I is:
3. Designing a parallel DNN network structure, and constructing a loss function of DNN, wherein the method comprises the following steps of:
(1) Taking into account the V2I, V V transmit power And the local calculation proportion lambda m of the vehicle is used for designing a parallel DNN framework capable of simultaneously deciding a plurality of variables.
The parallel DNN architecture is shown in FIG. 5 and includes a power decision unit for outputtingA local calculation scaling decision unit is used to calculate lambda m.
(2) According to Lagrange's theorem, a dual variable is introduced, and a Lagrange expression is deduced. And takes this as a loss function of the training DNN.
Solving the dual problem of the original problem (Lagrangian function) to obtain an iterative expression of the dual variables (multipliers). And feeds back the multiplier to the DNN network parameters (original variables).
4. Training DNN parameters comprises the steps of:
(1) The DNN parameters, dual variables, are initialized.
The number of samples was 100000, where the training samples were 70000 and the validation samples were 30000.
The DNN parameter is initialized with Xavier and the dual variable initial value is set to 0.
(2) And updating the dual variables and the neural network parameters by adopting a small-batch gradient method.
The training procedure for DNN was performed using Python 3.6 with Tensorflow.1.7.1. The learning rate was set to 0.001, and 4 hidden layers were assumed for DNN, each containing 200 units, with a mini-band size of 100. The parameters are updated by adopting the training method of Adam. Meanwhile, due to feedback of dual variables, the loss function is dynamically updated.
DNN reasoning, obtaining the transmitting power of V2I and V2V according to the trained DNNAnd locally calculating the ratio lambda m.
6. According to the hungarian algorithm, an optimal spectrum matching decision x m,k is obtained.
(1) The most favorable solution for each possible V2I-V2V pair is obtained.
(2) And sequencing each possible solution to push out an optimal spectrum allocation decision.
When the transmission power of any V2I-V2V pairAnd the local calculation ratio lambda m is known. The problem becomes a bipartite graph problem which can be efficiently solved by using the hungarian algorithm.
To illustrate the effectiveness of the proposed learning-based resource allocation algorithm, the present invention presents the following comparison algorithm.
A large time scale based resource allocation algorithm (Resource allocation under large-SCALE CHANNEL (RALS)) in which it is assumed that fast changing small time scale channel states are ignored, i.e. the channel state depends only on slow changing large time scale channel information states.
The resource allocation algorithm (Resource allocation for VEHICLE WITH no computing unit) for which the vehicle is not equipped with a computing unit assumes that the vehicle is not equipped with a computing unit, i.e. the original perceived data is transferred between vehicles to the base station.
Fig. 6 shows the average data upload delay at the maximum V2V transmit power increase. It was observed that the average delay for all schemes is very large when the V2V maximum transmit power is relatively small. This is because V2V communications can accept less interference from V2I links. As the maximum V2V transmit power increases, the average delay slowly decreases as V2V reliability is met, primarily affecting the V2I throughput dependent performance. Furthermore, it was found that the algorithm performs better in terms of delay than RALS and RANC at different maximum V2V transmit powers.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When used in whole or in part, is implemented in the form of a computer program product comprising one or more computer instructions. When loaded or executed on a computer, produces a flow or function in accordance with embodiments of the present invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk Solid STATE DISK (SSD)), etc.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (4)
1. The imperfect CSI-oriented resource allocation method in the ultra-reliable Internet of vehicles is characterized by comprising the following steps of: firstly, utilizing the corresponding heterogeneous service requirements of V2I and V2V links; then, converting the V2V reliability constraint and the V2I throughput into a mode capable of being calculated by a robust method; secondly, training a parallel DNN architecture by using a feedback control-based method to obtain power control and local calculation allocation decisions; finally, the optimal spectrum allocation decision is deduced based on the Hungary algorithm;
Assuming that one V2V only multiplexes one V2I link, meanwhile, one V2I link can only be shared with one V2V link, firstly solving the power and the local calculation proportion of each possible V2I-V2V pair, and then obtaining the optimal spectrum allocation decision according to the Hungary algorithm;
(1) Description of imperfect channel:
The CSI defining the m-thV2I link is denoted as h m,B＝gm,Bαm,B, where g m,B denotes the small time scale part in the CSI, a m,B is the large time scale parameter; assuming that the small time scale parameter g m,B meets the exponential distribution of the unit mean value, the channel state of the V2V link, the interference of the m-thV2I link to the k-thV2V link, and the interference of the k-thV2V link to the m-thV2I link are defined identically and are respectively expressed as h k,hm,k,hk,B;
