Equipment deployment optimization method for different road condition perception coverage problems

Abstract

The invention provides an equipment deployment optimization method for different road condition perception coverage problems, and the method comprises the following steps: analyzing different road types, and building a grid map library and a roadside equipment model library; in order to ensure the rationality and robustness of an optimization result, a two-stage method is adopted to carry out deployment optimization for a sensor and MEC key equipment; in the first stage, a sensor in an RSU is deployed and designed based on an NSGA-II algorithm, and optimal configuration of the sensor in different road scenes is ensured; in the second stage, a hybrid expert model of clustering and model predictive control MPC is used for deploying MEC equipment, and a topological relation between roadside equipment is established so as to adapt to communication requirements of an actual road scene; and based on the obtained roadside equipment deployment scheme, carrying out deployment in an actual scene. According to the method, optimization deployment can be carried out for various different scenes and sensor requirements on the basis of an optimization algorithm, and the robustness of data transmission is relatively high.

Description

Technical Field

[0001] The present invention provides a device deployment optimization method for different road condition perception coverage issues, belonging to the technical field of traffic road detection. Background Art

[0002] Currently, roadside equipment deployment is mostly based on manual experience. On the one hand, installation based on existing equipment may not meet the requirements of actual roads and fail to achieve efficient coverage of roads. On the other hand, excessive redundancy and empirical installation will lead to high deployment costs. In order to solve this problem, an automatic and highly robust deployment method is urgently needed in the current vehicle-road collaboration field.

[0003] To address the issue of perception coverage, scholars have analyzed the two-dimensional perception models of different sensor types, and further defined coverage overlap as the intersection of the fields of view of multiple visual sensors. If a point is captured by the fields of view of at least two visual sensors, it is considered to be covered. For example, methods based on the art gallery problem choose to minimize the number of sensors while avoiding blind spots.

[0004] Road Side Unit (RSU) is an important hardware for vehicle-road collaboration to carry out perception and other functions. Global RSU equipment deployment involves many considerations.

[0005] Existing technologies such as (Zhao, Y., Wang, Z., Sun, C. (2024). Deployment Optimization of Roadside Sensing Units Based on NSGA-II for Vehicle Infrastructure Cooperated Autonomous Driving. In: Jia, L., Easa, S., Qin, Y. (eds) Developments and Applications in SmartRail, Traffic, and Transportation Engineering. ICSTTE2023. Lecture Notes in Electrical Engineering,vol. 1209. Springer, Singapore. https: / / doi.org / 10.1007 / 978-981-97-3682-9_30), based on the investigation and research of current perception devices, based on the self-built sensor performance model, on the two-dimensional grid map of typical roads represented by highways and intersections, genetic algorithms and improved genetic algorithms (NSGA-Ⅱ) were used to further carry out multi-objective optimization for the three target values ​​of coverage, redundancy and cost while meeting the given coverage and redundancy. In order to further verify the model algorithm, actual road scene experiments were carried out. The design inputs were the perception sensor type, two-dimensional grid map and basic parameters of the optimization algorithm, and the design output was the Pareto solution set for the optimal deployment of perception devices. This technology has the following disadvantages:

[0006] 1. It only involves the deployment of roadside sensors, and does not consider the deployment of computing devices and communication equipment such as edge computing units (MEC) and signal repeaters. The overall architecture is not complete and does not conform to actual scenarios.

[0007] 2. The application scenarios of roadside sensor algorithms are relatively poor and cannot fully cover the needs of current high-speed and urban scenarios.

[0008] 3. The topological relationship of the sensor position and signal transmission in the optimization results is not further considered.

[0009] This paper (Tong, X.; Li, M.; Cui, Z. A. Hybrid Traffic Sensor Deployment Model with Communication Consideration for Highways. Appl. Sci. 2024, 14, 536. https: / / doi.org / 10.3390 / app14020536) proposes an optimization model for hybrid sensor deployment on highways, aiming to improve the accuracy of traffic monitoring and take into account communication requirements. The model uses 0-1 planning technology to integrate sensor deployment and server placement problems and solves them through a two-step search algorithm. This technology has the following disadvantages:

[0010] 1. The algorithm focuses on highways and lacks consideration of urban roads;

[0011] 2. The highway deployment algorithm focuses on the deployment of one-dimensional coordinates and only considers the distance between sensors. The resulting algorithm architecture is not applicable to the deployment of two-dimensional devices in urban environments.

[0012] 3. The sensors studied in this algorithm are limited to cameras, and the differences in perception performance and parameters of multiple sensors are not considered. There is a lack of consideration for hybrid sensors, and the robustness required for different engineering projects is relatively weak. Summary of the invention

[0013] The technical problems to be solved by the present invention are:

[0014] Technical Problem 1: Traditional roadside sensing equipment deployment methods result in high total costs. The present invention provides a more cost-effective deployment solution based on a multi-expert model optimization algorithm.

[0015] Technical Problem 2: The distribution range of the existing deployment method is unreasonable. The present invention implements a more reasonable device distribution strategy based on an improved genetic algorithm and a hybrid expert model.

[0016] Technical Problem 3: The traditional method has a large impact of subjective factors and poor robustness. The present invention is committed to reducing the impact of subjective factors on the deployment plan and has shown higher performance in scenarios such as roundabouts and ramps.

[0017] Technical Problem 4: Coverage is difficult to quantify. The present invention aims to propose a deployment method capable of quantifying coverage.

[0018] Technical Problem 5: Existing solutions rarely start from the performance of the sensor itself and combine it with edge computing units to carry out research on roadside equipment deployment. This invention aims to develop a deployment algorithm based on a sensor performance model under a multi-layer architecture.

[0019] The specific technical solution is:

[0020] A device deployment optimization method for different road condition perception coverage issues includes the following steps:

[0021] S1, analyze different road types and establish a raster map library and roadside equipment model library;

[0022] S2, in order to ensure the rationality and robustness of the optimization results, a two-stage approach is adopted to carry out deployment optimization for sensors and MEC key equipment;

[0023] In the first stage, the deployment design of sensors in RSU is carried out based on the NSGA-Ⅱ algorithm to ensure the optimal configuration of sensors in different road scenarios;

[0024] In the second stage, a hybrid expert model of clustering and model predictive control (MPC) is used to deploy MEC devices and establish topological relationships between roadside devices to meet the communication needs of actual road scenarios.

[0025] S3, based on the roadside equipment deployment plan, carry out deployment in actual scenarios.

[0026] Further:

[0027] In S1, it includes planning for different road types, sensor requirements, and cost budgets;

[0028] First, the road type is selected. The current deployment algorithm covers scenarios including highways, intersections, T-junctions, ramps, and roundabouts. After determining which of the above scenarios the road to be deployed belongs to, obtain a high-definition map of the current road and rasterize it in two dimensions, which will be used for calculating the Occupancy grid in the subsequent optimization process.

[0029] Then, based on current sensor needs and cost requirements, set the sensor selection range and the upper limit of the total deployment quantity.

[0030] In S2, the RSU deployment optimizer is used in the first phase;

[0031] In the first stage, based on the NSGA-Ⅱ algorithm, coverage and cost are used as optimization targets. The location coordinates, sensing radius, FOV, and deployment cost of the sensors in the RSU are deployed and optimized to obtain the optimal solution for RSU deployment.

[0032] The optimization variables are set as: the maximum sensing distance of the sensor, FOV (including horizontal and vertical), deployment cost, lens orientation, and maximum communication bandwidth;

[0033] Next, set the constraints for the optimization process:

[0034] α≥90% (1)

[0035] β ≥ 30% (2)

[0036]

[0037] Where α represents the coverage rate, β represents the redundancy rate, represents the distance between the i-th and j-th sensors;

[0038] Set the optimization variables:

[0039] Coverage:

[0040]

[0041] in, is the weight corresponding to the i-th grid in the map Q, is the coverage score corresponding to the i-th grid in the map Q;

[0042] Then, the improved NSGA-Ⅱ algorithm is applied to carry out the deployment design of roadside RSU equipment;

[0043] In the population initialization module, according to the preset population size N, several deployment schemes consisting of N RSU individuals are randomly generated. Each deployment scheme is a population with N RSUs. Each RSU has several sensors. The fitness of the deployment scheme is equal to the sum of the fitness of each RSU.

[0044] The individual fitness function value is the basis for comparison and elimination between individuals in the genetic algorithm population, and is also the most critical content of the genetic algorithm. Different individuals are sorted according to the size of their individual fitness. Those at the end of the sequence will be eliminated, thereby ensuring that the algorithm converges to the direction of high overall fitness; the fitness of the population individuals is:

[0045]

[0046] Among them B n is the individual fitness, S n is the grid score newly covered by the current individual, G n is the grid score of the new redundancy of the current individual, R n is the detection radius of the current device; the higher the fitness value of an individual, the higher the score of the newly covered grid, which means that the net increase in grid score per unit area is higher, and thus its coverage performance is relatively better; finally, the total score of the grid map is taken as the fitness score of the population, and it participates in the external competition of the population, while the internal population is based on B n To engage in intra-population competition;

[0047] The population selection part adopts a tournament-based selection method, which randomly selects two deployment plans for comparison, and the deployment plan with higher fitness is selected to generate offspring;

[0048] The process of generating offspring deployment schemes includes crossover and mutation. The crossover operation uses single-point crossover to randomly exchange the types, coordinates and parameters of some sensors in the RSUs of the two deployment schemes. The mutation operation uses single-point mutation to randomly modify the types, coordinates and parameters of the sensors in the RSUs in the deployment schemes. The RSUs with lower individual fitness are more likely to mutate.

[0049] Finally, the deployment plans are sorted according to their fitness scores, and only the top 50 deployment plans are retained. When the number of iterations meets the requirements, the optimal deployment plan is output.

[0050] In S2, the MEC deployment optimizer is used in the first phase;

[0051] The second stage is based on the hybrid expert model. First, the weighted sum of the location and communication bandwidth of the RSU is used as the topological distance, and the K-mean clustering algorithm is used to perform clustering to find the potential topological relationship between the RSUs. The maximum number of clusters is set according to the number of RSUs and then the loop begins. In the loop, the location coordinates, maximum communication bandwidth, and communication coverage radius of the MEC in the current cluster are optimized according to the constraints and initial values, the deployment cost is calculated, and the included RSUs are indexed to form a topological structure. Finally, the deployment costs of clusters of different numbers are compared, and the cluster with the lowest cost is selected as the optimal solution output. This part is based on the MPC model predictive control algorithm, and the K-mean clustering algorithm is added. By designing the optimization target, limiting the constraints, and giving the initial conditions, the deployment algorithm is optimized in a rolling manner. The algorithm inputs the optimal deployment plan obtained in the previous step, which is the coordinates and maximum communication bandwidth of each RSU.

[0052] First, the K-means clustering scheme is used to find the relationship between each RSU; K-means is to find the center of K clusters through iteration so that the sum of the distances from the data points in each cluster to the center of the cluster is minimized; V is a set of n observations, where each observation vi is a d-dimensional vector; the optimization goal of the K-means algorithm is to minimize the following objective function:

[0053]

[0054] Where: k is the number of clusters; Z i is the i-th cluster, containing all observations belonging to this cluster; u i is the center of the ith cluster, calculated as the mean of all observations in the cluster; represents the square of the Euclidean distance between the observation and the cluster center;

[0055] Based on the coordinates and parameters of the RSU obtained in the previous stage, the number of clusters is preset, and the number of clusters is traversed from 1 to the maximum number of clusters in turn to find the intrinsic topological connection between the RSUs; in each iteration, the optimal deployment solution of MEC under the current number of clusters is found; after traversing all solutions, the solution with the lowest deployment cost is selected as the optimal solution output;

[0056] Next, establish the roadside MEC library and set the MEC maximum bearer bandwidth to the following array:

[0057] W s =[w s1 ,w s2 ,...] (7)

[0058] Next, the optimal solution (x, y, bandwith) obtained in the previous stage is used as input:

[0059] P=[p1,p2,...,p n ] (8)

[0060] The deployment scheme of MEC when there are k types of RSU clusters is Zk:

[0061] Z k =[z k1 ,z k2 ,...,z kk ] (9)

[0062]

[0063] (1) Constraints

[0064] ① The total amount of data transmitted by the RSU combination in the current area must be lower than the upper limit that MEC can process:

[0065]

[0066] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes;

[0067] ②The MEC in the jth cluster should satisfy:

[0068]

[0069] When cell i is included in the coverage of the server MEC on the jth cell u, its index is set to the id of the current RSU:

[0070]

[0071] Where d is the distance between RSU and MEC server, R u is the coverage radius of the server MEC on the jth unit u;

[0072] ③ The total cost of RSUs in a single region must be lower than a fixed upper limit:

[0073]

[0074] Among them, c k is the cost of each deployment solution, W is the cost of the MEC server deployment solution;

[0075] ④The spatial deployment location of RSU should be limited to a certain area to meet the needs of actual scenarios:

[0076]

[0077] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes;

[0078] (3) Optimization goals

[0079] Find an optimal solution that satisfies the constraints while minimizing the total cost of RSU:

[0080] Q opt =min{Q(Z k )|Z k ∈Z} (16)

[0081] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u Indicates the maximum bearer bandwidth of the MEC server located at unit u, index indicates the topological relationship between the RSU at position i and the MEC server located at unit u, where index can contain one or more MEC indexes;

[0082] (4) Output the properties of the solution for each loop:

[0083] R k1 、w k1 、c k1 ;

[0084] …

[0085] R kj 、w kj 、c kj ;

[0086] …

[0087] R kk 、w kk 、c k

[0088] The first two columns are the position coordinates of the RSU in the jth class at the kth iteration, R kj is the communication coverage, w kj is the maximum carrying bandwidth, c kj is the deployment cost; for the kth cycle, Qk:

[0089]

[0090] Arrange them in order according to the size of Qk, and select the optimal solution corresponding to the smallest Qk in the iteration process as the final solution.

[0091] In S3, joint deployment verification is carried out based on the RSU optimization deployment scheme obtained in the first and second stages;

[0092] The deployment plan was selected to carry out simulation in CARLA and MATLAB software, the sensors in CARLA were selected as the sensor library to be deployed, and the map was selected and converted into a raster map to complete the input preprocessing;

[0093] Next, deployment optimization was carried out in MATLAB, first through the RSU deployment optimizer and then the MEC deployment optimizer; the final MEC deployment design was obtained.

[0094] The technical effects of the technical solution of the present invention are as follows:

[0095] 1. The present invention can carry out optimized deployment for a variety of different scenarios and sensor requirements based on the optimization algorithm. The algorithm covers five scenarios and five road levels, including highways, intersections, T-junctions, ramps and roundabouts. The sensor library includes a variety of popular products in the market, including lidar, cameras, and millimeter-wave radars;

[0096] 2. The present invention integrates the perception characteristics of multiple sensors to carry out RSU deployment, and the coverage rate of the scene can be maintained at more than 90%;

[0097] 3. While designing the physical architecture of RSU and MEC, the present invention can establish an end-to-end communication topology architecture between RSU and MEC devices, avoiding excessive load on RSU during transmission peak due to a single architecture, and has high data transmission robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 It is a schematic diagram of the overall framework of the deployment method of the present invention;

[0099] Figure 2 It is a schematic diagram of the overall architecture design of the present invention;

[0100] Figure 3 It is a simulation result diagram of RSU deployment scheme in multiple scenarios of the embodiment;

[0101] Figure 4 Schematic diagram of the crossroads equipment and architecture deployment in the embodiment

[0102] Figure 5 Graphs showing simulation results for various road scenarios of the embodiment. DETAILED DESCRIPTION

[0103] The specific technical solution of the present invention is explained in conjunction with embodiments.

[0104] The overall framework of the deployment method of the present invention is as follows Figure 1 As shown in the figure, the overall architecture design is as follows Figure 2 As shown in the figure, it is based on a two-stage hybrid expert optimization model. First, based on the current road environment and the actual needs of sensors, costs, etc., a two-dimensional grid map of the current road is generated, and the range of sensors and costs currently used are set; secondly, the above content is used as a known quantity and input into the RSU deployment optimizer to obtain the RSU deployment plan for the current road; then, the RSU deployment plan is input into the MEC deployment optimizer to obtain the MEC deployment plan for the current road; finally, a joint deployment is carried out based on the obtained RSU and MEC and the optimization target is verified.

[0105] Some of the schemes adopted a variety of research methods and technical means during the research process:

[0106] S1, analyzed different road types, established a raster map library and a roadside equipment model library;

[0107] S2, in order to ensure the rationality and robustness of the optimization results, a two-stage method is designed to carry out deployment optimization for key equipment such as sensors and MEC. In the first stage, the deployment design of sensors in RSU is carried out based on the NSGA-Ⅱ algorithm to ensure the optimal configuration of sensors in different road scenarios; in the second stage, the hybrid expert model of clustering and model predictive control (MPC) is used to deploy MEC equipment and establish the topological relationship between roadside equipment to meet the communication needs of actual road scenarios.

[0108] S3, based on the roadside equipment deployment plan, carry out deployment in actual scenarios.

[0109] Specific:

[0110] In S1, input preprocessing includes:

[0111] The design of equipment deployment plans requires planning based on different road types, sensor requirements and cost budgets.

[0112] The first step is to select the road type. The current deployment algorithm covers scenarios including highways, intersections, T-junctions, ramps, and roundabouts. After determining which of the above scenarios the road to be deployed belongs to, obtain a high-definition map of the current road through OpenStreet software or Amap and rasterize it in two dimensions, which will be used to calculate the Occupancy grid in the subsequent optimization process. The grid map can show the current features and provide the potential deployment location of the sensor and the area to be covered.

[0113] Then, based on the current sensor needs and cost requirements, set the sensor selection range and the upper limit of the total number of deployments. The most important component of the RSU is the sensor. The sensor library established by the present invention can flexibly add and shield the types of sensors according to the needs of the actual engineering scenarios, and can further find target sensors with suitable maximum perception radius and field of view (Field of View, FOV) as needed. Cost limits can be used as constraints in the optimization process to ensure that the optimization process will not produce solutions that exceed the actual engineering requirements. The types of sensors and various parameters are shown in Table 1 below:

[0114] Table 1 Equipment types and related parameters

[0115]

[0116] For example, for intersection scenes, pure vision-based algorithms have significantly higher requirements for visual cameras, fisheye cameras, etc.; for highways, millimeter-wave radar has a higher advantage in speed measurement. The cost cap limits the scale of the current deployment to ensure that the cost is within a reasonable range and can be further optimized. Based on the current road conditions, the present invention regenerates a new sensor device library for the subsequent optimizer to carry out optimization design, and the cost cap will participate in the optimization process as a constraint.

[0117] In S2, the RSU deployment optimizer used in the first phase is as follows:

[0118] The first stage is based on the NSGA-Ⅱ algorithm, with coverage and cost as the optimization objectives. The location coordinates, perception radius, FOV, and deployment cost of the sensors in the RSU are deployed and optimized to obtain the optimal solution for RSU deployment.

[0119] In view of certain parameters of the sensor, such as the number of point clouds generated per second, reflection intensity or image capture cycle, the industry often sets them according to inherent parameters, so the optimization variables of the present invention are set to: the maximum sensing distance of the sensor, FOV (including horizontal and vertical), deployment cost, lens orientation and maximum communication bandwidth, as shown in Table 2 below:

[0120] Table 2 Optimal solution data structure

[0121]

[0122] Next, set the constraints for the optimization process:

[0123] α≥90% (1)

[0124] β ≥ 30% (2)

[0125]

[0126] Where α represents the coverage rate, β represents the redundancy rate, Represents the distance between the i-th and j-th sensors

[0127] Set the optimization variables:

[0128] Coverage:

[0129]

[0130] in, is the weight corresponding to the i-th grid in the map Q, is the coverage score corresponding to the i-th grid in the map Q, where the coverage score calculation is shown in Table 3:

[0131] Table 3 Grid scoring algorithm

[0132]

[0133] Then the improved NSGA-Ⅱ algorithm is applied to carry out the deployment design of roadside RSU equipment. The corresponding road grid information has been obtained by processing the map information. The improved NSGA-Ⅱ algorithm is developed from the genetic algorithm and has population initialization, fitness function setting, population selection operation, population crossover operation, population mutation operation, and population elimination operation.

[0134] In the population initialization module, according to the preset population size N, several deployment schemes consisting of N RSU individuals with the structure in Table 3 are randomly generated. Each deployment scheme is a population with N RSUs. Each RSU has several sensors. The fitness of the deployment scheme is equal to the sum of the fitness of each RSU.

[0135] The individual fitness function value is the basis for comparison and elimination between individuals in the genetic algorithm population, and is also the most critical content of the genetic algorithm. Different individuals are sorted according to the size of their individual fitness, and those at the end of the sequence will be eliminated, thereby ensuring that the algorithm converges to a direction with a higher overall fitness. Here, the fitness of the population individuals designed by the present invention is:

[0136]

[0137] Among them B n is the individual fitness, S n is the grid score newly covered by the current individual, G n is the grid score of the new redundancy of the current individual, R n is the detection radius of the current device. The higher the fitness value of an individual, the higher the score of the newly covered grid, which means that the net increase in grid score per unit area is higher, and thus its coverage performance is relatively better. Finally, the total score of the grid map is taken as the fitness score of the population, and it participates in the external competition of the population, while the internal population is based on B n Competition within the population.

[0138] Similar to the traditional genetic algorithm, the present invention carries out the operations of selection, crossover, mutation and elimination. The population selection part adopts the selection method based on the tournament, randomly selects two deployment plans for comparison, and the deployment plan with higher fitness is selected to generate offspring.

[0139] The process of generating offspring deployment plans includes crossover and mutation. The crossover operation uses single-point crossover to randomly exchange the types, coordinates and parameters of sensors in some RSUs in the two deployment plans; the mutation operation uses single-point mutation to randomly modify the types, coordinates and parameters of sensors in the RSUs in the deployment plan, where the RSUs with lower individual fitness are more likely to mutate. The above operations can discard RSUs with poor deployment performance in the deployment plan and achieve faster convergence. In addition, the patent chooses to introduce the concept of "external population" to avoid the optimization process from falling into the local optimal trap.

[0140] Finally, according to the fitness scores of the deployment plans, the deployment plans are sorted and only the top 50 deployment plans are retained. When the number of iterations meets the requirements, the optimal deployment plan is output. The simulation results are shown in Figure 2. Figure 3 shown.

[0141] In S2, the second phase uses the MEC deployment optimizer, and the specific method is as follows:

[0142] The second stage is based on the hybrid expert model. First, the weighted sum of the location and communication bandwidth of the RSU is used as the topological distance, and the K-mean clustering algorithm is used to perform clustering to find the potential topological relationship between the RSUs. The maximum number of clusters is set according to the number of RSUs and the loop begins. In the loop, according to the constraints and initial values, the location coordinates, maximum communication bandwidth, and communication coverage radius of the MEC in the current cluster are optimized, the deployment cost is calculated, and the included RSUs are indexed to form a topological structure. Finally, the deployment costs of clusters of different numbers are compared, and the cluster with the lowest cost is selected as the optimal solution output. This part is based on the MPC model predictive control algorithm, and the K-mean clustering algorithm is added. By designing the optimization target, limiting the constraints, and giving the initial conditions, the deployment algorithm is optimized. The algorithm inputs the optimal deployment solution obtained in the previous step, mainly the coordinates and maximum communication bandwidth of each RSU.

[0143] First, the K-means clustering scheme is used to find the relationship between each RSU. The core idea of ​​K-means is to find the center of K clusters through iteration so that the sum of the distances from the data points in each cluster to the center of the cluster is minimized. Let V be a set of n observations, where each observation vi is a d-dimensional vector. The optimization goal of the K-means algorithm is to minimize the following objective function:

[0144]

[0145] Where: k is the number of clusters. Zi is the i-th cluster, containing all observations belonging to that cluster. u iis the center of the ith cluster, usually calculated as the mean of all observations in the cluster. represents the square of the Euclidean distance between the observation and the cluster center.

[0146] Here, the present invention obtains the coordinates and parameters of the RSU in the previous stage, and traverses from 1 to the maximum number of clusters in sequence by presetting the number of clusters, so as to find the intrinsic topological connection between the RSUs. In each iteration, the optimal deployment solution of MEC under the current number of clusters is found. After traversing all solutions, the present invention selects the solution with the minimum deployment cost as the optimal solution output.

[0147] Next, establish the roadside MEC library and set the MEC maximum bearer bandwidth to the following array:

[0148] W s =[w s1 ,w s2 ,...] (7)

[0149] Next, the optimal solution (x, y, bandwith) obtained in the previous stage is used as input:

[0150] P=[p1,p2,...,p n ] (8)

[0151] The deployment scheme of MEC when there are k types of RSU clusters is Zk:

[0152] Z k =[z k1 ,z k2 ,...,z kk ] (9)

[0153]

[0154] (1) Constraints

[0155] ① The total amount of data transmitted by the RSU combination in the current area must be lower than the upper limit that MEC can process:

[0156]

[0157] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes;

[0158] ②The MEC in the jth cluster should satisfy:

[0159]

[0160] When cell i is included in the coverage of the server MEC on the jth cell u, its index is set to the id of the current RSU:

[0161]

[0162] Where d is the distance between RSU and MEC server, R u is the coverage radius of the server MEC on the jth unit u;

[0163] ③ The total cost of RSUs in a single region must be lower than a fixed upper limit:

[0164]

[0165] Among them, c k is the cost of each deployment solution, W is the cost of the MEC server deployment solution;

[0166] ④The spatial deployment location of RSU should be limited to a certain area to meet the needs of actual scenarios:

[0167]

[0168] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes;

[0169] (3) Optimization goals

[0170] Find an optimal solution that satisfies the constraints while minimizing the total cost of RSU:

[0171] Q opt =min{Q(Z k )|Z k ∈Z} (16)

[0172] Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F uIt represents the maximum bearing bandwidth of the MEC server located at unit u, and index represents the topological relationship between the RSU at position i and the MEC server located at unit u. Here, index can contain one or more MEC indexes.

[0173] (4) Output the properties of the solution of each cycle as shown in Table 4:

[0174] Table 4 MEC optimal solution data structure

[0175]

[0176] The first two columns are the position coordinates of the RSU in the jth class at the kth iteration, R kj is the communication coverage, w kj is the maximum carrying bandwidth, c kj is the deployment cost. For the kth cycle, Qk is:

[0177]

[0178] Arrange them in order according to the size of Qk, and select the optimal solution corresponding to the smallest Qk in the iteration process as the final solution.

[0179] In S3, joint deployment is adopted:

[0180] Based on the RSU optimized deployment scheme obtained in the second and third parts, joint deployment verification is carried out. The deployment scheme is selected to carry out simulation in CARLA and MATLAB software, the sensors in CARLA are selected as the sensor library to be deployed, and the map is selected and converted into a raster map to complete the input preprocessing. Then, deployment optimization is carried out in MATLAB, which passes through the RSU deployment optimizer and the MEC deployment optimizer. Finally, the MEC deployment design obtained by the present invention is as follows Figure 4 shown.

[0181] The simulation results are as follows Figure 5 As shown:

[0182] The green area is the covered area, the darker the color, the higher the coverage rate, the red point is the RSU deployment point, and the yellow point is the MEC deployment location. It can be seen that under RSU deployment, the coverage rate is higher than 90%, and the distance between MEC and RSU with which it has a topological relationship is moderate, and the location is reasonable.

[0183] Table 5 Simulation results of joint deployment scheme

[0184] type Coverage cost Four-lane intersection 93.211％ 19.2189 Two-lane T-junction 93.603％ 14.5447 Two-lane roundabout 91.054％ 14.1350 Two-lane ramp 94.198％ 12.4341

[0185] The total cost is shown in Table 5.

Claims

1. A device deployment optimization method for different road condition perception coverage issues, characterized in that: The following steps are involved: S1, analyze different road types and establish a raster map library and roadside equipment model library; S2, in order to ensure the rationality and robustness of the optimization results, a two-stage approach is adopted to carry out deployment optimization for sensors and MEC key equipment; In the first stage, the deployment design of sensors in RSU is carried out based on the NSGA-Ⅱ algorithm to ensure the optimal configuration of sensors in different road scenarios; In the second stage, a hybrid expert model of clustering and model predictive control (MPC) is used to deploy MEC devices and establish topological relationships between roadside devices to meet the communication needs of actual road scenarios. S3, based on the roadside equipment deployment plan, carry out deployment in actual scenarios.

2. According to claim 1, a device deployment optimization method for different road condition perception coverage issues is characterized in that: In S1, it includes planning for different road types, sensor requirements, and cost budgets; First, the road type is selected. The current deployment algorithm covers scenarios including highways, intersections, T-junctions, ramps, and roundabouts. After determining which of the above scenarios the road to be deployed belongs to, obtain a high-definition map of the current road and rasterize it in two dimensions, which will be used for calculating the Occupancy grid in the subsequent optimization process. Then, based on current sensor needs and cost requirements, set the sensor selection range and the upper limit of the total deployment quantity.

3. The device deployment optimization method for different road condition perception coverage problems according to claim 1 is characterized in that: In S2, the RSU deployment optimizer is used in the first phase; In the first stage, based on the NSGA-Ⅱ algorithm, coverage and cost are used as optimization targets. The location coordinates, sensing radius, FOV, and deployment cost of the sensors in the RSU are deployed and optimized to obtain the optimal solution for RSU deployment. The optimization variables are set as: maximum sensing distance of the sensor, FOV, deployment cost, lens orientation, and maximum communication bandwidth; Next, set the constraints for the optimization process: α≥90％ (1) β≥30％ (2) Where α represents the coverage rate, β represents the redundancy rate, represents the distance between the i-th and j-th sensors; Set optimization variables: coverage: in, is the weight corresponding to the i-th grid in the map Q, is the coverage score corresponding to the i-th grid in the map Q; Then, the improved NSGA-Ⅱ algorithm is applied to carry out the deployment design of roadside RSU equipment; In the population initialization module, according to the preset population size N, several deployment schemes consisting of N RSU individuals are randomly generated. Each deployment scheme is a population with N RSUs. Each RSU has several sensors. The fitness of the deployment scheme is equal to the sum of the fitness of each RSU. The individual fitness function value is the basis for comparison and elimination between individuals in the genetic algorithm population, and is also the most critical content of the genetic algorithm. Different individuals are sorted according to the size of their individual fitness. Those at the end of the sequence will be eliminated, thereby ensuring that the algorithm converges to the direction of high overall fitness; the fitness of the population individuals is: Among them B n is the individual fitness, S n is the grid score newly covered by the current individual, G n is the grid score of the new redundancy of the current individual, R n is the detection radius of the current device; the higher the fitness value of an individual, the higher the score of the newly covered grid, which means that the net increase in grid score per unit area is higher, and thus its coverage performance is relatively better; finally, the total score of the grid map is taken as the fitness score of the population, and it participates in the external competition of the population, while the internal population is based on B n To engage in intra-population competition; The population selection part adopts a tournament-based selection method, which randomly selects two deployment plans for comparison, and the deployment plan with higher fitness is selected to generate offspring; The process of generating offspring deployment schemes includes crossover and mutation. The crossover operation uses single-point crossover to randomly exchange the types, coordinates and parameters of some sensors in the RSUs of the two deployment schemes. The mutation operation uses single-point mutation to randomly modify the types, coordinates and parameters of the sensors in the RSUs in the deployment schemes. The RSUs with lower individual fitness are more likely to mutate. Finally, the deployment plans are sorted according to their fitness scores, and only the top 50 deployment plans are retained. When the number of iterations meets the requirements, the optimal deployment plan is output.

4. The device deployment optimization method for different road condition perception coverage problems according to claim 3 is characterized in that: In S2, the MEC deployment optimizer is used in the first phase; The second stage is based on the hybrid expert model. First, the weighted sum of the location and communication bandwidth of the RSU is used as the topological distance, and the K-mean clustering algorithm is used to perform clustering to find the potential topological relationship between the RSUs. The maximum number of clusters is set according to the number of RSUs and then the loop begins. In the loop, the location coordinates, maximum communication bandwidth, and communication coverage radius of the MEC in the current cluster are optimized according to the constraints and initial values, the deployment cost is calculated, and the included RSUs are indexed to form a topological structure. Finally, the deployment costs of clusters of different numbers are compared, and the cluster with the lowest cost is selected as the optimal solution output. This part is based on the MPC model predictive control algorithm, and the K-mean clustering algorithm is added. By designing the optimization target, limiting the constraints, and giving the initial conditions, the deployment algorithm is optimized in a rolling manner. The algorithm inputs the optimal deployment plan obtained in the previous step, which is the coordinates and maximum communication bandwidth of each RSU. First, the K-means clustering scheme is used to find the relationship between each RSU; K-means is to find the center of K clusters through iteration so that the sum of the distances from the data points in each cluster to the center of the cluster is minimized; V is a set of n observations, where each observation vi is a d-dimensional vector; the optimization goal of the K-means algorithm is to minimize the following objective function: Where: k is the number of clusters; Z i is the i-th cluster, containing all observations belonging to this cluster; u i is the center of the ith cluster, calculated as the mean of all observations in the cluster; represents the square of the Euclidean distance between the observation and the cluster center; Based on the coordinates and parameters of the RSU obtained in the previous stage, the number of clusters is preset, and the number of clusters is traversed from 1 to the maximum number of clusters in turn to find the intrinsic topological connection between the RSUs; in each iteration, the optimal deployment solution of MEC under the current number of clusters is found; after traversing all solutions, the solution with the lowest deployment cost is selected as the optimal solution output; Next, establish the roadside MEC library and set the MEC maximum bearer bandwidth to the following array: IN s =[in s1 ,IN s2 ,...] (7) Next, the optimal solution (x, y, bandwith) obtained in the previous stage is used as input: P=[p1,p2,…,p n ] (8) The deployment scheme of MEC when there are k types of RSU clusters is Zk: WITH k =[of k1 ,With k2 ,...,With kk ] (9) (1) Constraints ① The total amount of data transmitted by the RSU combination in the current area must be lower than the upper limit that MEC can process: Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes; ②The MEC in the jth cluster should satisfy: When cell i is included in the coverage of the server MEC on the jth cell u, its index is set to the id of the current RSU: Where d is the distance between RSU and MEC server, Ru is the coverage radius of server MEC on the jth unit u; ③ The total cost of RSUs in a single region must be lower than a fixed upper limit: Among them, c k is the cost of each deployment solution, W is the cost of the MEC server deployment solution; ④The spatial deployment location of RSU should be limited to a certain area to meet the needs of actual scenarios: Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u represents the maximum bearer bandwidth of the MEC server located at unit u, index represents the topological relationship between the RSU at position i and the MEC server located at unit u, where index contains one or more MEC indexes; (3) Optimization goals Find an optimal solution that satisfies the constraints while minimizing the total cost of RSU: Q opt =min{Q(Z k )|Z k ∈Z} (16) Among them, i represents the i-th cycle, j represents the j-th cluster, and a k represents the bandwidth required by the kth RSU, Z u Indicates that there is a MEC at unit u, F u Indicates the maximum bearer bandwidth of the MEC server located at unit u, index indicates the topological relationship between the RSU at position i and the MEC server located at unit u, where index can contain one or more MEC indexes; (4) Output the properties of the solution for each loop: … … The first two columns are the position coordinates of the RSU in the jth class at the kth iteration, R kj is the communication coverage, w kj is the maximum carrying bandwidth, c kj is the deployment cost; for the kth cycle, Qk: Arrange them in order according to the size of Qk, and select the optimal solution corresponding to the smallest Qk in the iteration process as the final solution.

5. A device deployment optimization method for different road condition perception coverage issues according to any one of claims 1 to 4, characterized in that: In S3, joint deployment verification is carried out based on the RSU optimization deployment scheme obtained in the first and second stages; The deployment plan was selected to carry out simulation in CARLA and MATLAB software, the sensors in CARLA were selected as the sensor library to be deployed, and the map was selected and converted into a raster map to complete the input preprocessing; Next, deployment optimization was carried out in MATLAB, first through the RSU deployment optimizer and then the MEC deployment optimizer; the final MEC deployment design was obtained.

Images