A multi-objective optimization roadside unit network deployment method in intelligent transportation system

Abstract

The present invention provides a multi-objective optimized roadside unit networking deployment method in a smart transportation system. The method comprises: obtaining the road network data of the target area; analyzing the communication performance of the roadside unit RSU in the target area according to the road network data of the target area, and obtaining the RSU coverage distance data; constructing a MOP-based RSU deployment problem model based on cost and total vehicle coverage time according to the RSU coverage distance data in the target area, and using the NSGA‑II algorithm to solve the MOP-based RSU deployment problem model to obtain the RSU networking deployment strategy in the target area. The present invention designs a multi-objective RSU networking deployment strategy based on the NSGA‑II algorithm, which can achieve a higher total vehicle coverage time with the smallest possible deployment cost investment, and provides a series of deployment solutions that meet the requirements for the RSU networking deployment project, which can be selected and dynamically adjusted according to actual conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic system management, and in particular to a roadside unit networking deployment method for multi-objective optimization in an intelligent traffic system. Background Art

[0002] As social and economic levels continue to improve, cars, as an important means of transportation for daily commuting and short-distance travel, account for an increasing proportion of contemporary residents' consumption. The huge number of cars and the increasing travel mileage have also greatly increased carbon emissions.

[0003] Accelerating the construction of a green, efficient, safe, and intelligent ITS (Intelligent Traffic System) is a key move to improve traffic efficiency and ensure travel safety. In ITS, vehicles are no longer simple means of transportation, but mobile nodes in the city. However, relying solely on vehicle-to-vehicle (V2V) communication, vehicles cannot perceive the entire traffic system in a more comprehensive and detailed manner and cannot meet the needs of smart transportation. Building the Internet of Vehicles (IoV), promoting the realization of the vehicle-to-everything (V2X) communication mode, and accelerating the transformation of the "individual intelligence" of a single vehicle to the "group intelligence" of vehicle-road collaboration are the inevitable trend of the development of modern smart transportation systems and the only way to achieve transportation strategic goals.

[0004] RSU (Road Side Unit) plays a key role in realizing V2X communication. It can improve communication quality and stability, alleviate the problem of shortage of storage and computing resources, provide access to heterogeneous networks and cloud servers, and assist vehicles in more comprehensive and detailed perception of road environment and other vehicles, making coordinated scheduling, route planning, accident reporting, emergency warning and other vehicle-road collaborative applications a reality, improving traffic efficiency while ensuring travel safety, and making vehicle driving more intelligent.

[0005] However, the main challenge in deploying RSUs is to balance QoS (Quality of Service) and RSU deployment costs. The QoS of ITS mainly involves coverage, throughput, latency, and energy consumption. Studies have shown that the QoS of different RSU deployment schemes depends largely on factors such as the number of RSUs, the shape of the road where they are located, the particularity of the road section (such as accident-prone sections), the wireless access method, the vehicle's mobility pattern, and the temporal and spatial distribution. The complex and changing road environment affects the performance of vehicle-to-Roadside Unit (V2R) communication with RSUs, which in turn affects QoS. Considering the cost investment of RSU deployment and the relatively fixed location once installed, it is of great significance to formulate RSU deployment optimization strategies and determine the deployment location, number, and type of RSUs before the project is implemented to promote the construction of ITS with vehicle-road collaboration. The research on RSU networking optimization strategies can provide decision-making departments with deployment strategies that meet cost and QoS requirements for specific application scenarios as a preliminary guide for project implementation.

[0006] At present, the related research on RSU networking deployment strategy in the existing technology describes the RSU deployment problem as a combinatorial optimization problem. Under given scenarios and parameters, the optimal combination of RSU installation locations is determined in the target area under study to achieve higher QoS at a lower cost. There is a scheme to study the deployment strategy of a fixed number of RSUs to maximize the connection time between RSUs and vehicles, and to express the RSU deployment problem as a 0-1 knapsack problem. Degree centrality and closeness centrality are introduced as cost functions. It is believed that the centrality of the candidate position is positively correlated with the total connection time between the RSU and the vehicle at the candidate position, and the maximization of connection time is converted into maximization of node centrality.

[0007] A proposal has been made to use a Multi-Objective Evolutionary Algorithm (MOEA) for QoS and cost to determine the deployment location of RSUs in urban areas, and to use the time that vehicles are connected to RSUs as a QoS metric. The proposed MOEA has higher performance than the greedy algorithm, but the resulting deployment solution does not have a good uniform coverage of the map, but is more concentrated on sections with more vehicles.

[0008] With the development of edge computing technology, some scholars have studied the deployment of RSUs as edge nodes, and jointly optimized RSU deployment and task offloading. One scheme aims to minimize the delay of task offloading and studies the optimal RSU deployment problem under uncertain road traffic density and vehicle task requirements. Particle Swarm Optimization (PSO) is used to determine the optimal deployment location of RSUs in the region, and Monte Carlo Stochastic Simulation (MCSS) outputs the expectation of task offloading delay at a given location. Due to the high computational complexity of PSO and MSCC, artificial neural networks are used to effectively predict the complex internal relationship between the uncertain environment and the output of the optimization objective function, which greatly speeds up the problem-solving process.

[0009] In the related research on dynamic RSU deployment strategy, considering using moving nodes in IoV as RSU can reduce the initial installation cost to a certain extent. Dynamic RSU can be temporarily enabled according to the needs of business scenarios to achieve more flexible deployment configuration, and the later operation cost is also greatly reduced. For post-accident notification applications, a scheme inspired by the colony of social insects proposes to use vehicles equipped with DSRC equipment as temporary RSUs. As temporary RSUs, vehicles can stop briefly and take on the tasks of regular RSUs, that is, transmit information to nearby vehicles and act as a communication bridge for other vehicles in the network. The author designed a corresponding algorithm to determine which vehicles are used as temporary RSUs and how long the vehicles used as temporary RSUs stop to broadcast messages. Simulation results show that using vehicles as temporary RSUs in post-accident notification applications can greatly improve message accessibility (the proportion of vehicles receiving messages in the network) while greatly saving costs.

[0010] The shortcomings of the research solutions for RSU networking deployment in the above-mentioned prior art include: these solutions are mainly based on single-objective optimization in optimization problems, and single-objective optimization has obvious limitations, that is, only the only optimal solution under a specific optimization goal can be obtained, and it is impossible to provide a variety of solution options for actual projects. In addition, coverage distance is an important performance parameter of RSU. Most studies use theoretical data provided by manufacturers. This data is usually determined by the signal strength being higher than the set threshold, and does not take into account the impact of the communication environment on the coverage distance. Only in the accident reporting scenario will researchers consider the performance of V2R communication and data transmission, such as latency, throughput, and packet delivery rate. Often, such studies will ignore road topology due to their focus on communication, and use a relatively simple grid-shaped road network to simulate and analyze the performance of the networking strategy. The actual road network topology is usually more complex, and the length, width, and vehicle density of the road section are different. Summary of the invention

[0011] An embodiment of the present invention provides a multi-objective optimized roadside unit networking deployment method in an intelligent transportation system, so as to effectively improve the RSU networking deployment efficiency in the intelligent transportation system.

