Intelligent network connection vehicle path planning and scheduling method and system based on vehicle-road cooperation

Abstract

The invention relates to the technical field of data processing, and discloses an intelligent network connection automobile path planning and scheduling method and system based on vehicle-road cooperation. The method comprises the steps of collecting data through a road side unit and a vehicle-mounted unit to establish a vehicle-road cooperative communication network, constructing a time delay weight matrix to carry out density clustering analysis to obtain a traffic flow state, establishing an inter-vehicle communication topology based on a clustering result, forming a formation organization through a negotiation function, generating a candidate path set in combination with a communication quality evaluation model, and carrying out vehicle-road cooperative communication. And outputting a global optimal scheduling scheme by adopting an improved particle swarm optimization algorithm. The technical problems that in the prior art, a vehicle-road cooperative communication mechanism is lacked, and large-scale vehicle cooperative path planning and scheduling cannot be achieved are solved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and system for intelligent connected vehicle path planning and scheduling based on vehicle-road cooperation. Background Technology

[0002] With the rapid development of intelligent connected vehicle technology, traditional single-vehicle independent path planning methods have been widely applied in in-vehicle navigation systems. These methods are mainly based on static map data and historical traffic information, employing classic path search algorithms such as Dijkstra's algorithm and A* algorithm to calculate the shortest or fastest path for a single vehicle. Existing vehicle navigation systems can provide drivers with recommended routes from origin to destination based on road network topology, real-time traffic flow, and road condition information, and possess dynamic replanning capabilities to cope with emergencies such as traffic congestion or road closures.

[0003] However, existing technologies have significant shortcomings: First, single-vehicle navigation lacks global traffic coordination capabilities. When a large number of vehicles use the same "optimal" route simultaneously, it can easily cause local road overload and new congestion points, leading to a decline in overall traffic efficiency. Second, traditional navigation systems cannot perceive and utilize real-time status information of road infrastructure, lack effective communication mechanisms with roadside equipment, and find it difficult to obtain accurate traffic flow distribution and road capacity data. Third, existing methods mainly consider distance and time factors in the route planning process, ignoring emerging needs such as inter-vehicle communication quality, cooperative driving energy consumption, and platooning organization.

[0004] The fundamental problem with existing technologies lies in the lack of a systematic solution for vehicle-road cooperation, which fails to achieve deep integration between vehicles and road infrastructure. Specifically, existing technologies cannot establish a real-time communication network topology between vehicles and roadside units, lack traffic flow clustering analysis capabilities based on communication latency and spatial distribution, cannot organize vehicles to form cooperative platoons and conduct distributed negotiation, lack a multi-objective path optimization mechanism that incorporates communication quality constraints, and cannot handle the real-time cooperative scheduling and dynamic optimization problems of large-scale vehicles. Summary of the Invention

[0005] This application provides a method and system for intelligent connected vehicle path planning and scheduling based on vehicle-road cooperation, which solves the technical problems of lacking vehicle-road cooperative communication mechanism and being unable to realize large-scale vehicle cooperative path planning and scheduling in the prior art. By establishing a vehicle-road cooperative communication network, traffic flow clustering analysis, vehicle negotiation platooning, and improving particle swarm optimization algorithm, the cooperative scheduling efficiency and global optimization capability of intelligent connected vehicles are significantly improved.

[0006] In a first aspect, this application provides a method for intelligent connected vehicle path planning and scheduling based on vehicle-road cooperation, the method comprising:

[0007] Step S1: Collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the vehicle-road cooperative communication network topology data;

[0008] Step S2: Construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area;

[0009] Step S3: Based on the traffic flow spatiotemporal state clustering results, establish a workshop communication topology map, and conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure;

[0010] Step S4: Establish a communication quality assessment model based on the vehicle cooperative formation organization structure, combine the communication quality attenuation function with the path weight calculation, and generate a set of vehicle-road cooperative candidate paths;

[0011] Step S5: Input the set of vehicle-road cooperative candidate paths into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the global optimal scheduling scheme for vehicle-road cooperative is output.

[0012] Secondly, this application provides a vehicle-road cooperative intelligent connected vehicle path planning and scheduling system, the vehicle-road cooperative intelligent connected vehicle path planning and scheduling system comprising:

[0013] The data acquisition module is used to collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the topology data of the vehicle-road cooperative communication network.

[0014] The analysis module is used to construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area.

[0015] The topology module is used to establish a workshop communication topology map based on the traffic flow spatiotemporal state clustering results, and to conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure.

[0016] The generation module is used to establish a communication quality assessment model based on the vehicle cooperative formation organization structure, and combine the communication quality attenuation function with the path weight calculation to generate a set of vehicle-road cooperative candidate paths.

[0017] The output module is used to input the set of vehicle-road cooperative candidate paths into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the algorithm outputs the globally optimal vehicle-road cooperative scheduling scheme.

[0018] Thirdly, a vehicle-road cooperative intelligent connected vehicle path planning and scheduling device is provided, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the vehicle-road cooperative intelligent connected vehicle path planning and scheduling device to execute the above-described vehicle-road cooperative intelligent connected vehicle path planning and scheduling method.

[0019] Fourthly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the above-described intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation.

[0020] The technical solution provided in this application solves the fundamental problem of traditional single-vehicle navigation lacking global traffic information perception by collecting vehicle status information and roadside perception data through the roadside unit coverage area and combining it with vehicle individual demand parameters reported by the on-board unit. This enables vehicles to obtain complete road traffic status and infrastructure support information. By calculating the local reachability density, the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area are obtained, overcoming the technical bottleneck of existing technologies' inability to accurately identify traffic flow distribution characteristics and providing scientific data support for vehicle cooperative organization. Through vehicle negotiation functions, inter-vehicle negotiation is achieved, and the vehicle cooperative formation organization structure is obtained, solving the core problem of vehicles operating independently and lacking coordination in traditional navigation systems. This is the first time that intelligent negotiation and self-organized formation between vehicles have been realized. By combining the communication quality attenuation function with path weight calculation to generate a set of vehicle-road cooperative candidate paths, the limitation of traditional path planning that only considers distance and time is overcome, and communication quality is innovatively used as the core constraint for path selection. By introducing particle swarm optimization and distributed computing of the roadside unit guidance term to output the globally optimal vehicle-road cooperative scheduling scheme, the computational complexity problem of large-scale vehicle cooperative scheduling is solved, achieving an effective balance between global optimality and real-time performance.

[0021] The introduction of density clustering analysis algorithms enables the system to adaptively identify areas with different traffic densities, providing a precise spatial distribution basis for platooning organization and significantly improving the organizational efficiency and stability of vehicle cooperation. The design of the vehicle negotiation function considers three dimensions: speed matching degree, target similarity, and communication quality, ensuring that vehicles within the platoon have a good foundation for cooperation and effectively reducing the frequency of platoon disbandment and regrouping. The establishment of a communication quality attenuation function allows path evaluation to accurately reflect the actual needs of vehicle-road cooperative communication, avoiding cooperation failures caused by selecting paths in communication blind spots. The innovative design of the roadside unit guidance term in the improved particle swarm optimization algorithm enables the optimization process to dynamically respond to real-time perception information from roadside infrastructure, significantly improving the algorithm's adaptability and convergence performance in complex traffic environments. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of an embodiment of the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation in this application.

[0024] Figure 2 This is a schematic diagram illustrating the distribution of candidate path selection under different weight coefficient configurations in the embodiments of this application;

[0025] Figure 3 This is a schematic diagram of an embodiment of the intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation in this application.