(2) Characterization of V2V reliability constraints:
for safety-related traffic, the V2V reliability constraint is expressed in the form of a probability, Wherein the method comprises the steps of Representing the probability of communication disruption, 1-p 0 representing the reliability requirement;
(3) Establishing a V2I time delay expression:
SINR of the m-thV2I user is expressed as V2I Capacity is expressed as/>For non-safety related business, partial original data is locally compressed to reduce small data volume, thereby reducing transmission load and time delay; d m represents the size of data, λ m represents the local calculation ratio, f l is the vehicle local CPU calculation frequency, and β represents the local calculated original data compression ratio; the local computation latency is denoted/>The time delay of directly transmitting the original data isThe uploading time delay after local data compression is/>When T local＞Traw is reached, the time delay of the m-th V2I link is T local+Tcompressed, otherwise T raw+Tcompressed; the delay of the m-th V2I link is expressed as
In addition, the V2I and V2V links have maximum transmit power constraints, denoted as
Converting the V2V reliability constraint into a computable form by using a feasible domain conversion method, and obtaining a computable expression of the V2I throughput by using a robust SINR correction method:
(1) The integration-based feasible region transformation method transforms the V2V reliability constraint into a computable form:
Since g k and g m,k satisfy independent unit index distributions, the reliability constraint can be translated by integration, namely:
(2) The method of correcting the V2I expression by using the robust SINR is converted into a mode which can be calculated:
The introduction of the auxiliary variable γ c, the V2I expression is obtained by solving the following problem:
Rm＝max E[log2(1+γc)]
Wherein, Epsilon represents the outage probability allowed by the V2I link, and the calculated expression of V2I is deduced as follows:
designing a parallel DNN network structure, and constructing a loss function of DNN, wherein the method comprises the following steps of:
(1) Taking into account the V2I, V V transmit power The local calculation proportion lambda m of the vehicle is adopted to design a parallel DNN framework capable of simultaneously deciding a plurality of variables;
Parallel DNN architectures, e.g. comprising a power decision unit for outputting A local calculation proportion decision unit is used for calculating lambda m;
(2) According to Lagrange theorem, introducing dual variables, deducing a Lagrange expression, and taking the Lagrange expression as a loss function of training DNN;
Solving the dual problem of the original problem, obtaining an iterative expression of the dual variable, and feeding back multipliers to DNN network parameters;
Training DNN parameters comprises the steps of:
(1) Initializing DNN parameters and dual variables: the number of samples is 100000, wherein the training samples are 70000, and the verification samples are 30000; initializing DNN parameters by using an Xavier, and setting the initial value of the dual variable to 0;
(2) The dual variable and the neural network parameters are updated by adopting a small-batch gradient method: the training procedure for DNN was performed using Python 3.6 with Tensorflow.1.7.1; the learning rate is set to 0.001, 4 hidden layers are assumed to exist in DNN, each hidden layer contains 200 units, and the mini-band size is 100; updating parameters by adopting an Adam training method; meanwhile, due to feedback of dual variables, the loss function is dynamically updated;
DNN reasoning: acquiring the transmitting power of V2I and V2V according to the trained DNN And locally calculating a ratio lambda m;
according to the hungarian algorithm, an optimal spectrum matching decision x m,k is obtained: (1) The most probable solution of each possible V2I-V2V pair is obtained, each possible solution is sequenced, and the optimal spectrum allocation decision is deduced.
2. A system for allocating imperfect CSI-oriented resources in an ultra-reliable internet of vehicles implementing the imperfect CSI-oriented resource allocation method in an ultra-reliable internet of vehicles of claim 1, characterized in that the imperfect CSI-oriented resource allocation system in an ultra-reliable internet of vehicles comprises:
the vehicle networking model building module is used for building a vehicle networking model of diversified services;
The calculable expression acquisition module is used for converting the V2V reliability constraint into a calculable form by using a feasible domain conversion method and obtaining a calculable expression of the V2I throughput by using a robust SINR correction method;
The DNN network structure construction module is used for designing a parallel DNN network structure and constructing a loss function of DNN;
The DNN parameter training module is used for training DNN parameters;
A DNN reasoning module for acquiring the transmitting power of V2I and V2V according to the trained DNN And locally calculating a ratio lambda m;
the spectrum matching decision acquisition module is used for acquiring an optimal spectrum matching decision x m,k according to the Hungary algorithm.
3. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the imperfect CSI oriented resource allocation method in the ultra-reliable internet of vehicles as claimed in claim 1.
4. An information data processing terminal, characterized in that the information data processing terminal is configured to implement a resource allocation system for imperfect CSI in an ultra-reliable internet of vehicles according to claim 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111384794.2A CN114257988B (en) | 2021-11-22 | 2021-11-22 | Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111384794.2A CN114257988B (en) | 2021-11-22 | 2021-11-22 | Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114257988A CN114257988A (en) | 2022-03-29 |
| CN114257988B true CN114257988B (en) | 2024-05-03 |
Family
ID=80792901
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111384794.2A Active CN114257988B (en) | 2021-11-22 | 2021-11-22 | Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114257988B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118139014B (en) * | 2024-04-08 | 2024-07-26 | 重庆理工大学 | MIoV joint interference resource allocation method and system based on robust learning |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018125686A2 (en) * | 2016-12-30 | 2018-07-05 | Intel Corporation | Methods and devices for radio communications |
| CN109302709A (en) * | 2018-09-14 | 2019-02-01 | 重庆邮电大学 | The unloading of car networking task and resource allocation policy towards mobile edge calculations |
| CN111328133A (en) * | 2018-12-17 | 2020-06-23 | 上海大学 | V2X resource allocation method based on deep neural network |
| WO2020131932A2 (en) * | 2018-12-17 | 2020-06-25 | Apple Inc. | Methods for simultaneous support of resource selection modes and configuration mechanisms for nr v2x sidelink |
| WO2020150495A1 (en) * | 2019-01-16 | 2020-07-23 | Apple Inc. | Sideline connection establishment design to support unicast and groupcast communication for nr v2x |
| WO2020168304A1 (en) * | 2019-02-14 | 2020-08-20 | Apple Inc. | Methods for simultaneous support of and switching between scheduled and ue autonomous resource selection modes for nr v2x sidelink |
| CN111800244A (en) * | 2019-04-01 | 2020-10-20 | 英特尔公司 | Design of physical side loop feedback channel of NR-V2X |
| CN112954651A (en) * | 2021-03-12 | 2021-06-11 | 南京航空航天大学 | Low-delay high-reliability V2V resource allocation method based on deep reinforcement learning |
| TW202139638A (en) * | 2020-02-24 | 2021-10-16 | 美商高通公司 | Channel state information (csi) learning |
| CN113595768A (en) * | 2021-07-07 | 2021-11-02 | 西安电子科技大学 | Distributed cooperative transmission algorithm for guaranteeing control performance of mobile information physical system |
| CN113613207A (en) * | 2020-06-12 | 2021-11-05 | 南京理工大学 | Internet of vehicles spectrum sharing method based on multi-agent reinforcement learning |
-
2021
- 2021-11-22 CN CN202111384794.2A patent/CN114257988B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018125686A2 (en) * | 2016-12-30 | 2018-07-05 | Intel Corporation | Methods and devices for radio communications |
| CN109302709A (en) * | 2018-09-14 | 2019-02-01 | 重庆邮电大学 | The unloading of car networking task and resource allocation policy towards mobile edge calculations |
| CN111328133A (en) * | 2018-12-17 | 2020-06-23 | 上海大学 | V2X resource allocation method based on deep neural network |
| WO2020131932A2 (en) * | 2018-12-17 | 2020-06-25 | Apple Inc. | Methods for simultaneous support of resource selection modes and configuration mechanisms for nr v2x sidelink |
| WO2020150495A1 (en) * | 2019-01-16 | 2020-07-23 | Apple Inc. | Sideline connection establishment design to support unicast and groupcast communication for nr v2x |
| WO2020168304A1 (en) * | 2019-02-14 | 2020-08-20 | Apple Inc. | Methods for simultaneous support of and switching between scheduled and ue autonomous resource selection modes for nr v2x sidelink |
| CN111800244A (en) * | 2019-04-01 | 2020-10-20 | 英特尔公司 | Design of physical side loop feedback channel of NR-V2X |
| TW202139638A (en) * | 2020-02-24 | 2021-10-16 | 美商高通公司 | Channel state information (csi) learning |