[0012] In order to achieve the above object, the present invention adopts the following technical scheme.

[0013] A multi-objective optimized roadside unit network deployment method in an intelligent transportation system, comprising:

[0014] Obtain road network data of the target area;

[0015] Analyze the communication performance of the roadside unit RSU in the target area according to the road network data of the target area, and obtain RSU coverage distance data;

[0016] According to the RSU coverage distance data in the target area, a MOP-based RSU deployment problem model based on cost and total coverage vehicle time is constructed, and the NSGA-II algorithm is used to solve the MOP-based RSU deployment problem model to obtain the RSU networking deployment strategy in the target area.

[0017] Preferably, the obtaining of road network data of the target area includes:

[0018] The road network data osm file is obtained from the open street map OpenStreetMap. The osm file contains road information, building information and traffic information with data labels and classification. The main road data is extracted from the original osm file according to the data labels using the JOSM tool. After removing redundant information from the main road data, the data is converted into a SUMO-readable net.xml file that retains the road topology information. The net.xml file represents the sub-lanes of the main road in the form of geographical separation, and sets the intersection of each sub-lane as a node. The netedit program of SUMO is used to merge the adjacent single geographically separated sub-lanes and nodes in the net.xml file to obtain a map file that retains the lane information and topological connection of the main road.

[0019] Preferably, analyzing the communication performance of the roadside unit RSU in the target area according to the road network data of the target area to obtain the RSU coverage distance data includes:

[0020] PDR is used as a communication performance measurement indicator, and an RSU communication performance analysis model based on C-V2X mode 4 is designed. The RSU communication performance analysis model is used to simulate and calculate the PDR under four error type scenarios, and PDR-distance curves under different traffic environments and transmission powers are drawn. The four error types include: errors caused by half-duplex communication, errors caused by the received signal power being lower than the receiver power perception threshold, errors caused by propagation, and errors caused by data packet collision. The RSU communication performance analysis model analyzes the impact of different traffic density, transmission power, propagation conditions and modulation coding methods on the RSU communication performance according to the simulation calculation results, and obtains RSU coverage distance related data under different communication environments according to the analysis results of the RSU communication performance; it is set that when PDR>95%, the data packet can be transmitted correctly, and the vehicle is within the coverage range of the RSU, and the distance value when PDR=95% in the PDR-distance curve is read, and the distance value is used as the RSU coverage distance.

[0021] Preferably, the RSU deployment problem model based on MOP based on cost and total coverage vehicle time is constructed according to the RSU coverage distance data in the target area, including:

[0022] The road between two nodes in the road network is defined as a road segment. Assume that there are N road segments in the road network and the set of all road segments is S = {s1, s2, ..., s N}, there are M nodes in total, and the set of all nodes is P = {p1, p2, ..., p M}, for road segment s i (s i ∈S) corresponds to a set of certain nodes {p k ,p t}(p k ,p t ∈P) and traffic density β i , R={R1,R2,...,R q} is a set of RSUs available for deployment, with a total of q types. Each road section is used as a candidate location for RSU deployment. i The deployment scheme on dp i Indicates that dp i Contains the deployment location and type information of the deployed RSU. The RSU deployment strategy of the entire road is expressed as DP = {dp1, dp2, ..., dp N};

[0023] Set the deployment policy dp i The corresponding deployment cost is totalcost i , totalcost i Contains device cost determined by device typei and the installation location determines the deployment cost i , deployment strategy dp i The RSU in the road section s j The coverage length on is len i,j , the vehicle moves in the road network, and the sum of the time the vehicle stays within the coverage of the RSU is used as an indicator to measure the performance of the deployment scheme. The deployment scheme dp i On road sections j The total time of all vehicles covered is totaltime i,j , the MOP-based RSU deployment problem model based on cost and total coverage vehicle time is formulated as:

[0024] Minimize the total deployment cost min totalcost DP , maximize the RSU coverage time of the vehicle maxtotaltime DP ;

[0025]

[0026]

[0027] Preferably, the adopting of the NSGA-II algorithm to solve the MOP-based RSU deployment problem model to obtain the RSU networking deployment strategy in the target area includes:

[0028] Set segments i The corresponding node is {p k ,p t}, road segment s i The total length is l k,t , assuming that three RSUs with different transmission powers are provided, represented by R1, R2, and R3 respectively, then the section s i The set of RSUs available for deployment is R = {R1, R2, R3}, and the deployment strategy dp i It includes information about the device type and deployment location, encoded using floating point numbers, with the integer part recorded as int i , the decimal part is recorded as dec i , then the road segment s i The deployment strategy on is encoded as dp i =int i +dec i The overall deployment strategy DP is encoded as a vector consisting of a set of floating point numbers, and the integer part int i The value is any integer from 0 to 3, indicating the road section s i The type of RSU deployed on the server, int i=q(q∈{1,2,3}) represents the road segment s i The RSU type deployed on the q , int i =0 means that there is no RSU deployed on this road section, and the decimal part dec i (dec i ∈[0,1)) indicates the specific location of the RSU in the road section, indicating that the RSU is located in the road section s i Into node p k The distance is dec i ·l k,t at the location of

[0029] For the deployment scheme DP = {dp1, dp2, ..., dp N} Each element dp i , calculate the device cost devicecost respectively i and deployment cost i , add the equipment cost and installation cost to get dp i Deployment cost totalcost i , according to formula (4-17-1), the deployment costs of all elements in DP are summed up to obtain the total deployment cost of the deployment scheme DP;

[0030] For road segment s i Deployment strategy dp i , which is on section s j Total time of covered vehicles totaltime i,j By RSU on section s j Coverage distance, road section s j The number of vehicles and the speed of vehicles are jointly determined. Assuming the deployment plan dp i The RSU in the road section s j The coverage distance on is len i,j , assuming that road segment s j The traffic density is β j vehicles / m, the vehicle speed is v j , then the number of vehicles NV within the coverage of RSU i,j It is expressed as:

[0031] NV i,j =β j ·len i,j (4-18)

[0032] For each car, assume its speed is v j , then a single vehicle on road section s j When driving inside, the deployment plan dp iThe time covered by the RSU in is expressed as:

[0033]

[0034] Road sections i Deployment strategy dp i On road sections j The total time for all vehicles covered is:

[0035] totaltime i,j =NV i,j time i,j (4-20)

[0036] Overall deployment plan DP = {dp1, dp2, ..., dp N The total time that the corresponding RSU covers the vehicle is:

[0037]

[0038] In the NSGA-II algorithm, gen represents the number of generations. max represents the maximum number of generations as the algorithm stopping condition, P and Q represent the parent population and the child population respectively. Each individual in the population represents a deployment strategy, including the number, type and deployment location of RSU. Assume that the population size is N size After merging the parent and child populations, calculate the objective function value of each individual according to formula (4-17-1) and (4-17-2), sort all individuals by non-dominated order according to Pareto rank, and add individuals to the new population P in order from small to large Pareto rank. new , if we add individuals with Pareto level i, P new The number of individuals is less than N size , and after adding individuals with Pareto level i+1, P new The number of individuals is greater than N size , then the individuals with Pareto level i+1 are sorted in descending order using the crowding operator, and individuals with low crowding are eliminated until P new The number of individuals with the remaining Pareto level i+1 added to the size , and finally obtain the individuals, that is, obtain the RSU network deployment strategy in the final target area;

[0039] The definition of Pareto dominance is as follows:

[0040] Definition 1 (Dominance Relation): For minimizing MOP, any two decision vectors x a ,x b ;

[0041] If (1) All have f i (x a )≤f i (x b ) is established;

[0042] (2) So that f j (x a ) <f j (x b ) is established;

[0043] Then x a Dominate x b ;

[0044] If (1) So that f i (x a )≤f i (x b ) is established;

[0045] (2) So that f j (x a )>f j (x b ) is established;

[0046] Then x a With x b They do not dominate each other;

[0047] Definition 2 (Pareto optimal solution): For a decision vector, if there is no other decision vector that can dominate it, then the decision vector is called a non-dominated solution or a Pareto optimal solution;

[0048] Definition 3 (Pareto set): For a given set of optimal solutions, if the solutions in this set are mutually non-dominated, then this solution set is called a Pareto set;

[0049] Definition 4 (Pareto front): The set of images of each solution in the Pareto set in the objective function space is called the Pareto front.

[0050] Definition 5 (Pareto rank): In any set of feasible solutions, the Pareto rank of a non-dominated solution is defined as 1. When a non-dominated solution is deleted from the set, the Pareto rank of the new non-dominated solution generated from the remaining solutions is defined as 2. Similarly, the Pareto rank of each solution in the feasible solution set can be obtained.

[0051] It can be seen from the technical solutions provided by the above embodiments of the present invention that the present invention designs a multi-objective RSU networking deployment strategy based on the NSGA-II algorithm for the RSU networking deployment problem in the actual road network, which improves the total vehicle coverage time of the solution while saving costs, and can provide a series of networking strategies that meet the requirements for project implementation to facilitate dynamic adjustment. A higher total vehicle coverage time can be achieved with the smallest possible deployment cost, providing a series of deployment solutions that meet the requirements for RSU networking deployment projects for selection and dynamic adjustment based on actual conditions.

[0052] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0054] Figure 1 A processing flow chart of a multi-objective optimized roadside unit networking deployment method in an intelligent transportation system provided by an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of obtaining and simplifying the road network of Yizhuang Demonstration Area provided in an embodiment of the present invention;

[0056] Figure 3 A schematic diagram of an RSU communication performance analysis model provided by an embodiment of the present invention;

[0057] Figure 4 A PDR and four error occurrence probabilities δ provided in an embodiment of the present invention HD ,δ SEN ,δ PRO ,δ COL Schematic diagram of the curve of the change with the distance between the vehicle and the RSU;

[0058] Figure 5 An MCS#1 provided in an embodiment of the present invention, β=0.1 vehicle / m, λ=10 Hz, different transmission powers (P t =20, 23, 30dBm) under PDR variation with distance;

[0059] Figure 6 An MCS#1,P provided in an embodiment of the present invention t=20dBm,λ=10Hz, PDR vs. distance curves at different traffic densities (β=0.1, 0.2, 0.3 vehicles / m);

[0060] Figure 7 A flow chart of an NSGA-II algorithm provided in an embodiment of the present invention;

[0061] Figure 8 A schematic diagram of a Xiong'an Rongdong RSU networking deployment strategy based on the NSGA-II algorithm provided in an embodiment of the present invention;

[0062] Fig. 9 This is a schematic diagram of the Beijing Yizhuang RSU networking deployment strategy based on the NSGA-II algorithm. DETAILED DESCRIPTION

[0063] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0064] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items.

[0065] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0066] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0067] The embodiment of the present invention proposes a static W-RSU networking deployment strategy in an actual road network. By analyzing the communication conditions to obtain coverage data of RSU communication performance, the RSU deployment optimization problem is described as the MOP (Multi-Objective Optimization Problem) of cost and QoS. The total time of RSU coverage of vehicles in the road network is used as the QoS measurement indicator. A deployment optimization strategy based on NSGA-II is proposed, which provides project decision makers with a series of RSU networking deployment solutions with more cost-effectiveness and better QoS performance.

[0068] The embodiment of the present invention provides a multi-objective optimized roadside unit networking deployment method in a smart transportation system based on the actual road network of the Beijing Yizhuang High-level Autonomous Driving Demonstration Area and the Xiong'an New Area Rongdong area obtained from OpenStreetMap, which is simplified and relevant road topology information is extracted. Based on the road network data, the C-V2X mode 4 communication model is adopted to analyze the impact of environmental factors such as traffic flow, road topology, communication distance between RSU and vehicles, and equipment factors such as RSU transmission power, data packet size, and data packet sending frequency on RSU communication performance. The PDR is calculated based on a variety of data transmission errors as a communication energy evaluation indicator, and then the RSU coverage of different types of equipment at different locations is obtained based on the analysis of RSU communication performance.

[0069] In the strategy design phase, the RSU deployment optimization problem is formulated as a MOP. The deployment cost and the sum of the time that the vehicle is within the RSU coverage area are used as the two sub-goals of the MOP. The NSGA-Ⅱ algorithm is used to solve the problem and obtain an approximate optimal solution set. Based on traffic data, moving vehicles are generated, and the ratio of the time that the vehicle is within the RSU coverage area to the total time that the vehicle is moving in the road network is calculated, which is used as an online evaluation indicator for the deployment plan. By combining algorithm performance evaluation, online evaluation of the plan, and decision maker preferences, decision makers can choose the most appropriate deployment plan.

[0070] The embodiment of the present invention proposes a processing flow chart of a multi-objective optimized roadside unit network deployment method in an intelligent transportation system. Figure 1 As shown, the processing steps include the following:

[0071] Step S10: Acquire the road network data of the target area and simplify the road network data.

[0072] Before designing the RSU networking deployment strategy, it is necessary to obtain the road network data and simplify it to suit the RSU deployment strategy. Taking Yizhuang Demonstration Area as an example, the data acquisition and simplification process is as follows.

[0073] Figure 2A schematic diagram of obtaining and simplifying the road network of Yizhuang Demonstration Area provided by an embodiment of the present invention. Figure 2 (a). First, obtain the road network data osm file from OpenStreetMap, as shown in Figure 2 (b) is shown. The osm file is a collaborative online map project that constructs free content. Its goal is to create a world map with free content that can be edited by everyone, and to provide convenient navigation solutions for generally inexpensive mobile devices. The map data file suffix downloaded from the website is .osm. The osm file contains road information, building information, traffic information, etc., with data labels and classifications. Since RSU deployment needs to be carried out on roads, traffic information such as building information, bus stop locations, and some roads where RSU cannot be deployed, such as bicycle lanes, pedestrian overpasses, and roads under construction, are redundant information that needs to be deleted. JOSM is an osm map file editor developed based on Java. It can provide osm map file editing and convenient and powerful data label management, and supports the editing and conversion of various map files. First, JOSM is used to extract the main road data from the original osm file according to the data label. After removing redundant information from the main road data, it is converted into a net.xml file readable by SUMO. After removing redundant information and extracting main roads, the main road topology is retained, such as Figure 2 (c) As shown. For some complex road topologies such as large intersections and overpasses, the sub-lanes of the main road are represented in the form of geographical separation in net.xml, and the intersections of each sub-lane are set as nodes. However, these nodes that are very close to each other are often located at the same large intersection in reality. The separation in net.xml is not conducive to node selection when specifying the RSU deployment strategy. Therefore, the netedit program of SUMO is further used to merge the adjacent single geographically separated sub-lanes and nodes. The final map file only retains the lane information and topological connections of the main road, as shown in Figure 2 (d) as shown.

[0074] Step S20: Analyze the communication performance of the RSU in the target area according to the road network data to obtain the RSU coverage distance data.

[0075] Most RSU deployment-related studies usually consider the deployment of a single type of RSU, that is, the coverage distance of the RSU is a fixed value. Some studies believe that the transmission power will affect the coverage distance of the RSU, and consider using RSUs with different transmission powers in the deployment plan. The greater the transmission power, the greater the coverage distance, and the corresponding equipment cost will also increase. However, these studies have not taken into account the impact of the complex and changeable communication environment around the RSU deployment location on its communication performance, which in turn affects the coverage distance of the RSU.

[0076] Figure 3 A schematic diagram of an RSU communication performance analysis model provided by an embodiment of the present invention is shown in FIG. Figure 3 As shown, in order to get rid of the limitations of using theoretical values ​​of coverage distance in RSU deployment strategy research and obtain RSU coverage distance data suitable for different communication scenarios, the present invention adopts PDR (Packet Delivery Ratio) as a communication performance measurement indicator, designs an RSU communication performance analysis model based on C-V2X Mode 4, simulates and analyzes the impact of different traffic density, transmission power, propagation conditions, modulation and coding methods on RSU communication performance, and obtains RSU coverage distance related data under different communication environments based on the analysis of communication performance.

[0077] PDR means the percentage of data packets that are successfully and correctly transmitted. The successful transmission of data packets is a necessary prerequisite for RSU to provide various services to vehicles. Using PDR to measure V2R communication performance pays more attention to the reliability of communication and the impact of traffic congestion on V2R communication. It is more suitable for IoVs with spatiotemporal changes in traffic flow and safety-related applications. The calculation of PDR considers four mutually exclusive error types, as described below:

[0078] (Error 1) Errors caused by half-duplex communication: V2R communication uses half-duplex communication. Both the vehicle and the RSU can send and receive data. However, for the same node, data cannot be sent and received at the same time. In the same subframe, if a node is sending a data packet, it will not be able to receive data packets sent to it by other nodes, resulting in data packet transmission errors. This error type is irrelevant to the distance between the vehicle and the RSU. Errors caused by duplex communication are represented by e HD It means that the probability of its occurrence is δ HD .

[0079] δ HD =Pr(e=e HD ) (4-1)

[0080] (Error 2) The received signal power is lower than the receiver power perception threshold. SEN If the received power is lower than this threshold, the receiver cannot perceive the data and decode it. The error caused by the receiver being unable to perceive the data due to the received power being lower than the perception threshold is SEN It means that the probability of its occurrence is δ SEN .

[0081] δ HD =Pr(e=e SEN |e≠e HD ) (4-2)

[0082] (Error 3) Error caused by propagation: The data packet propagation in the channel will be affected by noise, and even though the node reception rate is higher than P SEN However, due to the low signal-to-noise ratio (SNR), the data cannot be received correctly. In this case, the interference from other nodes is not considered, only the channel noise is considered. The error caused by propagation is represented by e PRO It means that the probability of its occurrence is δ PRO .

[0083] δ PRO =Pr(e=e PRO |e≠e HD ,e≠e SEN ) (4-3)

[0084] (Error 4) Error caused by packet collision: When two vehicles send packets to the RSU in the same subframe, packet collision will cause the signal to interference plus noise ratio (SINR) to be too low, and the RSU cannot receive the data correctly. This error does not include the errors in Error 1-Error 3. Errors caused by packet collision are represented by e COL It means that the probability of its occurrence is δ COL .

[0085] δ COL =Pr(e=e COL |e≠e HD ,e≠e SEN ,e≠e PRO ) (4-4)

[0086] Assume that the packet size is B bytes and the transmission frequency is λ Hz, that is, each vehicle sends λ packets per second, and the transmission power is P t , the received power is P r , the distance between the sender and the receiver is d t,r , the total channel bandwidth is 10MHz or 20MHz, divided into 2 or 4 sub-channels, and QPSK encoding is used. Considering four mutually exclusive error types, PDR can be calculated by formula (4-5):

[0087] PDR(d t,r )=(1-δ HD )·(1-δ SEN (d t,r ))·(1-δ PRO (d t,r ))·(1-δ COL (d t,r )) (4-5)

[0088] Where PDR(d t,r ),δ HD ,δ SEN (d t,r ),δ PRO (d t,r ),δ COL (d t,r ) respectively represent the distance between the transmitter and receiver as d t,r The packet delivery rate and the probability of four mutually exclusive errors when δ HD ,δ SEN (d t,r ),δ PRO (d t,r ),δ COL (d t,r ) is normalized to obtain Formula (4-6) can be expressed as

[0089]

[0090] in,

[0091]

[0092]

[0093]

[0094]

[0095] Errors caused by half-duplex

[0096] The distance between the sender and the receiver does not affect the probability of half-duplex errors. This error type occurs when the vehicle and RSU that need to communicate select the same subframe to send data. Assuming that the subframe length is 1ms and the data packet transmission frequency is λHz, then δ HD It can be approximately calculated by formula (4-7).

[0097]

[0098] Errors caused by received signal power being lower than the perception threshold

[0099] Error 2 and Error 3 are mutually exclusive, so when calculating δ SEN Path loss and shadow fading are considered, and the influence of noise is not added. The received signal power at distance d t,r The probability density function at Indicates that, as shown in formula (4-8), the received signal power is lower than P SEN The error probability caused can be expressed by formula (4-9)

[0100]

[0101]

[0102] Using the complementary error function, equation (4-9) can be rewritten as equation (4-10).

[0103]

[0104] Errors caused by propagation

[0105] When data packets propagate in the channel, they will be affected by noise. If the received SNR is too low, even if the received power is greater than P SEN , the receiver cannot receive the signal correctly. Considering the mutual exclusion relationship between different errors, calculate δ PRO The interference from other nodes is not considered. The distance between the transceivers is d t,r When the noise power is N0, the SNR at the receiving node is:

[0106] SNR(d t,r )=P r -N0=P t -Loss path (d t,r )-Loss shadow -N0 (4-11)

[0107] SNR(d t,r ) is subject to the mean value P t -Loss path (d t,r )-N0, with a standard deviation of σ shadow Normal distribution, distance d t,r The probability density function of SNR at is:

[0108]

[0109] There are two possible reasons for low SNR: one is that the received power is too low, and the other is that the channel noise is too high. Considering the mutual exclusivity of error types, calculate δ PRO Considering the error caused by propagation noise when the received power is higher than the perception threshold, it is necessary to exclude the case where the received power is lower than the perception threshold. This type of data transmission error has been calculated in formula (4-10). r >P SEN In the case of , the probability density function of SNR is obtained as:

[0110]

[0111] The probability of error due to propagation is:

[0112]

[0113] Errors caused by packet collision

[0114] When the vehicle v t When sending data to RSU, if there is an interfering vehicle v i Data is transmitted on the same sub-channel in the same sub-frame, and v i If the interference is large enough to cause packet loss, a packet collision will occur, resulting in an error in which the data cannot be received correctly. i The probability of packet collision error is It can be calculated by formula (4-15)

[0115]

[0116] where p SIM (d t,i ) is the vehicle v t and v i The probability of using the same sub-channel in the same sub-frame to transmit data, p INT (d t,r ,d i,r ) is the vehicle v i The probability that the interference is large enough to cause the SINR to be too low for the receiver to receive correctly, d t,i ,d t,r ,d i,r The vehicle v t With interference vehicle v i The distance between t The distance to the RSU and v i The distance between the vehicle and the RSU. Then, all interfering vehicles cause the vehicle v t The probability of packet collision error between RSU and RSU can be expressed as:

[0117]

[0118] (2) Simulation results

[0119] The V2R communication performance analysis model based on C-V2X mode 4 proposed in this invention is simulated and implemented in Matlab. Simulation of PDR and the probability of occurrence of four error types under different scenarios δ HD ,δ SEN ,δ PRO ,δ COL As the distance between the vehicle and the RSU changes. Assuming the traffic density β = 0.1 vehicle / m, the transmission power P t=20dBm, transmission frequency λ=10Hz, and MCS#1 modulation coding mode is used as the standard scenario to analyze the impact of traffic density, transmission frequency, transmission power, and modulation mode on PDR and the probability of occurrence of four errors. The meaning of each parameter in the simulation and the parameter settings are shown in Table 1.

[0120] Table 1 Communication performance analysis simulation parameters

[0121]

[0122] PDR and four error probability δ HD ,δ SEN ,δ PRO ,δ COL The curve of the distance between the vehicle and the RSU is as follows: Figure 4 shown.

[0123] It is believed that PDR>95% can ensure that most data packets are transmitted correctly, the communication reliability between the vehicle and the RSU is high, and a higher quality service can be obtained. At this time, the vehicle is within the coverage range of the RSU. The PDR-distance curve under different traffic environments and transmission powers is simulated, and the distance when PDR=95% is read, and this distance is used as the RSU coverage distance.

[0124] From the simulation results, it can be seen that the PDR of V2R communication decreases with the increase of distance. This decrease is mainly due to the increase in the probability of data transmission error when the received power is lower than the perception threshold. SEN The probability of data transmission error caused by packet collision δ COL When the distance is less than 200m, the PDR does not decrease significantly with the increase of distance. When the distance is greater than 200m, the PDR decreases more with the increase of distance. When the distance increases from 0 to 200m, the PDR only decreases by about 3.6%. The PDR at 200m is 95.4%. When the distance increases from 200m to 400m, the PDR decreases by about 15%. When the distance between the RSU and the vehicle is 500m, the PDR is only 29.4%. When the distance between the RSU and the vehicle is less than 290m, the PDR can be guaranteed to be > 90%, and when the distance between the RSU and the vehicle is less than 210m, the PDR can be guaranteed to be > 95%.

[0125] Figure 5 An MCS#1 provided in an embodiment of the present invention, β=0.1 vehicle / m, λ=10 Hz, different transmission powers (P t =20, 23, 30dBm) under PDR and distance variation curve. Figure 5 As shown in Figure 2, the transmission power has a great influence on the RSU coverage. Assuming that a vehicle performing V2R communication is considered to be within the coverage of the RSU if its PDR>90%, P t=20dBm, the distance between RSU and vehicle is less than 280m to ensure PDR>90%, P t =23dBm, this range is extended to 310m. t When increased to 30dBm, the range reaches 370m, which shows that increasing the transmit power can effectively increase the coverage of the RSU, but this is also accompanied by an increase in cost.

[0126] Figure 6 An MCS#1,P provided in an embodiment of the present invention t =20dBm,λ=10Hz, PDR vs. distance curves at different traffic densities (β=0.1, 0.2, 0.3 vehicles / m). Figure 6 As shown in the figure, traffic density at different distances has an impact on PDR. When the distance is the same, the higher the traffic density, the smaller the PDR. Traffic density has the greatest impact on PDR in the range of 250-400m. At the same distance, the difference in PDR when β = 0.1 and β = 0.3 shows a trend of first increasing and then decreasing with the increase of distance. The difference is the largest at 370m, reaching 0.365. Traffic density has a great impact on the coverage range of RSU. When β = 0.1, the V2R communication distance is less than 280m to ensure PDR>90%; when β increases to 0.2, this distance is reduced to 200m; in the high traffic density scenario of β = 0.3, the RSU coverage range is 150m.

[0127] The above analysis further proves that the method proposed in the present invention for analyzing the communication performance of RSU and obtaining RSU coverage distance data is more scientific and more in line with actual engineering needs than using the theoretical data of RSU coverage distance provided by equipment manufacturers.

[0128] Acquisition of RSU coverage distance data: Based on the RSU communication performance analysis model, the present invention uses PDR as an indicator to determine the RSU coverage range. Combined with relevant data, we believe that there are about 6 vehicles per 100 meters on a single lane in the Yizhuang Demonstration Area, that is, β = 0.06, and about 4 vehicles per 100 meters on a single lane in the Xiongan New Area, that is, β = 0.04. Combined with the road topology, most of the roads in the two road networks are two-lane or four-lane, and the largest trunk road can reach 6 lanes. Based on the above settings, combined with the C-V2X mode 4 communication performance analysis model, the coverage range of RSUs with different transmission powers in different regions and different road topologies can be obtained, as shown in Table 2.

[0129] Table 2 Comparison of RSU coverage, road topology and transmission power

[0130]

[0131] Step S30: construct an MOP-based RSU deployment problem model based on cost and total coverage vehicle time according to the RSU coverage distance data in the target area, and use the NSGA-II algorithm to solve the MOP-based RSU deployment problem model to obtain the RSU networking deployment strategy in the target area.

[0132] The road between two nodes in the road network is defined as a road segment, and the road segment is used as a candidate location for RSU deployment. Assume that there are N road segments in the road network, and the set of all road segments is S = {s1, s2, ..., s N}, there are M nodes in total, and the set of all nodes is P = {p1, p2, ..., p M}, for road segment s i (s i ∈S) corresponds to a set of certain nodes {p k ,p t}(p k ,p t ∈P) and traffic density β i R={R1,R2,...,R q} is a set of RSUs available for deployment, with a total of q types. Different types of RSUs have different equipment costs and transmission powers. Each road section can be used as a candidate location for RSU deployment. i The deployment scheme on dp i It indicates that the deployment location and the type of deployed RSU are included. Then the entire deployment strategy can be expressed as DP = {dp1, dp2, ..., dp N}.

[0133] For deployment strategy dp i , and its corresponding deployment cost is totalcost i , including the device cost devicecost determined by the device type i and the installation location determines the deployment cost i . Deployment strategy dp i The RSU in the road section s j The coverage length on is len i,j , the vehicle moves in the road network, and the sum of the time the vehicle stays within the coverage of the RSU is used as an indicator to measure the performance of the deployment scheme. The deployment scheme dp i On road sections j The total time of all vehicles covered is totaltime i,j , which is related to the vehicle speed, RSU coverage distance and traffic density. Then the MOP-based RSU deployment problem model based on cost and total coverage time of vehicles can be expressed as:

[0134] Minimize the total deployment cost min totalcost DP , maximize the RSU coverage time of the vehicle maxtotaltime DP ;

[0135]

[0136]

[0137] Solving the above-mentioned MOP-based RSU deployment problem model using a Pareto-dominance-based method can well preserve the original data and not lose relevant information, thereby ensuring the quality of the solution. Evolutionary algorithms are a type of commonly used algorithm for solving MOP problems based on the Pareto-dominance idea, among which NSGA-Ⅱ uses fast non-dominated sorting to speed up convergence, merges parent and child generations, and introduces an elite retention strategy, which expands the search space and ensures the diversity and uniformity of the solution. It is a MOEA algorithm with good performance and is widely used in solving MOP. The present invention uses NSGA-Ⅱ to solve the MOP-based RSU deployment problem.

[0138] An NSGA-II algorithm process provided by an embodiment of the present invention is as follows: Figure 7 As shown, gen represents the generation number, gen max represents the maximum number of generations as the algorithm stop condition, P and Q represent the parent population and the child population respectively, and each individual in the population represents a deployment strategy, including the number, type and deployment location of RSU. The core of the algorithm is the elite retention strategy with fast non-dominated sorting and congestion calculation. Assume that the population size is N size After merging the parent and child populations, calculate the objective function value of each individual according to formulas (4-17-1) and (4-17-2). Sort all individuals by non-dominated order according to the Pareto rank, and add individuals to the new population P in order from small to large Pareto rank. new , if we add individuals with Pareto level i, P new The number of individuals is less than N size , and after adding individuals with Pareto level i+1, P new The number of individuals is greater than N size , then the individuals with Pareto level i+1 are sorted in descending order using the crowding operator, and individuals with low crowding are eliminated until P new The number of individuals with the remaining Pareto level i+1 added to the size The final individuals are obtained, that is, the RSU network deployment strategy in the final target area is obtained.

[0139] The definition of Pareto dominance is as follows:

[0140] Definition 1 (Dominance Relation): For minimizing MOP, any two decision vectors x a ,x b ;

[0141] If (1) All have f i (x a )≤f i (x b ) is established;

[0142] (2) So that f j (x a ) <f j (x b ) is established;

[0143] Then x a Dominate x b ;

[0144] If (1) So that f i (x a )≤f i (x b ) is established;

[0145] (2) So that f j (x a )>f j (x b ) is established;

[0146] Then x a With x b They do not dominate each other;

[0147] Definition 2 (Pareto optimal solution): For a decision vector, if there is no other decision vector that can dominate it, then the decision vector is called a non-dominated solution or a Pareto optimal solution;

[0148] Definition 3 (Pareto set): For a given set of optimal solutions, if the solutions in this set are mutually non-dominated, then this solution set is called a Pareto set;

[0149] Definition 4 (Pareto front): The set of images of each solution in the Pareto set in the objective function space is called the Pareto front.

[0150] Definition 5 (Pareto rank): In any set of feasible solutions, the Pareto rank of a non-dominated solution is defined as 1. When a non-dominated solution is deleted from the set, the Pareto rank of the new non-dominated solution generated from the remaining solutions is defined as 2. Similarly, the Pareto rank of each solution in the feasible solution set can be obtained.

[0151] The core of the NSGA-Ⅱ algorithm is the mapping from phenotype to genotype, and judging the quality of individuals by genotype, that is, encoding the deployment strategy, and designing the fitness function according to the encoded deployment strategy to calculate the fitness of different deployment strategies.

[0152] Deployment strategy encoding: For road segment s i , whose corresponding node is {p k ,p t}, the total length of the road section is l k,t Assume that three types of RSU are provided, with transmission powers of 20dBm, 23dBm, and 30dBm, respectively represented by R1, R2, and R3. Then the section s i The set of RSUs available for deployment is R = {R1, R2, R3}. Deployment strategy dp i The information about the device type and deployment location needs to be encoded using floating point numbers, with the integer part recorded as int i , the decimal part is recorded as dec i , then s i The deployment strategy on can be encoded as dp i =int i +dec i , the overall deployment strategy DP is encoded as a vector consisting of a set of floating point numbers. i The value is any integer from 0 to 3, indicating the road section s i The type of RSU deployed on the server, int i =q(q∈{1,2,3}) represents the road segment s i The RSU type deployed on the q , int i =0 means that there is no RSU deployed on this road section. i (dec i ∈[0,1)) indicates the specific location of the RSU in the road section, indicating that the RSU is located in the road section s i Into node p k The distance is dec i ·l k,t at the location.

[0153] Total cost model: For road segment s i Deployment strategy dp i , its deployment cost totalcost i Device cost i and equipment installation cost deployment cost iThe cost of the equipment itself is determined by the type of equipment. The cost of equipment with different transmission powers is different. Equipment with high transmission power is often equipped with more expensive RF circuits and higher gain antenna modules, and the cost is relatively high. The equipment installation cost is determined by the equipment installation location. This is because the inherent infrastructure status of different locations is different. For example, if an RSU is installed at an intersection, a traffic light can be used as a bracket. If an RSU is installed in the middle of a road without a traffic light, the installation cost will increase. The corresponding relationship between RSU equipment type, deployment location and deployment cost is shown in Table 3.

[0154] Table 3 Comparison of cost, RSU type and deployment location

[0155]

[0156] Deployment plan DP = {dp1, dp2, ..., dp N}Calculate the total cost of its deployment DP The process is shown in Algorithm 4-1. For each element dp i , lines 4 to 14 calculate the device cost devicecost i , lines 5 to 24 calculate the installation cost deploycost i , add the equipment cost and installation cost to get dp i Deployment cost totalcost i , traverse all elements in DP, and add up their deployment costs to get the total deployment cost of the deployment plan DP.

[0157]

[0158] Total coverage time model: For road segment s i Deployment strategy dp i , which is on section s j Total time of covered vehicles totaltime i,j By RSU on section s j Coverage distance, road section s j The number of vehicles and the speed of vehicles are jointly determined. Using the RSU coverage model proposed in Section 3.4, the coverage distance of RSU is related to the traffic density and RSU transmission power. Assuming the deployment scheme dp i The RSU in the road section s j The coverage distance on is len i,j Assume that road segment s j The traffic density is β j vehicles / m, the vehicle speed is v j , then the number of vehicles NV within the coverage of RSU i,j It can be expressed as:

[0159] NV i,j =β j ·len i,j (4-18)

[0160] For each car, assume its speed is v j , then a single vehicle on road section s j When driving inside, the deployment plan dp i The time covered by the RSU in can be expressed as:

[0161]

[0162] Furthermore, the road segment s i Deployment strategy dp i On road sections j The total time for all vehicles covered is:

[0163] totaltime i,j =NV i,j time i,j (4-20)

[0164] Overall deployment plan DP = {dp1, dp2, ..., dp N The total coverage time that can be achieved is:

[0165]

[0166] Step S40: RSU networking deployment strategy performance simulation.

[0167] (1) Visual display of RSU network deployment strategy

[0168] Xiongan Rongdong RSU network deployment strategy based on NSGA-II algorithm Figure 8 As shown in Figure 2, the Beijing Yizhuang RSU network deployment strategy based on the NSGA-II algorithm is as follows: Fig. 9 As shown, the darkest grayscale, the middle grayscale, and the lightest grayscale represent RSUs with transmission powers of 20dBm, 23dBm, and 30dBm, respectively.

[0169] (2) RSU Network Deployment Strategy Performance Evaluation

[0170] In order to evaluate the coverage time performance of the deployment scheme when facing real moving vehicles and verify the effectiveness of the scheme in practical applications, the present invention uses the SUMO traffic simulation platform to generate randomly moving vehicles and count the fixed time window T w The total time t that all vehicles travel in the road network move The time t when the vehicle is covered by RSU cover, calculate the coverage time ratio cmr = t cover / t move The simulation analysis of coverage time performance under different coverage rates is shown in Table 4. The simulation results based on traffic data show that the deployment strategy obtained by the NSGA-II algorithm can obtain results similar to the calculated total coverage time in dynamic traffic simulation, and is often better than the theoretical coverage rate, which proves the effectiveness of the algorithm and its robustness in the face of dynamic traffic data.

[0171] Table 4 Coverage time performance of NSGA-II deployment scheme based on traffic data

[0172]

[0173] In summary, compared with the existing RSU networking deployment strategy, the RSU networking deployment strategy proposed in the present invention can provide an RSU networking deployment solution for a real road network, rather than just staying in an abstract road network. Under the same circumstances, it has better cost-effectiveness and better QoS performance, can provide more diverse networking solutions, and is convenient for dynamic adjustment following project implementation. After the performance verification of the networking solution based on dynamic vehicle movement data, Henghao's performance is also demonstrated in actual dynamic traffic scenarios, providing a guarantee for the effect of project implementation.

[0174] The RSU networking strategy of the present invention is more suitable for actual road scenarios and better meets the actual engineering implementation needs. In terms of static algorithm performance evaluation and performance evaluation with the addition of dynamic traffic data, the RSU networking strategy of the present invention has better cost-effectiveness and performance.

[0175] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0176] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0177] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0178] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multi-objective optimized roadside unit network deployment method in an intelligent transportation system, characterized in that: include: Obtain road network data of the target area; Analyze the communication performance of the roadside unit RSU in the target area according to the road network data of the target area, and obtain RSU coverage distance data; According to the RSU coverage distance data in the target area, a MOP-based RSU deployment problem model based on cost and total coverage vehicle time is constructed, and the NSGA-II algorithm is used to solve the MOP-based RSU deployment problem model to obtain the RSU networking deployment strategy in the target area; The analyzing the communication performance of the roadside unit RSU in the target area according to the road network data of the target area to obtain the RSU coverage distance data includes: Using PDR as a communication performance measurement indicator, an RSU communication performance analysis model based on C-V2X mode 4 is designed. The RSU communication performance analysis model is used to simulate and calculate the probability of occurrence of four error types in different scenarios. The PDR is further calculated based on the probability of occurrence of the four error types, and the PDR-distance curve diagram under different traffic environments and transmission powers is drawn. The four error types include: errors caused by half-duplex communication, errors caused by the received signal power being lower than the receiver power perception threshold, errors caused by propagation, and errors caused by data packet collision. The RSU communication performance analysis model analyzes the impact of different traffic density, transmission power, propagation conditions and modulation coding methods on the RSU communication performance based on the simulation calculation results, and obtains RSU coverage distance related data under different communication environments based on the analysis results of the RSU communication performance; it is set that when PDR>95%, the data packet can be transmitted correctly, and the vehicle is within the coverage range of the RSU, and the distance value when PDR=95% in the PDR-distance curve is read, and the distance value is used as the RSU coverage distance; The RSU deployment problem model based on MOP based on cost and total coverage vehicle time is constructed according to the RSU coverage distance data in the target area, including: The road between two nodes in the road network is defined as a road segment. Assume that there are N road segments in the road network and the set of all road segments is S = {s1, s2, ..., s N }, there are M nodes in total, and the set of all nodes is P = {p1, p2, ..., p M }, for road segment s i Corresponding to a certain set of nodes {p k ,p t } and traffic density β i , where s i ∈S,p k ∈P,p t ∈P, R={R1,R2,...,R q } is a set of RSUs available for deployment, with a total of q types. Each road segment is used as a candidate location for RSU deployment. i The deployment strategy on dp i Indicates that dp i Contains the deployment location and type information of the deployed RSU. The RSU deployment strategy of the entire road is expressed as DP = {dp1, dp2, ..., dp N }; Set the deployment policy dp i The corresponding deployment cost is totalcost i , totalcost i Contains device cost determined by device type i and the installation location determines the deployment cost i , deployment strategy dp i The RSU in the road section s j The coverage length on is len i,j , the vehicle moves in the road network, and the sum of the time the vehicle stays within the coverage of the RSU is used as an indicator to measure the performance of the deployment strategy. The deployment strategy dp i On road sections j The total time of all vehicles covered is totaltime i,j , the MOP-based RSU deployment problem model based on cost and total coverage vehicle time is formulated as: Minimize the total deployment cost min totalcost DP , maximize the RSU coverage time of the vehicle max totaltime DP ; 2. The method according to claim 1, characterized in that The step of obtaining the road network data of the target area includes: Obtain the road network data osm file from the open street map OpenStreetMap, the osm file contains road information, building information and traffic information with data label annotation and classification, use the JOSM tool to extract the main road data from the original osm file according to the data label, remove the redundant information of the main road data, and convert it into a SUMO-readable net.xml file that retains the road topology information. The net.xml file represents the sub-lanes of the main road in the form of geographical separation, and sets the intersection of each sub-lane as a node. Use the netedit program of SUMO to merge the adjacent single geographically separated sub-lanes and nodes in the net.xml file to obtain a map file that retains the lane information and topological connection of the main road.

3. The method according to claim 1, characterized in that The NSGA-II algorithm is used to solve the MOP-based RSU deployment problem model to obtain the RSU network deployment strategy in the target area, including: Set road segments i The corresponding node is {p k ,p t }, road segment s i The total length is l k,t , assuming that three RSUs with different transmission powers are provided, represented by R1, R2, and R3 respectively, then the section s i The set of RSUs available for deployment is R = {R1, R2, R3}, and the deployment strategy dp i It includes information about the device type and deployment location, encoded using floating point numbers, with the integer part recorded as int i , the decimal part is recorded as dec i , then the road segment s i The deployment strategy on is encoded as dp i =int i +dec i The overall deployment strategy DP is encoded as a vector of floating point numbers, the integer part int i The value is any integer from 0 to 3, indicating the road section s i The type of RSU deployed on the server, int i =q, q∈{1,2,3} represents the road section s i The RSU type deployed on the q , int i =0 means that there is no RSU deployed on this road section, and the decimal part dec i ,dec i ∈[0,1) indicates the specific location of the RSU in the road section, indicating that the RSU is located in the road section s i Into node p k The distance is dec i ·l k,t at the location of For deployment strategy DP = {dp1, dp2, ..., dp N } Each element dp i , calculate the device cost devicecost respectively i and deployment cost i , add the equipment cost and installation cost to get dp i Deployment cost totalcost i , according to formula (4-17-1), the deployment costs of all elements in DP are summed up to obtain the total deployment cost of the deployment strategy DP; For road segment s i Deployment strategy dp i , which is on section s j Total time of covered vehicles totaltime i,j By RSU on section s j Coverage distance, road section s j The number of vehicles and the speed of vehicles are jointly determined. Assuming that the deployment strategy dp i The RSU in the road section s j The coverage distance on is len i,j , assuming that road segment s j The traffic density is β j vehicles / m, the vehicle speed is v j , then the number of vehicles NV within the coverage of RSU i,j It is expressed as: NV i,j =β j len i,j (4-18) For each car, assume its speed is v j , then a single vehicle on section s j When driving inside, the strategy dp is deployed i The time covered by the RSU in is expressed as: Road sections i Deployment strategy dp i On road sections j The total time for all vehicles covered is: total time i,j =NV i,j ·hour i,j (4-20) Overall deployment strategy DP = {dp1, dp2, ..., dp N The total time that the corresponding RSU covers the vehicle is: In the NSGA-II algorithm, gen represents the number of generations. max represents the maximum number of generations as the algorithm stopping condition, P and Q represent the parent population and the child population respectively. Each individual in the population represents a deployment strategy, including the number, type and deployment location of RSU. Assume that the population size is N size After merging the parent and child populations, calculate the objective function value of each individual according to formula (4-17-1) and (4-17-2), sort all individuals by non-dominated order according to Pareto rank, and add individuals to the new population P in order from small to large Pareto rank. new , if we add individuals with Pareto level i, P new The number of individuals is less than N size , and after adding individuals with Pareto level i+1, P new The number of individuals is greater than N size , then the individuals with Pareto level i+1 are sorted in descending order using the crowding operator, and individuals with low crowding are eliminated until P new The number of individuals with the remaining Pareto level i+1 added to the size , and finally obtain the individuals, that is, obtain the RSU network deployment strategy in the final target area; The definition of Pareto dominance is as follows: Definition 1, Dominance Relation: For minimizing MOP, any two decision vectors x a ,x b ; If (1) All have f i (x a )≤f i (x b ) is established; (2) So that f j (x a ) <f j (x b ) is established; Then x a Dominate x b ; If (1) So that f i (x a )≤f i (x b ) is established; (2) So that f j (x a )>f j (x b ) is established; Then x a With x b They do not dominate each other; Definition 2, Pareto optimal solution: For a decision vector, if there is no other decision vector that can dominate it, then the decision vector is called a non-dominated solution or a Pareto optimal solution; Definition 3, Pareto set: For a given set of optimal solutions, if the solutions in this set are mutually non-dominated, then this solution set is called a Pareto set; Definition 4, Pareto front: The set of images of each solution in the Pareto set in the objective function space is called the Pareto front; Definition 5, Pareto rank: In any set of feasible solutions, the Pareto rank of the non-dominated solution is defined as 1. When the non-dominated solution is deleted from the set, the Pareto rank of the new non-dominated solution generated from the remaining solutions is defined as 2. This can be deduced by analogy to obtain the Pareto rank of each solution in the feasible solution set.

Images