[0026] Figure 4 This is a schematic block diagram of the intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation in an embodiment of the present invention. Detailed Implementation

[0027] This application provides a method and system for intelligent connected vehicle path planning and scheduling based on vehicle-road cooperation. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0028] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation in this application includes:

[0029] Step S1: Collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the vehicle-road cooperative communication network topology data;

[0030] Step S2: Construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area;

[0031] Step S3: Based on the traffic flow spatiotemporal state clustering results, establish a workshop communication topology map, and conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure;

[0032] Step S4: Establish a communication quality assessment model based on the vehicle cooperative formation organization structure, combine the communication quality attenuation function with the path weight calculation, and generate a set of vehicle-road cooperative candidate paths;

[0033] Step S5: Input the set of candidate paths for vehicle-road cooperation into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the globally optimal scheduling scheme for vehicle-road cooperation is output.

[0034] It is understood that the executing entity of this application can be an intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation, or it can be a terminal or a server; the specific implementation is not limited here. This application's embodiments use a server as an example for illustration.

[0035] Specifically, the roadside unit uses LiDAR to perform a 360-degree scan, detecting the three-dimensional coordinates of vehicles within its coverage area, while millimeter-wave radar measures the instantaneous speed and driving direction angle of the vehicles. The onboard unit reports the vehicle's remaining battery percentage, destination GPS coordinates, and vehicle priority level to the roadside unit via the Dedicated Short Range Communication (DSRC) protocol. During data preprocessing, a Kalman filter algorithm is used to remove noise from the raw roadside sensing data, and timestamp alignment and coordinate system transformation are performed on data acquired from different sensors. When establishing network connectivity, the RSSI (Representative Signal Strength Index) value and data transmission delay time between vehicle nodes and roadside unit nodes are calculated to form a network topology matrix. Connectivity verification checks for valid communication paths between nodes, identifies redundant links, and marks communication blind spots.

[0036] When constructing the weight matrix, communication delay data and spatial distance parameters between vehicles are extracted. The delay weight coefficient is set to 0.6, and the distance weight coefficient to 0.4. A weighted summation is used to obtain the comprehensive weight value. Density clustering employs an improved version of the DBSCAN algorithm. Based on the statistical distribution characteristics of communication delay values, the standard deviation and mean are calculated to determine a cluster radius of 50 meters, with a minimum number of vehicle nodes to be included. During neighborhood search, distance calculations are performed on each vehicle node within the roadside unit coverage area. Nodes meeting the cluster radius condition are grouped into the same neighborhood. The core point identification criterion is that the number of vehicle nodes in the neighborhood is greater than or equal to the minimum number of included points. Core points are expanded outwards to connect density-reachable boundary points, forming initial cluster groups. Local reachability density is calculated by statistically analyzing the k-th nearest neighbor distance for each vehicle. The value of k is set to 3, and the formula is: local density equals k divided by the average of the k nearest neighbor distances.

[0037] The candidate connection set for workshop communication is obtained by identifying vehicle nodes within the same cluster and calculating the relative positional relationship and Euclidean distance between vehicles. When constructing the communication topology, a communication signal strength threshold of -70 dB and a relative distance threshold of 200 meters are set; communication links are established between vehicle nodes that meet these conditions. The vehicle negotiation function calculation includes three dimensions: speed matching degree is calculated by subtracting the ratio of the vehicle speed difference to the maximum allowed speed from 1; target similarity is obtained by calculating the cosine of the angle between the vehicle's destination direction; and communication quality is determined by the ratio of the signal strength between vehicles to the maximum system signal strength. The distributed negotiation strategy requires each vehicle node to broadcast negotiation parameters containing its own speed, destination, and communication quality to its neighboring vehicles in the communication topology. After receiving response messages from neighboring vehicles, the optimal negotiation partner is selected based on the negotiation function calculation results. Formation grouping processing groups successfully negotiated vehicles into the same formation, establishing dedicated communication links within the formation to share path planning information and scheduling instructions.

[0038] When establishing the communication quality attenuation function, transmission distance and transmission delay data are extracted based on the communication link parameters of vehicles in the platoon, and the communication coverage quality of different road segments is calculated using an exponential attenuation model. Path weight calculation employs a multi-objective weighting method, with distance cost weighting coefficient set at 0.3, time cost weighting coefficient at 0.25, communication cost weighting coefficient at 0.25, and energy consumption cost weighting coefficient at 0.2. Communication cost is calculated by subtracting the communication quality value from 1 and then multiplying by the road segment length. Energy consumption cost is based on a vehicle dynamics model, considering parameters such as vehicle weight, acceleration, and drag coefficient. Path combination analysis uses an improved shortest path search algorithm, adding communication quality constraints to the traditional Dijkstra algorithm to select road segments whose communication quality meets the requirements of vehicle-road cooperation. Candidate path generation provides 3 to 5 alternative paths for each vehicle platoon, with each path labeled with an expected communication quality score and a comprehensive cost value.

[0039] During particle swarm initialization, each candidate path is encoded as a particle's position vector. The particle swarm size is set to twice the number of candidate paths, and the initial velocity parameters are generated using a random distribution. The fitness evaluation function comprehensively considers four factors: total travel time, communication latency, energy consumption, and road network congestion, calculating the fitness value of each particle through weighted summation. The individual optimal position records the path selection scheme corresponding to each particle's historical best fitness, while the global optimal position records the best scheme for the entire particle swarm. The roadside unit guidance term is an innovative feature of this algorithm, dynamically adjusting the particle's search direction based on real-time traffic flow, road conditions, and event information perceived by roadside units. The particle velocity update formula adds a roadside unit guidance term to the standard formula, with a weight coefficient set to 1.2. Distributed processing by edge computing nodes allocates optimization tasks for different regions to corresponding edge servers, accelerating algorithm convergence through parallel computing. During iterative optimization, the global optimal solution and the roadside unit guidance term are updated every 10 iterations based on the latest perception data from the roadside units. The convergence criterion is that the fitness value changes by less than 0.01 over 20 consecutive iterations or the maximum number of iterations (100) is reached. In the scheduling scheme generation phase, the position vector of the globally optimal particle is extracted, the corresponding optimal path is assigned to each vehicle formation, and detailed scheduling instructions containing path point sequences, expected transit time, suggested speed and formation coordination parameters are generated.

[0040] In one specific embodiment, step S1 includes:

[0041] The roadside unit uses lidar and millimeter-wave radar deployed in the roadside unit to scan and detect vehicles in the coverage area in real time, and obtains raw roadside perception data such as vehicle position coordinates, instantaneous speed and driving direction.

[0042] The raw roadside perception data is denoised and standardized before being combined with the remaining battery power, destination coordinates and priority parameters reported by the on-board unit through the vehicle-road communication protocol to obtain vehicle status information fusion data.

[0043] The network connection relationship between vehicle nodes and roadside unit nodes is established based on vehicle status information fusion data. The topology matrix of the vehicle-road cooperative communication network is obtained by calculating the communication signal strength and transmission delay parameters between nodes.

[0044] Connectivity verification and redundant link identification are performed on the vehicle-road cooperative communication network topology matrix. Valid communication links are selected and network coverage blind spots are marked to obtain vehicle-road cooperative communication network topology data.

[0045] Specifically, in the real-time scanning and detection process of lidar and millimeter-wave radar, lidar employs a rotating scanning method, performing 10 360-degree scans per second and emitting laser pulses with a wavelength of 905 nanometers. It calculates the precise distance between the vehicle and the roadside unit by measuring the laser reflection time, and obtains the vehicle's two-dimensional coordinate position by combining the scanning angle information. Millimeter-wave radar operates in the 77 GHz band, utilizing the Doppler effect to measure the radial velocity component of the vehicle, and obtains the instantaneous velocity value of the vehicle relative to the roadside unit through continuous wave frequency modulation technology. The driving direction detection is calculated by analyzing the vector direction of the vehicle's position change over two consecutive scanning cycles, forming a raw roadside perception data matrix containing X-coordinate, Y-coordinate, velocity vector, and direction angle. The roadside unit sorts and organizes the detected vehicle data according to timestamps, with each data entry containing basic parameters such as vehicle identifier, position coordinates, velocity information, and detection time.

[0046] The denoising and standardization preprocessing operations employ a Kalman filter algorithm to eliminate noise in the raw roadside perception data. The Kalman filter predicts the vehicle's position and velocity at the next moment based on the vehicle's motion model, compares the predicted values ​​with the actual measured values, and corrects measurement errors using a weighted averaging method. Standardization converts data from different sensors into a unified coordinate system and units of measurement. Position coordinates are converted to a relative coordinate system with the roadside unit as the origin, and velocity data is uniformly converted to meters per second. The onboard unit sends a reporting message containing the vehicle's unique identifier, current battery percentage, destination GPS coordinates, and vehicle priority value to the roadside unit via a dedicated short-range communication (DSRC) protocol, with a reporting frequency set to once per second. The data fusion process matches and associates the roadside perception data with the onboard reported data according to the vehicle identifier, generating complete vehicle status information fused data including position, speed, direction, battery level, destination, and priority. A time synchronization mechanism ensures that the timestamp error between the roadside perception data and the onboard reported data is controlled within 100 milliseconds, avoiding fusion errors caused by data inconsistency.

[0047] The network connection establishment process is based on the location information from the vehicle status information fusion data. The Euclidean distance between each vehicle node and the roadside unit node is calculated using the formula: the square root of the sum of the squares of the coordinate differences between the two points. Received Signal Strength Indicator (RSSI) technology is used for communication signal strength measurement. The roadside unit sends a test signal to the vehicle node, and the vehicle node receives the signal, measures the signal power, and feeds it back to the roadside unit. The signal strength is expressed in dBm. Transmission delay is measured using the Round Trip Time (RTT) method. The roadside unit sends a timestamped test data packet to the vehicle node, and the vehicle node immediately returns an acknowledgment message upon receiving it. The roadside unit calculates the time difference between the sending time and the receiving acknowledgment message time, divides it by 2 to obtain the one-way transmission delay. The vehicle-road cooperative communication network topology matrix is ​​stored in adjacency matrix form. The rows and columns of the matrix represent roadside unit nodes and vehicle nodes, respectively. Matrix elements contain three parameters: distance between corresponding nodes, signal strength, and transmission delay. After the matrix is ​​constructed, a valid connection relationship between nodes is determined based on communication distance thresholds and signal strength thresholds. Node pairs that meet the conditions are marked as connected in the matrix.

[0048] Connectivity verification and redundant link identification employ a depth-first search algorithm to traverse the vehicle-to-infrastructure (V2I) network topology matrix. Starting from any roadside unit node, it visits all vehicle nodes along valid connections, recording the access paths and connection states. Connectivity verification checks whether each vehicle node has at least one communication path to a roadside unit; unreachable vehicle nodes are marked as isolated nodes. Redundant link identification analyzes multi-path connections in the network topology. When multiple independent communication paths exist between two nodes, the link with the strongest signal strength or lowest transmission delay is selected as the primary connection, and other links are marked as redundant backup links. Network coverage blind spot marking identifies the coordinate range of areas lacking roadside unit coverage or with signal strength below the usable threshold by analyzing the communication coverage range of roadside units and the distribution of vehicle nodes. The V2I network topology data includes a list of valid communication links, a list of isolated nodes, a list of redundant links, and coverage blind spot coordinates, forming a complete description of the network connection status.

[0049] In one specific embodiment, step S2 includes:

[0050] Based on the topology data of the vehicle-road cooperative communication network, the communication delay and spatial distance parameters between vehicles are extracted. By setting delay weight coefficients and distance weight coefficients, a weighted calculation is performed to obtain the vehicle-road communication delay weight matrix.

[0051] Based on the vehicle-road communication delay weight matrix, the clustering radius and minimum number of included points threshold are set, and density clustering is performed on the vehicle set within the roadside unit coverage area to obtain vehicle density distribution clusters.

[0052] The distance to the kth nearest neighbor of each vehicle in the vehicle density distribution cluster is statistically calculated. By calculating the local reachability density value, dense and sparse traffic flow areas are identified, and a traffic flow density heat map is obtained.

[0053] The traffic flow density heat map is processed by calibrating the cluster center coordinates and dividing the cluster boundaries. Combined with vehicle number statistics and average communication delay calculation, the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area are obtained.

[0054] Specifically, the process of extracting vehicle-to-vehicle communication latency and spatial distance parameters involves reading the transmission delay and Euclidean distance values ​​between vehicle nodes from the vehicle-to-infrastructure (V2I) cooperative communication network topology data. The transmission delay parameter is obtained through round-trip time measurement, reflecting the real-time performance of the inter-vehicle communication link. The spatial distance parameter is calculated using the difference in vehicle position coordinates, reflecting the physical distribution relationship between vehicles. The latency weighting coefficient is set to 0.6, and the distance weighting coefficient is set to 0.4. The weighting is based on the importance of latency to path planning decisions in V2I cooperative communication. The weighted calculation process multiplies the communication latency value of each pair of vehicle nodes by the latency weighting coefficient and the spatial distance value by the distance weighting coefficient, then adds the two products to obtain the comprehensive weight value. The V2I communication latency weighting matrix is ​​stored in symmetric matrix form, with the matrix dimension equal to the total number of vehicle nodes. The matrix element value represents the comprehensive communication cost between the corresponding pair of vehicle nodes; a smaller value indicates better inter-vehicle communication conditions, making it suitable for forming a cooperative convoy.

[0055] The density clustering process determines clustering parameters based on the weight distribution characteristics in the vehicle-to-infrastructure (V2I) communication delay weight matrix. The cluster radius is calculated using the standard deviation and mean of all elements in the statistical weight matrix. When the ratio of the standard deviation to the mean exceeds 0.3, the cluster radius is set to the mean minus one standard deviation; when the ratio is below 0.3, the cluster radius is set to half the mean. The minimum number of included points threshold is set to 5 vehicle nodes. This threshold is determined based on the minimum effective size of the V2I cooperative formation; formations with fewer than 5 vehicles cannot form an effective cooperative scheduling effect. The density clustering algorithm uses the DBSCAN principle, traversing each vehicle node within the roadside unit coverage area and calculating the number of neighboring nodes within the cluster radius of that node. The criterion for determining neighboring nodes is that the weight value between two nodes is less than the cluster radius threshold. Core point identification marks vehicle nodes with a number of neighboring nodes greater than or equal to the minimum number of included points threshold as core points. Boundary point identification marks nodes within the core point cluster radius but with insufficient neighboring nodes as boundary points. Noise point identification marks isolated nodes that are neither core points nor boundary points as noise points. Vehicle density distribution clusters are generated by expanding from core points. Starting from each core point, its neighboring nodes are recursively added to the same cluster until no new density-reachable nodes can be found.

[0056] The local reachability density calculation and traffic flow region identification process calculates the distance to the k-th nearest neighbor for each vehicle node in the vehicle density distribution cluster. The value of k is set to 3, and the nearest neighbor distance is obtained by sorting the values ​​of the corresponding rows in the weight matrix; the third smallest weight value is the third nearest neighbor distance. Local reachability density calculation uses a reciprocal method, dividing the k-th value by the arithmetic mean of the k nearest neighbor distances to obtain the local density value of the node. A higher density value indicates a denser vehicle distribution in the area. Dense traffic flow regions are identified by comparing the local reachability density values ​​of each node. Nodes with density values ​​higher than 1.5 times the overall average density and their neighbors are marked as dense regions. Sparse traffic flow regions are identified by marking nodes with density values ​​lower than 0.5 times the overall average density and their neighbors as sparse regions. The traffic flow density heat map is generated using a gridding method, dividing the roadside unit coverage area into 10-meter by 10-meter grid units. The density value of each grid is calculated by the average local reachability density of all vehicle nodes within that grid. The density value is represented by color depth, with dark areas corresponding to high-density traffic flow and light areas corresponding to low-density traffic flow.

[0057] The cluster center identification and boundary delineation process calculates the geometric center coordinates of each vehicle density distribution cluster in the traffic flow density heat map. The center coordinates are obtained by weighted averaging of the coordinates of all vehicle nodes within the cluster, with the weights being the local reachability density values ​​of each node. The cluster boundary delineation uses the convex hull algorithm to find the smallest convex polygon that can enclose all vehicle nodes within the cluster; the coordinates of the convex hull vertices are the key points of the cluster boundary. Vehicle quantity statistics record the total number of vehicle nodes in each cluster and the distribution of different vehicle types, including the proportion of different vehicle types such as cars, trucks, and buses. Average communication delay is calculated by extracting the communication delay values ​​between all pairs of vehicle nodes within the cluster and calculating the arithmetic mean to obtain the communication performance index of that cluster. The traffic flow spatiotemporal state clustering results for the roadside unit coverage area include complete descriptive information such as cluster identifier, center coordinates, boundary range, vehicle quantity, average communication delay, and density level, forming a quantitative expression of traffic flow distribution.

[0058] In one specific embodiment, based on the vehicle-road communication delay weight matrix, a clustering radius and a minimum number of included points threshold are set, and density clustering is performed on the vehicle set within the roadside unit coverage area to obtain vehicle density distribution clusters, including:

[0059] Based on the statistical analysis of the distribution characteristics of communication delay values ​​in the vehicle-road communication delay weight matrix, the clustering radius parameter of 50 meters was determined by calculating the standard deviation and mean of the communication delay, and the density clustering radius threshold was obtained.

[0060] Based on the density clustering radius threshold and the characteristics of vehicle communication coverage, the minimum number of points to be included is set to 5 vehicle nodes. Neighborhood search processing is performed on all vehicle nodes in the roadside unit coverage area to obtain the vehicle node neighborhood relationship matrix.

[0061] Vehicle nodes that meet the density condition in the vehicle node neighborhood relationship matrix are marked as core points. Clustering expansion is performed by connecting the density-reachable boundary points outward from the core points to obtain the initial vehicle density cluster group.

[0062] Noise points and outliers are removed from the initial vehicle density clusters. By merging adjacent clusters and relabeling the cluster numbers, vehicle density distribution clusters are obtained.

[0063] Specifically, the statistical analysis process for the distribution characteristics of communication delay values ​​extracts the communication delay values ​​between all vehicle node pairs from the vehicle-to-infrastructure communication delay weight matrix, forming a one-dimensional data sequence for statistical calculation. The standard deviation calculation process first calculates the arithmetic mean of all delay values, then calculates the squared difference between each delay value and the mean, sums all the squared differences, divides by the total number of data points minus 1, and finally takes the square root of the result to obtain the standard deviation. The mean is calculated by adding all communication delay values ​​and dividing by the total number of data points, reflecting the overall level of communication delay. The cluster radius parameter is determined using the three sigma principle in statistics. When the ratio of the standard deviation to the mean is less than 0.4, it indicates that the data distribution is relatively concentrated, and the cluster radius is set to the mean minus the standard deviation. When the ratio is greater than or equal to 0.4, it indicates that the data distribution is relatively dispersed, and the cluster radius is set to the mean multiplied by a fixed coefficient of 0.8. The density cluster radius threshold is calculated and determined to be 50 meters. This threshold reflects the maximum physical distance for effective communication coordination between vehicles; vehicle nodes exceeding this distance cannot establish stable cooperative relationships.

[0064] The neighborhood search and relation matrix construction process is based on a density clustering radius threshold of 50 meters and the characteristics of vehicle communication coverage. The minimum number of included points is set to 5 vehicle nodes, a value determined by the minimum effective size of vehicle-road cooperative formations, as formations with fewer than 5 vehicles are unlikely to produce significant cooperative scheduling effects. The neighborhood search process traverses each vehicle node within the roadside unit's coverage area, searching for all neighboring vehicle nodes within the clustering radius, centered on that node. The search process calculates the Euclidean distance between the target node and candidate neighbor nodes; when the distance is less than or equal to the clustering radius threshold, the candidate node is marked as a neighbor of the target node. The vehicle node neighborhood relation matrix is ​​stored in adjacency list form. The row index of the matrix represents the vehicle node identifier, and the column element of the corresponding row records the identifiers and distance values ​​of all neighbor nodes of that node. After the relation matrix is ​​constructed, the number of neighbors for each vehicle node is counted. Nodes with a number of neighbors greater than or equal to the minimum number of included points are qualified to become core points; nodes with a number of neighbors less than the threshold are treated as boundary points or noise points.

[0065] The density condition determination in the core point labeling and cluster expansion process involves comparing the number of neighboring vehicle nodes with a minimum threshold of contained points. Vehicle nodes meeting the condition are labeled as core points, which have the ability to expand outwards into new clusters. The cluster expansion process starts from each core point and uses a depth-first search algorithm to expand outwards, adding all neighboring nodes of the core point to the current cluster group. Then, it checks whether newly added neighboring nodes are also core points. If they are, the expansion continues; otherwise, they are marked as boundary points and the expansion stops. Density reachability refers to all nodes reachable from a core point through a series of connecting paths, including core points and boundary points, but excluding noise points. The expansion process is recursively performed, ensuring that all nodes within a cluster group have a direct or indirect density reachability relationship with at least one core point. Initial vehicle density clusters are formed through the expansion process. Each cluster group contains one or more core points and all their density-reachable boundary points. There are no density reachability relationships between different cluster groups, ensuring the independence of the clustering results.

[0066] The noise point removal and cluster group optimization process involves several steps. Noise point identification is performed by checking if a vehicle node belongs to any initial cluster group. Isolated nodes that are neither core points nor boundary points are marked as noise points. Isolated point processing removes noise points from the clustering results to prevent isolated vehicles from interfering with subsequent path planning and scheduling decisions. Isolated vehicles will adopt independent path planning strategies and will not participate in collaborative scheduling. Adjacent cluster merging is determined by calculating the minimum distance between the boundary points of different cluster groups. When the minimum distance is less than 1.2 times the cluster radius threshold, the two cluster groups are considered spatially adjacent and eligible for merging. Merging processing combines adjacent small-scale clusters into large-scale clusters to avoid reduced collaborative efficiency due to over-segmentation. The merged clusters have recalculated center coordinates and boundary ranges. Cluster numbering is relabeled using a continuous integer sequence, starting from 1, assigning a unique identifier to each final cluster group. The numbering order is arranged from most to least numerous vehicles within the cluster group, facilitating the subsequent priority processing of large-scale vehicle platoons. The vehicle density distribution cluster includes a cluster identifier, a list of core points, a list of boundary points, coordinates of the cluster center, coordinates of the boundary range, and the total number of vehicles.

[0067] In one specific embodiment, step S3 includes:

[0068] Based on the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area, vehicle clustering distribution information is extracted. By identifying vehicle nodes within the clusters and calculating the relative positional relationships between vehicles, a set of candidate connections for vehicle-to-vehicle communication is obtained.

[0069] Based on the communication signal strength and relative distance parameters of vehicle nodes in the candidate connection set of workshop communication, the workshop communication topology map is obtained by setting a communication signal strength threshold of -70 dB and a relative distance threshold of 200 meters for screening.

[0070] The vehicle node speed difference, destination direction angle, and communication quality parameters in the workshop communication topology are normalized. The vehicle negotiation score matrix is ​​obtained by calculating the speed matching degree as 1 minus the ratio of the speed difference to the maximum speed, the target similarity as the cosine of the destination direction angle, and the communication quality as the ratio of the signal strength to the maximum strength.

[0071] Vehicle nodes that meet the negotiation conditions in the vehicle negotiation scoring matrix are grouped into formations. By establishing communication links and a shared scheduling information mechanism for vehicles within the formation, a vehicle collaborative formation organizational structure is obtained.

[0072] Specifically, basic data such as cluster identifiers, core point lists, boundary point lists, and cluster center coordinates are retrieved from the spatiotemporal state clustering results of traffic flow within the roadside unit coverage area. Vehicle node identification within each cluster involves traversing the vehicle list of each cluster to extract key parameters such as unique vehicle identifiers, current position coordinates, velocity vectors, and driving direction angles. The calculation of relative positional relationships between vehicles employs vector geometry methods, using the first vehicle as the reference origin to calculate the position offset vectors of other vehicles within the same cluster relative to the reference vehicle. These offset vectors include X-axis and Y-axis components. Relative distance calculation uses the Euclidean distance formula, taking the square root of the sum of the squares of the differences in the position coordinates of two vehicles to obtain the straight-line distance between them. Relative azimuth angle calculation uses the arctangent function, calculating the azimuth angle between vehicles based on the X and Y components of the position offset vectors, with an angle range from 0 to 360 degrees. The vehicle-to-vehicle communication candidate connection set contains connection information for all vehicle node pairs within the cluster. Each connection record includes detailed parameters such as the starting vehicle identifier, target vehicle identifier, relative distance, relative azimuth angle, and connection establishment timestamp.

[0073] The communication signal strength and distance parameter screening process is based on vehicle node pairs in the candidate connection set for workshop communication, extracting two key screening parameters: communication signal strength and relative distance. Communication signal strength is measured using Received Signal Strength Indicator (RSSI) technology. The initiating vehicle sends a test signal of standard power to the target vehicle. After receiving the signal, the target vehicle measures the received power and feeds it back to the initiating vehicle. Signal strength is expressed in decibels (dBm), with values ​​closer to zero indicating stronger signals. The communication signal strength threshold is set to -70 dB, determined based on the minimum reliability requirements of vehicle-to-infrastructure (V2I) communication. Vehicle connections with signal strength below this threshold cannot guarantee data transmission stability and real-time performance. The relative distance threshold is set to 200 meters, determined based on the effective coverage of the Dedicated Short Range Communication (DSRC) protocol. Vehicle connections exceeding 200 meters pose a risk of communication interruption. The screening process uses a dual-condition judgment: vehicle connections that simultaneously meet the conditions of a signal strength greater than or equal to -70 dB and a relative distance less than or equal to 200 meters are retained; connections that do not meet either condition are removed from the candidate set. The workshop communication topology is represented by an undirected graph structure, with vehicle nodes as vertices and vehicle connections that meet the screening criteria as edges. Each edge is labeled with signal strength and distance values.

[0074] The normalization and negotiation scoring matrix calculation process extracts three negotiation evaluation parameters from the workshop communication topology: speed difference, destination direction angle, and communication quality. The speed difference is calculated as the absolute difference between the current speeds of two vehicles, reflecting their motion synchronization; a smaller speed difference indicates a better fit for coordinated formation. The destination direction angle is calculated as the difference between the azimuth angles of the destinations of two vehicles, ranging from 0 to 180 degrees; a smaller angle indicates closer proximity of the vehicles' destinations. Communication quality is represented by the ratio of signal strength to the maximum signal strength recorded by the system, typically -30 dB; a ratio closer to 1 indicates better communication quality. Normalization converts the three parameters to values ​​between 0 and 1. The speed matching degree is calculated by subtracting the ratio of the speed difference to the system's maximum permissible speed difference from 1. A matching degree of 1 is achieved when the speed difference is zero, and 0 when the speed difference equals the maximum permissible difference. Target similarity is calculated using the cosine of the angle between the destination and the target direction. A cosine value of 1 indicates perfect similarity when the angle is 0 degrees, while a cosine value of 0 indicates complete dissimilarity when the angle is 90 degrees. Communication quality normalization is directly applied using the ratio of signal strength to maximum signal strength, with a value naturally ranging from 0 to 1. The vehicle negotiation scoring matrix is ​​stored in a square matrix format, with row and column indices corresponding to vehicle node identifiers. Matrix elements contain a comprehensive score for three components: speed matching degree, target similarity, and communication quality.

[0075] In the negotiation condition determination and formation organization structure construction process, vehicle node pairs in the vehicle negotiation scoring matrix that meet the negotiation conditions must simultaneously satisfy three criteria: speed matching degree greater than 0.7, target similarity greater than 0.6, and communication quality greater than 0.5. A greedy algorithm strategy is used for formation grouping, prioritizing the vehicle node pair with the highest comprehensive score as the formation core, and then gradually adding other vehicle nodes that meet the negotiation conditions with the core node. The formation size is controlled between 3 and 8 vehicles; formations with fewer than 3 vehicles have limited collaborative effects, while formations with more than 8 vehicles have excessively high communication complexity and are difficult to manage. A star topology is used to establish vehicle communication links within the formation. The vehicle with the highest comprehensive score is selected as the formation leader, and other vehicles establish direct communication links with the formation leader, who is responsible for coordinating path planning and scheduling decisions within the formation. The shared scheduling information mechanism includes key information such as the formation member list, formation leader identifier, formation target area, estimated arrival time, and cooperative driving speed. The formation leader periodically broadcasts updated information to member vehicles, and member vehicles report their status changes to the formation leader. The vehicle collaborative formation organizational structure records the formation identifier, formation leader vehicle, member vehicle list, formation communication topology, collaborative parameters, and establishment timestamp.

[0076] In one specific embodiment, step S4 includes:

[0077] Based on the platoon communication link parameters in the vehicle cooperative platooning organizational structure, a communication quality attenuation function is established by calculating the transmission distance and transmission delay, and the communication coverage quality of road segments in the road network is evaluated to obtain road segment communication quality evaluation data.

[0078] Based on road segment communication quality assessment data combined with road segment distance, travel time and energy consumption parameters, a comprehensive weight assessment result of the path is obtained by weighting the calculation by setting weight coefficients for distance, time, communication and energy consumption.

[0079] The road segments whose communication quality meets the requirements of vehicle-road cooperation are analyzed by combining the comprehensive weight evaluation results of the paths. Multiple candidate paths are generated for each vehicle platoon and communication quality and cost parameters are labeled to obtain a set of candidate paths for vehicle-road cooperation.

[0080] Specifically, the process of calculating transmission distance and delay, and establishing the communication quality attenuation function, extracts platoon communication link parameters from the vehicle collaborative formation organizational structure, including key data such as signal strength, transmission delay, and connection stability between the formation leader and member vehicles. Transmission distance is calculated based on the real-time position coordinates of vehicles within the platoon, using the Euclidean distance formula between two points to calculate the straight-line distance between the formation leader and each member vehicle. The distance value reflects the physical path length of signal propagation. Transmission delay is obtained through round-trip time (RTT) measurement. The formation leader sends timestamp data packets to member vehicles, and the member vehicles immediately return acknowledgment messages upon receiving them. The formation leader records the time difference between sending and receiving and divides it by 2 to obtain the one-way transmission delay. The communication quality attenuation function is established using an exponential attenuation model. The initial communication quality is set to 1.0 to represent the ideal state. The attenuation function decreases exponentially with increasing transmission distance and delay, with the distance attenuation coefficient set to 0.02 per meter and the delay attenuation coefficient set to 0.05 per millisecond. The road network segment communication coverage quality assessment divides each road into 100-meter-long segment units, calculates the expected communication quality value when vehicle convoys pass through the segment, and considers factors such as the coverage range of roadside units within the segment, the impact of signal obstructions, and vehicle density. The road segment communication quality assessment data includes detailed information such as segment identifiers, start coordinates, end coordinates, expected communication quality, signal coverage range, and communication blind spot markers.

[0081] The multi-parameter weighting calculation and path comprehensive evaluation process is based on road segment communication quality assessment data, combined with road segment distance, travel time, and energy consumption parameters to conduct multi-dimensional path evaluation. The road segment distance parameter directly uses the physical length of the road segment, calculated from the distance between the starting and ending coordinates; a longer distance indicates a higher path cost. The travel time parameter is calculated based on the road segment speed limit and expected traffic density; travel time equals the road segment distance divided by the average vehicle speed on that road segment, with speed limits and traffic congestion affecting the calculation. The energy consumption parameter is calculated based on a vehicle dynamics model, considering factors such as vehicle mass, road gradient, driving resistance, and acceleration mode. Energy consumption calculation includes both cruising energy consumption and additional energy consumption during acceleration and deceleration. The weighting coefficients reflect the priority ranking of the vehicle-road cooperative scenario: a distance weighting coefficient of 0.3 reflects the basic path cost; a time weighting coefficient of 0.25 reflects traffic efficiency requirements; a communication weighting coefficient of 0.25 highlights the core characteristics of vehicle-road cooperation; and an energy consumption weighting coefficient of 0.2 takes into account environmental protection and energy conservation needs. The weighted calculation process multiplies each of the four parameters by its corresponding weight coefficient and then sums the results to obtain the overall path weight evaluation result. The smaller the value, the better the overall path performance. The overall path weight evaluation result is stored in matrix form, with row indices corresponding to the starting road nodes, column indices corresponding to the target road nodes, and matrix elements recording the overall weight values ​​of the corresponding paths.

[0082] In the path combination analysis and candidate path set generation process, the road segments whose communication quality meets the requirements of vehicle-road cooperation are selected using a threshold filtering method. The communication quality threshold is set at 0.6; road segments below this threshold cannot guarantee the cooperative communication needs of the vehicle platoon. The path combination analysis employs an improved shortest path algorithm, adding communication quality constraints to the traditional Dijkstra algorithm. During the search process, only road segments that meet the communication quality requirements are considered as valid path nodes. The path search starts from the current position of the vehicle platoon and ends at the platoon's common target area, using dynamic programming to find all path combinations that meet the constraints. Multiple candidate paths are generated for each vehicle platoon, with 3 to 5 different candidate paths. The path generation strategy includes different preferences such as shortest distance priority, shortest time priority, lowest energy consumption priority, and best communication quality priority. The candidate path labeling process calculates a communication quality score and a comprehensive cost parameter for each path. The communication quality score is the weighted average of the communication quality of all road segments on the path, with the weight being the length ratio of each road segment. The comprehensive cost parameter directly uses the path comprehensive weight evaluation result. The candidate path set for vehicle-road cooperative systems includes path identifier, path node sequence, total distance, estimated travel time, estimated energy consumption, communication quality score, overall cost, and generation timestamp.

[0083] Figure 2 This is a schematic diagram of the candidate path selection distribution under different weight coefficient configurations in the embodiments of this application.

[0084] This paper illustrates the frequency distribution of candidate path selection for different types of routes under various weighting coefficient configurations during vehicle-road cooperative path planning. Scheme A, employing a time-priority configuration, achieves a shortest path selection frequency of 65%, reflecting an emphasis on traffic efficiency. Scheme B, prioritizing communication, increases the optimal communication path selection frequency to 55%, ensuring the quality of vehicle-road cooperative communication. Scheme C, prioritizing energy consumption, achieves a minimum energy consumption path selection frequency of 45%, meeting the requirements of green travel. Scheme D, employing a balanced configuration, shows a relatively balanced selection of the three path types, comprehensively considering the weight allocation across four dimensions: distance, time, communication, and energy consumption. This distribution chart verifies the effectiveness of the comprehensive path weighting evaluation algorithm in this invention. Different weighting coefficient settings can flexibly adapt to different vehicle-road cooperative application scenarios, providing personalized path selection strategies for intelligent connected vehicle platooning.

[0085] In one specific embodiment, step S5 includes:

[0086] Each path in the candidate path set for vehicle-road cooperative communication is encoded as a particle position vector for particle swarm initialization. By setting the particle swarm size and initial velocity parameters, the initial particle swarm of the improved vehicle-road cooperative particle swarm optimization algorithm is obtained.

[0087] The weighted fitness values ​​of total travel time, communication delay, energy consumption, and road network congestion penalty are calculated based on the path selection scheme of each particle in the initial particle swarm. The particle swarm fitness evaluation results are obtained by recording the individual optimal position and the global optimal position.

[0088] Based on the particle swarm fitness evaluation results and the real-time sensing data of the roadside unit, the roadside unit guidance terms are updated. Iterative calculations and distributed processing of edge computing nodes are performed through particle velocity and position update formulas to obtain optimized iterative convergence results.

[0089] The optimization iteration convergence results are processed to extract the global optimal solution and generate a scheduling scheme. By assigning the optimal path and coordination parameters to each vehicle platoon, the global optimal scheduling scheme for vehicle-road cooperation is obtained.

[0090] Specifically, each path in the vehicle-road cooperative candidate path set is encoded as a particle position vector. The encoding method uses an integer sequence, the length of which is equal to the total number of vehicle platoons. The value of each element in the sequence corresponds to the candidate path number selected by that platoon. The particle position vector encoding process traverses the candidate path set, assigning a unique integer identifier to the candidate paths of each vehicle platoon. The three candidate paths of the first platoon are numbered 1, 2, and 3; the candidate paths of the second platoon are numbered 4, 5, 6, and 7, and so on, forming a continuous numbering sequence. The particle swarm size is set based on the total number of candidate paths and computational resource constraints, typically set to 1.5 times the total number of candidate paths, ensuring sufficient search space coverage and keeping computational complexity within a reasonable range. The initial velocity parameters are generated using a random distribution. Each component of the velocity vector takes a random number between -3 and +3, with negative values ​​indicating movement to a path with a smaller number and positive values ​​indicating movement to a path with a larger number. The initial particle swarm of the improved vehicle-road cooperative particle swarm optimization algorithm includes basic attributes such as particle identifier, position vector, velocity vector, individual optimal position, individual optimal fitness value, and particle activity state. After the particle swarm is initialized, each particle represents a complete vehicle platoon path allocation scheme, and the particle swarm covers the path combination space of all platoons.

[0091] The fitness calculation and particle swarm evaluation process is based on the path selection scheme of each particle in the initial particle swarm, calculating a weighted fitness value that includes four objectives: total travel time, communication delay, energy consumption, and network congestion penalty. Total travel time is calculated by summing the estimated travel times of each formation's path corresponding to the particle's position vector. Travel time is calculated based on path length and expected speed, taking into account the impact of traffic flow on speed. Communication delay is calculated by analyzing the communication quality of each formation's chosen path. Communication delay equals 1 minus the average path communication quality value multiplied by the baseline delay time; lower communication quality results in a larger delay penalty. Energy consumption is calculated based on the distance, gradient, and expected acceleration mode of each formation's path, using a vehicle dynamics model to calculate the expected energy consumption of each formation, and then summing the energy consumption values ​​of all formations. Network congestion penalty is calculated by analyzing the overlap of different formations' path selections. When multiple formations choose the same road segment, a congestion penalty is incurred, and the penalty value is proportional to the square of the number of overlapping formations. The weighted fitness value is calculated by multiplying each of the four objectives by a weighted coefficient and then summing the results. The total travel time has a weight of 0.3, communication delay has a weight of 0.3, energy consumption has a weight of 0.2, and congestion penalty has a weight of 0.2. The individual optimal position records the position vector corresponding to the best fitness value in each particle's historical iterations, while the global optimal position records the position vector corresponding to the best fitness value of the entire particle swarm. The particle swarm fitness evaluation result includes complete information such as the current fitness value of each particle, the individual optimal record, the global optimal record, and the statistical distribution of fitness values.

[0092] The roadside unit guidance term update and iterative calculation process updates the unique roadside unit guidance term parameters in the improved algorithm based on the particle swarm fitness evaluation results and real-time roadside unit perception data. The real-time roadside unit perception data includes traffic flow, average speed, congestion level, and information on sudden events for each road segment at the current moment. The data is updated every 10 seconds, reflecting the dynamic changes in traffic conditions. The roadside unit guidance term calculation re-evaluates the adaptability of each candidate path based on the real-time perception data. When a new congestion occurs on a road segment traversed by a path, the guidance term value for that path decreases, and vice versa. The particle velocity update formula adds the roadside unit guidance term to the standard particle swarm algorithm. The new velocity equals the inertia weight multiplied by the current velocity, plus the individual learning factor multiplied by a random number multiplied by the difference between the individual's optimal position and the current position, plus the global learning factor multiplied by a random number multiplied by the difference between the global optimal position and the current position, plus the roadside guidance factor multiplied by a random number multiplied by the difference between the roadside guidance term and the current position. The inertia weight is set to 0.7 to control the degree to which particles maintain their current motion trend. The individual learning factor is set to 1.5 to control the intensity of particle movement towards the individual optimal position. The global learning factor is set to 1.5 to control the intensity of particle movement towards the global optimal position. The roadside guidance factor is set to 1.2 to control the degree of influence of real-time information from roadside units on particle motion. Distributed processing at edge computing nodes allocates particles from different regions to corresponding edge servers for parallel computation. Each edge node is responsible for fitness evaluation and position updates within a specific region. The computation results are aggregated at the central node for updating the global optimal solution. The convergence of the optimization iteration is determined by monitoring the change in the global optimal fitness value during consecutive iterations. The algorithm is considered converged when the fitness improvement after 20 consecutive iterations is less than 0.01.

[0093] The global optimal solution extraction and scheduling scheme generation process extracts the global optimal solution from the convergence results of the optimization iteration, decoding the optimal path selection for each vehicle formation from the final global optimal position vector. The decoding process maps each element of the position vector back to the candidate path number for the corresponding formation, and extracts the specific path node sequence, estimated travel time, and coordination parameters from the candidate path set based on the number. The scheduling scheme generation process assigns detailed driving instructions to each vehicle formation, including the coordinates of the starting point, the sequence of node coordinates along the route, the coordinates of the destination, the suggested speed, the expected arrival time, and the formation coordination requirements. Optimal path allocation considers the coordination relationships between formations. When multiple formations' paths intersect, the scheduling scheme specifies the passage order and safety distance requirements to avoid conflicts between formations. Coordination parameters include key information such as formation leader responsibility allocation, member vehicle following distance, communication frequency settings, and emergency response plans. The vehicle-road cooperative global optimal scheduling scheme integrates the path allocation and coordination parameters of all formations to form a complete multi-formation cooperative driving plan. The scheduling scheme also includes management information such as generation time, validity period, and execution priority.

[0094] The above describes the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation in the embodiments of this application. The following describes the intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation in the embodiments of this application. Please refer to [link / reference]. Figure 3 One embodiment of the intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation in this application includes:

[0095] The data acquisition module is used to collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the topology data of the vehicle-road cooperative communication network.

[0096] The analysis module is used to construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area.

[0097] The topology module is used to establish a workshop communication topology map based on the traffic flow spatiotemporal state clustering results, and to conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure.

[0098] The generation module is used to establish a communication quality assessment model based on the vehicle cooperative formation organization structure, and combine the communication quality attenuation function with the path weight calculation to generate a set of vehicle-road cooperative candidate paths.

[0099] The output module is used to input the set of vehicle-road cooperative candidate paths into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the algorithm outputs the globally optimal vehicle-road cooperative scheduling scheme.

[0100] above Figure 3 The intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0101] Reference Figure 4 This invention also provides a vehicle-road cooperative intelligent connected vehicle path planning and scheduling device. This device can be a server, and its internal structure can be as follows: Figure 4As shown, the intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor in this computer design provides computing and control capabilities. The memory of the intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation is used to store the data corresponding to this embodiment. The network interface of the intelligent connected vehicle path planning and scheduling device based on vehicle-road cooperation is used for communication with external terminals via network connection. When the computer program is executed by the processor, it implements the above-described method.

[0102] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the intelligent connected vehicle path planning and scheduling equipment based on vehicle-road cooperation applied thereto.

[0103] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation.

[0104] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0105] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a vehicle-road cooperative intelligent connected vehicle path planning and scheduling device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent connected vehicle path planning and scheduling based on vehicle-road cooperation, characterized in that, The method includes: Step S1: Collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the vehicle-road cooperative communication network topology data; Step S2: Construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area; Step S3: Based on the traffic flow spatiotemporal state clustering results, establish a workshop communication topology map, and conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure; Step S4: Establish a communication quality assessment model based on the vehicle cooperative formation organization structure, combine the communication quality attenuation function with the path weight calculation, and generate a set of vehicle-road cooperative candidate paths; Step S5: Input the set of vehicle-road cooperative candidate paths into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the global optimal scheduling scheme for vehicle-road cooperative is output.

2. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 1, characterized in that, Step S1 includes: The roadside unit uses lidar and millimeter-wave radar deployed in the roadside unit to scan and detect vehicles in the coverage area in real time, and obtains raw roadside perception data such as vehicle position coordinates, instantaneous speed and driving direction. The raw roadside sensing data is denoised and standardized before being combined with the remaining battery power, destination coordinates and priority parameters reported by the vehicle unit through the vehicle-road communication protocol to obtain vehicle status information fusion data. Based on the vehicle status information fusion data, a network connection relationship between vehicle nodes and roadside unit nodes is established. By calculating the communication signal strength and transmission delay parameters between nodes, the topology matrix of the vehicle-road cooperative communication network is obtained. The connectivity verification and redundant link identification processes are performed on the vehicle-road cooperative communication network topology matrix. Valid communication links are selected and network coverage blind spots are marked to obtain vehicle-road cooperative communication network topology data.

3. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 1, characterized in that, Step S2 includes: Based on the vehicle-road cooperative communication network topology data, the communication delay and spatial distance parameters between vehicles are extracted. By setting delay weight coefficients and distance weight coefficients, a weighted calculation is performed to obtain the vehicle-road communication delay weight matrix. Based on the vehicle-road communication delay weight matrix, a clustering radius and a minimum number of included points threshold are set, and density clustering is performed on the vehicle set within the roadside unit coverage area to obtain vehicle density distribution clusters. The k-th nearest neighbor distance of each vehicle in the vehicle density distribution cluster is statistically calculated. By calculating the local reachability density value, dense and sparse traffic flow areas are identified, and a traffic flow density heat map is obtained. The traffic flow density heat map is processed by calibrating the cluster center coordinates and dividing the cluster boundaries. Combined with vehicle number statistics and average communication delay calculation, the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area are obtained.

4. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 3, characterized in that, The method involves setting a clustering radius and a minimum number of included points based on the vehicle-to-infrastructure communication delay weight matrix, and performing density clustering on the vehicle set within the roadside unit coverage area to obtain vehicle density distribution clusters, including: Based on the distribution characteristics of communication delay values ​​in the vehicle-road communication delay weight matrix, statistical analysis is performed. By calculating the standard deviation and mean of the communication delay, the clustering radius parameter is determined to be 50 meters, and the density clustering radius threshold is obtained. Based on the density clustering radius threshold and the vehicle communication coverage characteristics, the minimum number of points to be included is set to 5 vehicle nodes. Neighborhood search processing is performed on all vehicle nodes within the roadside unit coverage area to obtain the vehicle node neighborhood relationship matrix. Vehicle nodes that meet the density condition in the vehicle node neighborhood relationship matrix are marked as core points. Clustering expansion is performed by connecting the density-reachable boundary points outward from the core points to obtain the initial vehicle density clustering group. The initial vehicle density clusters are processed by removing noise points and handling outliers. By merging adjacent clusters and relabeling the cluster numbers, vehicle density distribution clusters are obtained.

5. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 1, characterized in that, Step S3 includes: Based on the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area, vehicle clustering distribution information is extracted. By identifying vehicle nodes within the clusters and calculating the relative positional relationships between vehicles, a set of candidate connections for vehicle-to-vehicle communication is obtained. Based on the communication signal strength and relative distance parameters of vehicle nodes in the candidate connection set of workshop communication, a workshop communication topology map is obtained by setting a communication signal strength threshold of -70 dB and a relative distance threshold of 200 meters for screening. The vehicle node speed difference, destination direction angle, and communication quality parameters in the workshop communication topology are normalized. The vehicle negotiation score matrix is ​​obtained by calculating the speed matching degree as 1 minus the ratio of the speed difference to the maximum speed, the target similarity as the cosine of the destination direction angle, and the communication quality as the ratio of the signal strength to the maximum strength. Vehicle nodes that meet the negotiation conditions in the vehicle negotiation scoring matrix are grouped into formations. By establishing communication links and a shared scheduling information mechanism for vehicles within the formation, a vehicle collaborative formation organizational structure is obtained.

6. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 1, characterized in that, Step S4 includes: Based on the platooning communication link parameters in the vehicle cooperative platooning organization structure, a communication quality attenuation function is established by calculating the transmission distance and transmission delay, and the communication coverage quality of road segments in the road network is evaluated to obtain road segment communication quality evaluation data. Based on the road segment communication quality assessment data, combined with road segment distance, travel time, and energy consumption parameters, a weighted calculation is performed by setting weight coefficients for distance, time, communication, and energy consumption to obtain the comprehensive weight assessment result of the path. The road segments whose communication quality meets the requirements of vehicle-road cooperation in the comprehensive weight evaluation results of the path are analyzed by path combination. Multiple candidate paths are generated for each vehicle platoon and communication quality and cost parameters are labeled to obtain a set of candidate paths for vehicle-road cooperation.

7. The intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation according to claim 1, characterized in that, Step S5 includes: Each path in the vehicle-road cooperative candidate path set is encoded as a particle position vector for particle swarm initialization. By setting the particle swarm size and initial velocity parameters, the initial particle swarm of the improved vehicle-road cooperative particle swarm optimization algorithm is obtained. Based on the path selection scheme of each particle in the initial particle swarm, the weighted fitness value of total travel time, communication delay, energy consumption and road network congestion penalty is calculated. By recording the individual optimal position and the global optimal position, the particle swarm fitness evaluation result is obtained. Based on the particle swarm fitness evaluation results and combined with the real-time sensing data of the roadside unit, the roadside unit guidance terms are updated. Iterative calculations and distributed processing of edge computing nodes are performed through particle velocity and position update formulas to obtain optimized iterative convergence results. The optimization iteration convergence results are processed to extract the global optimal solution and generate a scheduling scheme. By assigning the optimal path and coordination parameters to each vehicle platoon, the global optimal scheduling scheme for vehicle-road cooperation is obtained.

8. A vehicle-road cooperative intelligent connected vehicle route planning and scheduling system, characterized in that, For implementing the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation as described in any one of claims 1-7, the intelligent connected vehicle path planning and scheduling system based on vehicle-road cooperation includes: The data acquisition module is used to collect vehicle status information and roadside perception data through the coverage area of ​​the roadside unit, and combine them with the vehicle individual demand parameters reported by the on-board unit to establish the topology data of the vehicle-road cooperative communication network. The analysis module is used to construct a vehicle-road communication delay weight matrix based on the vehicle-road cooperative communication network topology data, calculate the local reachability density through density clustering analysis, and obtain the spatiotemporal state clustering results of traffic flow in the roadside unit coverage area. The topology module is used to establish a workshop communication topology map based on the traffic flow spatiotemporal state clustering results, and to conduct workshop negotiation through vehicle negotiation functions and distributed negotiation strategies to obtain the vehicle collaborative formation organization structure. The generation module is used to establish a communication quality assessment model based on the vehicle cooperative formation organization structure, and combine the communication quality attenuation function with the path weight calculation to generate a set of vehicle-road cooperative candidate paths. The output module is used to input the set of vehicle-road cooperative candidate paths into the improved vehicle-road cooperative particle swarm optimization algorithm. By introducing the particle swarm iterative optimization and distributed computing of the roadside unit guidance term, the algorithm outputs the globally optimal vehicle-road cooperative scheduling scheme.

9. A path planning and scheduling device for intelligent connected vehicles based on vehicle-road cooperation, characterized in that, It includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it causes the processor to execute the intelligent connected vehicle path planning and scheduling method based on vehicle-road cooperation as described in any one of claims 1 to 7.

Images