| CN113613207A (en) * | 2020-06-12 | 2021-11-05 | 南京理工大学 | Internet of vehicles spectrum sharing method based on multi-agent reinforcement learning |
| CN112954651A (en) * | 2021-03-12 | 2021-06-11 | 南京航空航天大学 | Low-delay high-reliability V2V resource allocation method based on deep reinforcement learning |
| CN113595768A (en) * | 2021-07-07 | 2021-11-02 | 西安电子科技大学 | Distributed cooperative transmission algorithm for guaranteeing control performance of mobile information physical system |
Non-Patent Citations (5)
| Title |
|---|
| Deep-Reinforcement-Learning-Based Mode Selection and Resource Allocation for Cellular V2X Communications;Xinran Zhang;《IEEE Internet of Things Journal》;20191227;全文 * |
| Energy-Efficient Joint Resource Allocation and User Association for Heterogeneous Wireless Networks with Multi-Homed User Equipments;Guanhua Chai;《2019 IEEE 89th Vehicular Technology Conference (VTC2019-Spring)》;20190627;全文 * |
| 人工智能在5G系统中的应用综述;章坚武;《电信科学》;20210531;全文 * |
| 基于车辆行为分析的车联网超可靠低时延通信关键技术;张海霞;刘文杰;薛彤;梁聪;;中兴通讯技术;20200219(第01期);全文 * |
| 面向不确定CSI随机接入网络的深度稳健资源分配;吴伟华;《通信学报》;20200731;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114257988A (en) | 2022-03-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111132077B (en) | Multi-access edge computing task unloading method based on D2D in Internet of vehicles environment | |
| Qin et al. | Federated learning and wireless communications | |
| CN111314889A (en) | Task unloading and resource allocation method based on mobile edge calculation in Internet of vehicles | |
| CN110839184B (en) | Method and device for adjusting bandwidth of mobile fronthaul optical network based on flow prediction | |
| CN108495340B (en) | Network resource allocation method and device based on heterogeneous hybrid cache | |
| CN110941675B (en) | Wireless energy supply edge calculation delay optimization method based on deep learning | |
| US10931538B2 (en) | Machine learning algorithms for quality of service assurance in network traffic | |
| CN112491957A (en) | Distributed computing unloading method and system under edge network environment | |
| Saravanan et al. | Design of deep learning model for radio resource allocation in 5G for massive iot device | |
| Zhang et al. | Endogenous security-aware resource management for digital twin and 6G edge intelligence integrated smart park | |
| CN114257988B (en) | Imperfect CSI-oriented resource allocation method and system in ultra-reliable Internet of vehicles | |
| CN114301794A (en) | LEOMEO double-layer satellite constellation-oriented interlayer link topology design method | |
| Liu et al. | Communication and computation efficient federated learning for Internet of vehicles with a constrained latency | |
| Zhao et al. | Multi-agent deep reinforcement learning based resource management in heterogeneous V2X networks | |
| Zheng et al. | Mobility-aware split-federated with transfer learning for vehicular semantic communication networks | |
| CN115866787A (en) | Network resource allocation method integrating terminal direct transmission communication and multi-access edge calculation | |
| CN117998420A (en) | DAG-based fragmented cascading distributed 6G network joint optimization method and device | |
| Kanaga Suba Raja et al. | Optimal resource allocation in 5G system using modified lion algorithm | |
| CN113784372A (en) | Joint optimization method for terminal multi-service model | |
| CN111343722A (en) | Cognitive radio-based energy efficiency optimization method in edge calculation | |
| CN118612754B (en) | Three-in-one terminal control system and method capable of intelligent networking | |
| CN116193405B (en) | Heterogeneous V2X network data transmission method based on DONA framework | |
| Xia | Wireless resource optimization in semantic communication-based cellular networks | |
| Zhang et al. | Latency guarantee for ubiquitous intelligence in 6g: A network calculus approach | |
| CN116489713A (en) | Resource allocation method based on secondary gradient descent |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |