Obstacle clearing vehicle intelligent patrol regulation method and system based on big data analysis

Abstract

The application discloses a method for intelligent patrol regulation of a wrecker based on big data analysis, and relates to the field of patrol regulation. The method comprises the following steps: acquiring a regional road network map; dividing a target region into a plurality of regional patrol grids; when receiving road wrecker information, integrating a target wrecker region; combining all road wreckers in the target wrecker region into an optimized game group, and acquiring wrecker configuration information and wrecker position information; real-time calling of wrecker traffic information; planning of regional wrecker routes for all road wreckers in the optimized game group; construction of a wrecker target function of all road wreckers, and screening of target wreckers in the optimized game group; marking of the regional wrecker routes corresponding to the target wreckers as target wrecker routes; real-time multi-scale modal decomposition of the wrecker traffic information, and construction of a traffic fluctuation propagation map; and generation of a wrecker patrol prompt item. The application can effectively improve the wrecker operation efficiency.

Description

Technical Field

[0001] This application relates to the field of patrol control, and in particular to a method and system for intelligent patrol control of clearing vehicles based on big data analysis. Background Technology

[0002] With the acceleration of urbanization and the continuous growth of motor vehicle ownership, urban road traffic pressure is becoming increasingly prominent. Various traffic accidents, vehicle breakdowns, cargo spills, and other emergencies occur frequently, leading to a decline in road capacity and even causing regional traffic congestion. Road clearing, as a crucial component of traffic emergency management, directly impacts the speed of road traffic recovery and overall operational order through its response efficiency and handling capabilities.

[0003] Traditional road clearing operations often rely on fixed patterns for dispatching road clearing vehicles, which can easily lead to unreasonable resource allocation, with some vehicles overloaded while others are idle. Furthermore, during peak traffic hours, relying on preset fixed dispatch patterns may cause secondary traffic congestion when road clearing vehicles enter the areas requiring clearing, potentially preventing vehicles from reaching the designated areas in a timely manner. This results in low efficiency in road clearing operations, traffic congestion, and even the possibility of secondary accidents. Summary of the Invention

[0004] This application provides a method and system for intelligent patrol and control of clearing vehicles based on big data analysis, which is used to improve the efficiency of clearing operations by clearing vehicles.

[0005] To achieve the above objectives, the embodiments of this application adopt the following technical solutions: Firstly, a method for intelligent patrol and control of clearing vehicles based on big data analysis is provided, the method comprising: Obtain the regional road network map of the target area; The target area is divided into multiple patrol grids based on the regional road network map; When road clearing information is received from any road clearing point within any area patrol grid, the area patrol grid and its corresponding adjacent patrol grids are integrated into a target clearing area. All road clearing vehicles within the target clearing area are grouped into an optimal game group, and the configuration and location information of the clearing vehicles in the optimal game group are obtained. Real-time access to traffic information related to clearing obstacles within the target area; By combining the location information of clearing vehicles and clearing traffic information, regional clearing routes are planned for all road clearing vehicles within the optimal game group; Based on regional clearing routes, road clearing information, and clearing vehicle configuration information, a clearing objective function for all road clearing vehicles is constructed, and target clearing vehicles within the optimization game group are selected according to the clearing objective function. Mark the area clearing route corresponding to the target clearing vehicle as the target clearing route; As the target clearing vehicle proceeds to the road clearing point along the target clearing route, the traffic information along the target clearing route is decomposed in real time using multi-scale modal analysis, and a traffic fluctuation propagation map along the regional clearing route is constructed based on the multi-scale modal analysis results. Based on the traffic fluctuation propagation map, a clearing patrol prompt is generated for the target clearing vehicle.

[0006] Optionally, combining the location information of clearing vehicles and clearing traffic information to plan regional clearing routes for all road clearing vehicles within the optimal game group includes the following steps: The target clearing area is divided into multiple clearing traffic grids based on the regional road network map; For any traffic clearing grid, the traffic multidimensional attributes of the traffic clearing grid are identified based on the regional road network map and traffic clearing information. These traffic multidimensional attributes include traffic topology attributes, traffic social attributes, traffic dynamic attributes, and traffic physical attributes. Based on the multidimensional attributes of traffic and the location information of the road clearing vehicle, multiple primary clearing routes are planned for the road clearing vehicle, and all primary clearing routes are integrated into a primary route set. For any clearing traffic grid that overlaps in area with any primary clearing route, the clearing traffic grid is marked as a clearing route grid; The grid accessibility of the clearing route grid is calculated based on the multi-dimensional traffic attributes of the clearing route grid. Based on the grid accessibility, all primary clearing routes in the primary route set are screened and optimized to obtain a candidate route set. If the total number of candidate clearance routes in the candidate route set is greater than the preset total threshold, then the clearance operation path segments of all candidate clearance routes in the candidate route set are extracted based on the road clearance points. Determine the obstacle removal driving posture parameters of the road clearing vehicle based on the regional road network map for the obstacle removal operation route; The clearance exit posture parameters of the road clearance vehicle are determined based on the pre-acquired clearance transportation destination and regional road network map; The attitude adjustment cost of the road clearing vehicle is calculated by combining the attitude parameters of the vehicle entering and exiting the clearing area, and the area clearing route of the road clearing vehicle is selected from the candidate route set based on the attitude adjustment cost.

[0007] Optionally, the grid accessibility of the clearing route grid is calculated based on the multi-dimensional traffic attributes of the clearing route grid, and the selection and optimization of all primary clearing routes in the primary route set are completed based on the grid accessibility to obtain the candidate route set, including the following steps: A grid traffic cost function for clearing routes is constructed by combining traffic topology and traffic physical properties. The grid traffic cost function includes travel distance cost, traffic light waiting cost, turning penalty, and gradient penalty. The traffic congestion index and traffic redundancy of the clearing route grid are calculated based on the traffic dynamic attributes, and the traffic risk coefficient of the clearing route grid is calculated by combining the traffic congestion index and traffic redundancy. The traffic sensitivity of the clearing route grid is determined based on the social attributes of traffic. For any clearing route grid, the grid accessibility of all clearing route grids is calculated by combining the grid access cost function, access risk coefficient and access sensitivity, and the accessibility verification of all clearing route grids is completed based on the grid accessibility. If the access verification of the clearing route grid fails, the route change verification of the clearing route grid is completed based on the traffic multidimensional attributes and the access verification results of the corresponding adjacent clearing route grids. If the route changeability check of the clearing route grid fails, then primary clearing routes that overlap with the clearing route grid in the primary route set will be removed. If the route changeability check of the obstacle removal route grid passes, then the primary obstacle removal routes that overlap with the obstacle removal route grid in the primary route set are locally optimized, and the primary obstacle removal routes that have completed local optimization are included in the primary route set to obtain the candidate route set.

[0008] Optionally, constructing a clearing objective function for all road clearing vehicles based on regional clearing routes, road clearing information, and clearing vehicle configuration information, and then selecting target clearing vehicles within the optimization game group based on the clearing objective function includes the following steps: Based on the road clearing information, the clearing needs of the target clearing area are broken down, and a clearing needs vector is constructed. For any road clearing vehicle in the optimization game group, the clearing function vector of the road clearing vehicle is constructed based on the clearing vehicle configuration information and the regional clearing route; The road clearing vehicle's clearance matching degree is calculated by combining the clearance demand vector and the clearance function vector; Based on the traffic multidimensional attributes of the traffic grid corresponding to the regional clearing route, the clearing cost information of the regional clearing route is calculated. Combined with the clearing cost information, clearing demand vector and clearing function vector, the clearing constraint set is defined, and the clearing closed convex set of the road clearing vehicle is constructed based on the clearing constraint set. The objective function for road clearing vehicles is constructed by combining information on clearing matching degree and clearing cost; By combining the closed convex set of obstacle removal and the obstacle removal objective function, the local optimal solution for all road clearing vehicles is calculated iteratively, and the target clearing vehicle in the optimization game group is selected based on the local optimal solution.

[0009] Optionally, the local optimal solution for all road clearing vehicles is calculated iteratively by combining the closed convex set of the clearing mechanism and the objective function of the clearing mechanism. The following steps are then taken to select the target clearing vehicle within the optimization game group based on the local optimal solution: The obstacle clearing constraints of all road clearing vehicles in the optimization game group are verified by using the obstacle clearing closed convex set, and the obstacle clearing decision variables are constructed by combining the obstacle clearing constraint verification results and the obstacle clearing function vector. By combining the obstacle removal decision variables and the obstacle removal objective function, and using the gradient descent method to iteratively calculate the local obstacle removal optimal solution for all road clearing vehicles; Based on the optimal solution of all local obstacle removal, complete the Lagrange multiplier exchange of all adjacent road obstacle removal workshops in the optimization game group, and based on the Lagrange multiplier exchange results, complete the iterative update of the optimal solution of all road obstacle removal vehicles to obtain the global optimal solution of obstacle removal; Extract the optimal solution function value of all global optimal solutions for obstacle clearing; Based on the optimal solution function value, the priority of all road clearing vehicles is sorted to obtain the priority of all road clearing vehicles. Target clearing vehicles within the optimized game group are selected based on the priority of clearing vehicles.

[0010] Optionally, performing multi-scale modal decomposition on the traffic information along the target clearance route in real time, and constructing a traffic fluctuation propagation map along the target clearance route based on the multi-scale modal decomposition results includes the following steps: The patrol grids of all areas overlapping the target clearance route are split into multiple area patrol sub-grids; For any patrol subgrid in a given area, calculate the time-series data of traffic density for the patrol subgrid based on the traffic information from clearing obstacles. An adaptive signal decomposition method is used to perform multi-scale mode decomposition on traffic density time series data to obtain multi-scale traffic signals; Multi-scale fluctuation characteristic parameters of multi-scale traffic signals are extracted, including multi-scale fluctuation amplitude, multi-scale fluctuation phase and multi-scale fluctuation frequency. All regional patrol subgrids are treated as traffic nodes, and all multi-scale fluctuation characteristic parameters are used as node attributes of the corresponding traffic nodes. The traffic complex impedance parameters of all traffic nodes are calculated based on the multi-scale fluctuation characteristic parameters. The coherence analysis algorithm was used to complete the coherence analysis of all multi-scale traffic signals and obtain the fluctuation coherence coefficients between all adjacent traffic nodes. Identify the timing of fluctuation events in multi-scale traffic signals based on multi-scale fluctuation characteristic parameters; By combining the wave coherence coefficient, traffic complex impedance parameters, and wave event time, a directed edge for propagation between all adjacent traffic nodes is constructed, resulting in a traffic wave propagation graph.

[0011] Optionally, using an adaptive signal decomposition method to perform multi-scale modal decomposition on the traffic density time series data to obtain multi-scale traffic signals includes the following steps: Traffic density time-series data is standardized to obtain standard density time-series data. The variational mode decomposition algorithm is used to perform macroscale decomposition of standard density time series data to obtain traffic macro-fluctuation signals and traffic macro-residual terms; The wavelet transform algorithm is used to perform mesoscale decomposition of the macroscopic residual term of traffic, and the mesoscopic fluctuation signal and mesoscopic residual term of traffic are obtained. The traffic mesoscopic residual term is decomposed at a microscale using the empirical mode decomposition algorithm to obtain the traffic micro-fluctuation signal; By integrating macroscopic, mesoscopic, and microscopic traffic fluctuation signals, multi-scale traffic signals are obtained.

[0012] Optionally, generating clearing patrol prompts for the target clearing vehicle based on the traffic fluctuation propagation map includes the following steps: The route stability of the target obstacle removal route is calculated based on traffic micro-fluctuation signals and using the Lyapunov exponent calculation algorithm. The route coupling risk value of the target obstacle removal route is calculated based on the time series data of route stability and traffic density. Based on the route coupling risk value, mark the key points of traffic fluctuations in the traffic fluctuation propagation diagram; Based on the traffic wave propagation map, phase consistency analysis of all key points of traffic waves was completed, and the traffic wave resonance zone of the target clearance route was located based on the phase consistency analysis results. Based on multi-scale traffic signals, the resonance criticality verification of the traffic fluctuation resonance zone is completed; If the traffic fluctuation resonance zone fails the resonance critical test, the proportion of fluctuation energy of all traffic nodes in the traffic fluctuation resonance zone is calculated based on the multi-scale fluctuation characteristic parameters. By combining the wave energy ratio and the resonance zone coordinates of the traffic wave resonance zone, phase resolution analysis of the traffic wave resonance zone is completed, and a clearing patrol prompt item is generated for the target clearing vehicle based on the phase resolution analysis results.

[0013] Secondly, this application provides a machine-readable storage medium storing instructions that cause a machine to execute the intelligent patrol control method for a clearing vehicle based on big data analysis as described in the first aspect.

[0014] Thirdly, this application provides an intelligent patrol and control system for clearing vehicles based on big data analysis, comprising: The memory is configured to store instructions; and The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the intelligent patrol and control method for clearing vehicles based on big data analysis as described in the first aspect.

[0015] The above technical solution divides the target area into multiple smaller patrol grids using a regional road network map. Each patrol grid consists of a certain number of roads, making the patrol range of road clearing vehicles more clearly defined, improving patrol coverage and efficiency. Furthermore, when a road clearing point appears, it facilitates rapid location of the corresponding patrol grid and quick mobilization of clearing vehicles from adjacent grids, shortening the response time of the clearing vehicles. In addition, to improve the speed at which grid clearing vehicles arrive at road clearing points and to ensure that the mobilized grid clearing vehicles can complete the clearing task, it is also necessary to plan regional clearing routes for each grid clearing vehicle based on the vehicle's location information and traffic information. This allows for precise calculation of the clearing matching degree and clearing cost information for each grid clearing vehicle, ensuring that the grid clearing vehicles can handle the clearing task at the road clearing point while minimizing their patrol and control costs. Simultaneously, the planned regional clearing routes also prevent road congestion when road clearing vehicles are heading to road clearing points, avoiding delays or failures in clearing tasks. Next, as the target clearing vehicle travels along the designated clearing route to the road clearing point, the system analyzes the traffic conditions along its route in real time based on traffic information. If new congestion or worsening traffic conditions occur, the system can generate advance patrol alerts, providing a basis for the target clearing vehicle to adjust its route or driving, minimizing delays and ensuring it reaches the road clearing point in the shortest possible time. In summary, this application, through pre-planning routes, screening clearing vehicles, and providing real-time traffic alerts, effectively improves the response speed and patrol control speed of clearing operations, thereby shortening road clearing time, increasing road clearing efficiency, reducing the impact of traffic congestion, and preventing secondary accidents.

[0016] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0017] Figure 1 A flowchart illustrating an intelligent patrol and control method for clearing vehicles based on big data analysis, provided for an embodiment of this application; Figure 2 A flowchart illustrating a method for planning a regional obstacle removal route provided in an embodiment of this application; Figure 3This is a flowchart illustrating a method for generating obstacle clearing patrol prompts, provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0020] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0021] Figure 1 The illustration shows a flowchart of a smart patrol and control method for clearing vehicles based on big data analysis, according to an embodiment of this application. Figure 1 As shown in the figure, this application provides an intelligent patrol and control method for clearing vehicles based on big data analysis. The method may include the following steps: S101. Obtain the regional road network map of the target area; S102. Divide the target area into multiple patrol grids based on the regional road network map; S103. When road clearing information is received from any road clearing point within any area patrol grid, the area patrol grid and the corresponding adjacent patrol grid are integrated into a target clearing area. S104. Combine all road clearing vehicles within the target clearing area into an optimal game group, and obtain the clearing vehicle configuration information and clearing vehicle location information of the optimal game group. S105. Real-time retrieval of traffic information related to clearing obstacles within the target area; S106. Combining the location information of clearing vehicles and clearing traffic information, regional clearing routes are planned for all road clearing vehicles within the optimized game group. S107. Based on the regional clearing routes, road clearing information, and clearing vehicle configuration information, construct the clearing objective function for all road clearing vehicles, and select the target clearing vehicles within the optimization game group according to the clearing objective function. S108. Mark the area clearing route corresponding to the target clearing vehicle as the target clearing route; S109. When the target clearing vehicle is heading to the road clearing point along the target clearing route, perform multi-scale modal decomposition on the clearing traffic information on the target clearing route in real time, and construct a traffic fluctuation propagation map on the regional clearing route based on the multi-scale modal decomposition results. S110. Generate clearing patrol prompts for target clearing vehicles based on traffic fluctuation propagation maps.

[0022] In this embodiment, the regional road network map can be obtained through a GIS geographic information system, a traffic management department database, or a high-precision map platform. It includes road topology information within the target area, such as road boundaries, intersection locations, number of lanes, road classification (arterial road, secondary arterial road, local road), social information (such as the distribution and names of buildings around the road, e.g., a certain middle school, a certain central hospital), and physical information (such as road width and height restrictions). Next, the target area is divided into multiple regional patrol grids based on the regional road network map. Regular rectangular grids can be used along the road direction (intersections, curves, and other special areas are adapted to polygonal grids), ensuring no overlap or gaps between grids and complete coverage of the target area. When the clearing vehicle control center receives road clearing information from any clearing point within any regional patrol grid, it merges the regional patrol grid and its adjacent regional patrol grids (i.e., adjacent patrol grids with overlapping edges) into the target clearing area. Next, the road clearing vehicles in the target clearing area that are on standby or empty are formed into an optimization game group, and the configuration information and location information of each road clearing vehicle in the optimization game group are retrieved. The configuration information of the clearing vehicles includes, but is not limited to, the type of clearing vehicle: such as flatbed towing type, heavy lifting type, comprehensive operation type, etc. The location information of the clearing vehicles refers to the real-time GPS positioning data of the road clearing vehicles.

[0023] Next, traffic information for clearing obstacles within the target area is collected using high-definition cameras, roadside radar, and other equipment, with a sampling frequency of 0.1 seconds per instance. Then, multiple primary clearing routes are planned based on the current location of each road clearing vehicle. These primary routes are then filtered and optimized based on the traffic information to obtain regional clearing routes. Next, clearing cost information is accurately calculated based on the regional clearing routes. Clearing demand vectors and clearing function vectors are constructed based on road clearing information and vehicle configuration information. Then, clearing constraints are defined based on clearing cost, total actual travel distance, total actual travel time, clearing demand vector, and clearing function vector, and all clearing constraints are converted into mathematically closed convex sets. Simultaneously, a clearing objective function is constructed with the goal of maximizing clearing matching degree and minimizing clearing cost. The clearing objective function includes travel distance cost, traffic light waiting cost, and travel time cost. Next, by combining the closed convex set of obstacle removal and the obstacle removal objective function, the local obstacle removal optimal solution for all road obstacle removal vehicles is calculated iteratively. After global optimization of the local obstacle removal optimal solution, the global obstacle removal optimal solution for all road obstacle removal vehicles is calculated. Based on the global obstacle removal optimal solution, all road obstacle removal vehicles are prioritized and the road obstacle removal vehicle with the highest priority is selected as the target obstacle removal vehicle.

[0024] The designated area clearing route for the target clearing vehicle is marked as the target clearing route. As the target clearing vehicle proceeds to the road clearing point along this route, real-time traffic density time-series data for the target clearing route is extracted based on retrieved traffic information. Multi-scale mode decomposition is then performed on this traffic density time-series data to obtain multi-scale traffic signals for the target clearing route. Multi-scale fluctuation feature parameters of the multi-scale traffic signals are extracted, including multi-scale fluctuation amplitude, phase, and frequency. All area patrol sub-grids are treated as traffic nodes, and all multi-scale fluctuation feature parameters are used as node attributes for the corresponding traffic nodes. Next, directed edges are constructed between adjacent traffic nodes that exhibit coherence and share the same fluctuation event, resulting in a traffic fluctuation propagation graph. Then, key traffic fluctuation points in the propagation graph are marked, generating clearing patrol prompts for the target clearing vehicle driver. This allows the driver to anticipate road conditions ahead and make optimal driving maneuvers, preventing traffic accidents and ensuring efficient completion of clearing operations.

[0025] In one embodiment, planning regional clearing routes for all road clearing vehicles within the optimization game group by combining clearing vehicle location information and clearing traffic information includes the following steps: S201. Divide the target clearing area into multiple clearing traffic grids according to the regional road network map; S202. For any traffic clearing grid, the traffic multidimensional attributes of the traffic clearing grid are identified based on the regional road network map and traffic clearing information. The traffic multidimensional attributes include traffic topology attributes, traffic social attributes, traffic dynamic attributes, and traffic physical attributes. S203. Based on the multi-dimensional attributes of traffic and the location information of the road clearing vehicle, multiple primary clearing routes are planned for the road clearing vehicle, and all primary clearing routes are integrated into a set of primary routes. S204. For any clearing traffic grid that overlaps in area with any primary clearing route, mark the clearing traffic grid as a clearing route grid. S205. Calculate the grid accessibility of the clearing route grid based on the multi-dimensional traffic attributes of the clearing route grid, and complete the screening and optimization of all primary clearing routes in the primary route set based on the grid accessibility to obtain a candidate route set. S206. If the total number of candidate clearance routes in the candidate route set is greater than the preset total threshold, then the clearance operation path segments of all candidate clearance routes in the candidate route set are extracted based on the road clearance points. S207. Determine the clearing and entry attitude parameters of the road clearing vehicle based on the regional road network map for the clearing operation route section; S207. Determine the clearing and exiting posture parameters of the road clearing vehicle based on the pre-acquired clearing and transportation destination and regional road network map; S208. Calculate the attitude adjustment cost of the road clearing vehicle by combining the attitude parameters of entering and exiting the clearing vehicle, and select the area clearing route of the road clearing vehicle from the candidate route set based on the attitude adjustment cost.

[0026] In this embodiment, the core elements of the regional road network map are first extracted. These core elements include road boundaries, intersection locations, number of lanes, and road grades, such as arterial roads, secondary arterial roads, local roads, and road network density distribution. Next, the grid division granularity is determined based on the location of the target area. If the target area is a core urban area, such as a region with a road network density ≥ 8 roads / km², the grid division granularity is smaller compared to suburban urban areas, such as a road network density < 4 roads / km². Furthermore, when dividing the regional road network map, roads of the same road grade are grouped into one grid; for example, multiple arterial roads are grouped into the same clearance traffic grid. Then, a unique identifier code is assigned to each clearance traffic grid, for example, G-XX-YY, where G is the Gth clearance traffic grid, XX is the horizontal coordinate of the two ends of the clearance traffic grid, and YY is the vertical coordinate of the two ends of the clearance traffic grid. This unique identifier code serves as the basic carrier for subsequent attribute labeling and route planning.

[0027] The traffic topology, social attributes, and physical attributes of each clearing traffic grid are extracted from the regional road network map. Traffic topology attributes include the number of one-way lanes, the number of two-way lanes, intersection turning rules (e.g., whether left turns or U-turns are prohibited), and road connectivity, primarily including the connection method with adjacent regional road network maps, the road segment length converted from the regional road network map scale, and the location and control cycle of traffic lights. Social attributes refer to the social value characteristics of the clearing traffic grid, i.e., whether it contains core functional areas such as hospitals and schools. Since road clearing vehicles are often large, they can easily cause traffic congestion or accidents; therefore, these core functional areas should be avoided as much as possible in subsequent route planning. Physical attributes include road width, road gradient (including uphill and downhill slopes and their corresponding gradient values), and road load-bearing limits within the clearing traffic grid. Simultaneously, dynamic traffic attributes are extracted from the clearing traffic information. These attributes include real-time traffic flow, average vehicle speed, and temporary traffic control information, such as whether there are construction closures or temporary restrictions. The clearing traffic information mainly refers to real-time traffic monitoring data obtained through traffic monitoring, vehicle-mounted terminals, and road condition platforms. The extracted multi-dimensional traffic attributes are then associated with the unique identifier code of each clearing traffic grid to facilitate subsequent traffic status analysis of each grid.

[0028] Next, based on traffic multidimensional attributes and the location information of the road clearing vehicle, multiple primary clearing routes are planned for the road clearing vehicle. Specifically, the starting point of the route is first determined based on the location information of the road clearing vehicle, and then the location of the road clearing point is used as the endpoint. Then, based on traffic multidimensional attributes, absolutely impassable road network areas are marked on the regional road network map, such as areas under construction closure, temporary traffic restrictions, or areas where large-scale events are being held, such as concerts. Next, a path planning algorithm is used to generate multiple primary clearing routes based on the shortest travel distance. Each primary clearing route contains a complete sequence of road network nodes (intersections, road segments) and a corresponding clearing traffic grid sequence. Commonly used path planning algorithms include Dijkstra's algorithm and improved A / B algorithm. Algorithms, etc., to improve A Taking the algorithm as an example, road intersections, the start and end points of the clearing route, and fixed obstacle points such as guardrails and construction barriers are abstracted as nodes. The passable road segments between nodes are abstracted as edges. Then, edge weights are assigned based on the actual travel distance of each passable road segment. Next, A is retained. Evaluation function framework for the algorithm: However, optimizing the node storage logic to support multi-path generation is about to... Defined as the distance from the starting point to the node The actual shortest distance will For nodes The heuristic distance to the destination, where the heuristic distance needs to satisfy acceptability, i.e. The distance should be less than or equal to the actual shortest distance. For example, Manhattan distance is used when the road network is urban roads, and Euclidean distance is used when the road network is suburban roads. This avoids overestimating the distance and causing the shortest path to be missed. The system stores the shortest path information for each node. The predecessor node that reaches the minimum actual distance is identified. Then, iterative multi-path search and expansion are performed. Initially, the starting point is added to the open list, and the actual distance from the starting point to itself is marked as 0. Simultaneously, the heuristic distance from the starting point to the destination is calculated to ensure that this distance does not exceed the actual shortest distance. At this point, the starting point has no predecessor node. Next, the node with the smallest sum of actual distance and heuristic distance is searched from the list of nodes to be explored and selected as the node n to be expanded. All passable neighboring nodes m of n are found, which must satisfy the clearance vehicle passage constraint, such as avoiding impassable road sections. If m has never been visited, the actual distance from the starting point to n is calculated plus the length of the road section from n to m. Simultaneously, the sum of the actual distance from the starting point to n and the heuristic distance of the neighboring node is calculated. Record n ​​as the predecessor node of m and add n to the list of nodes to be explored. If m has been visited and the newly calculated actual distance is smaller than its recorded actual distance, update m's actual distance to this smaller value, clear all predecessor nodes of m, and keep only n as its predecessor node. If m has been visited and the newly calculated actual distance is equal to its recorded actual distance, do not change its actual distance, record n ​​as the predecessor node of m, keep this path branch with equal shortest distance, and remove n from the list of nodes to be explored. When the destination is also removed from the list of nodes to be explored, stop the above steps, use the destination as the backtracking starting point, traverse all its predecessor nodes until backtracking to the starting point, and use all routes that can backtrack to the starting point during the destination backtracking process as primary obstacle clearing routes. Integrate all primary obstacle clearing routes into a primary route set.

[0029] Next, based on the coordinates of each traffic clearing grid, it is verified whether any primary clearing route passes through that grid. The area of ​​the primary clearing route passing through the grid is calculated as a percentage of the total area of ​​the grid. If this percentage is greater than a preset threshold, such as 5%, it is determined that the clearing grid overlaps with a primary clearing route, and the grid is marked as a clearing route grid. Then, the grid accessibility of each clearing route grid is calculated. Primary clearing routes that fail both accessibility verification and route changeability verification are removed from the primary route set. Primary clearing routes that fail accessibility verification but pass route changeability verification are locally optimized and added to the primary route set as candidate clearing routes. All primary clearing routes that pass accessibility verification are also added as candidate clearing routes to form a candidate route set. Each route in the candidate route set undergoes cost accounting, accessibility verification, and optimization adjustments to ensure basic feasibility.

[0030] If the total number of candidate clearance routes in the candidate route set exceeds a preset threshold (which can be 1), meaning there is more than one candidate clearance route, the candidate route with the lowest attitude adjustment cost can be selected as the area clearance route for the road clearing vehicle based on the clearance entry and exit attitude parameters. First, based on the location of the area clearance point, a clearance operation path segment is extracted from each candidate clearance route in the candidate route set. This can be the path segment traversed by candidate clearance routes within a preset range around the area clearance point, such as a radius of 50 meters. The traffic physical attributes of this clearance operation path segment are recorded, such as the work space, road boundaries, and traffic topology attributes, such as lane distribution. Using the road tangent direction of the area where the area clearance point is located as a reference, and combined with work space constraints such as shoulder width and obstacle location, the entry angle between the longitudinal axis of the road clearing vehicle and the road tangent direction when parking is determined. The optimal angle is 0-30°, with the vehicle parking in the forward direction. These parameters are then integrated into the clearance entry attitude parameters.

[0031] Next, based on the pre-acquired clearing and transportation endpoints, such as parking lots and repair stations, the exit posture parameters are determined. Using the direction of the clearing and transportation endpoint as a reference, and considering the topological connection between the area where the regional clearing point is located and the subsequent transportation route, the exit angle between the longitudinal axis of the road clearing vehicle and the direction of the transportation endpoint is determined when the road clearing vehicle starts after loading the obstacle vehicle. The optimal angle is 0-30°. The forward start, the adaptation relationship between the start position and the surrounding road network, and the presence of obstacles are considered to form the exit posture parameters. Then, combining the clearing entry posture parameters and the clearing exit posture parameters, the posture adjustment cost of the road clearing vehicle when performing clearing operations according to different candidate clearing routes is calculated. First, the turning cost of the road clearing vehicle is determined by combining the entry angle and the exit angle. For example, when the entry angle or exit angle is within the range of (0°, 30°), the cost index is 1; within the range of [30°, 90°), the cost index is 3; within the range of [90°~150°), the cost index is 5; and within the range of [150°, 180°), the cost index is 7. Then, the cost indices of the entry angle and the exit angle are added together and normalized to obtain the turning cost of the road clearing vehicle. Next, obstacle costs are assigned based on the presence of obstacles (such as curbs, green belts, etc.). For example, when there are no obstacles, the obstacle cost is set to 0; when there are obstacles, the obstacle cost is determined according to the number of obstacles. If the number of obstacles is less than or equal to 2, the obstacle cost is 1; if it is greater than 2, the obstacle cost is 2. After normalizing the obstacle cost, it is weighted and summed with the turning cost. For example, the weight of the obstacle cost is 0.4, and the weight of the turning cost is 0.6 to obtain the attitude adjustment cost. The candidate route with the lowest attitude adjustment cost from the candidate route set is selected as the area's obstacle removal route. This provides a data foundation for subsequent calculations of obstacle removal costs and selection of target obstacle removal vehicles.

[0032] In one embodiment, the grid accessibility of the clearing route grid is calculated based on the multi-dimensional traffic attributes of the clearing route grid, and the selection and optimization of all primary clearing routes in the primary route set are completed based on the grid accessibility to obtain a candidate route set, including the following steps: A grid traffic cost function for clearing routes is constructed by combining traffic topology and traffic physical properties. The grid traffic cost function includes travel distance cost, traffic light waiting cost, turning penalty, and gradient penalty. The traffic congestion index and traffic redundancy of the clearing route grid are calculated based on the traffic dynamic attributes, and the traffic risk coefficient of the clearing route grid is calculated by combining the traffic congestion index and traffic redundancy. The traffic sensitivity of the clearing route grid is determined based on the social attributes of traffic. For any clearing route grid, the grid accessibility of all clearing route grids is calculated by combining the grid access cost function, access risk coefficient and access sensitivity, and the accessibility verification of all clearing route grids is completed based on the grid accessibility. If the access verification of the clearing route grid fails, the route change verification of the clearing route grid is completed based on the traffic multidimensional attributes and the access verification results of the corresponding adjacent clearing route grids. If the route changeability check of the clearing route grid fails, then primary clearing routes that overlap with the clearing route grid in the primary route set will be removed. If the route changeability check of the obstacle removal route grid passes, then the primary obstacle removal routes that overlap with the obstacle removal route grid in the primary route set are locally optimized, and the primary obstacle removal routes that have completed local optimization are included in the primary route set to obtain the candidate route set.

[0033] In this embodiment, the actual travel distance of the road clearing vehicle through the clearing route grid according to the primary clearing route is calculated based on the road segment length in the traffic topology attributes. Then, the product of the unit mileage fuel consumption cost and the actual travel distance for each road clearing vehicle is calculated. This product is normalized and used as the travel distance cost. The unit mileage fuel consumption cost is expressed in yuan / km. Since fuel consumption differs between unloaded and loaded states, the actual unit mileage fuel consumption cost needs to be determined based on the actual load of each road clearing vehicle, which can be determined from the historical patrol records of each vehicle. Next, the traffic light waiting time is calculated based on the traffic signal location and control cycle in the traffic topology attributes. The red light time of all traffic signals can be directly used as the traffic light waiting time, which is then normalized and used as the traffic light waiting cost. Finally, based on the intersection turning rules in the traffic topology attributes and the turning radius requirements of the road clearing vehicle, a turning penalty is set. ,in This is the steering penalty coefficient. For travel distance costs, such as when there are no-turn zones, = Turning is permitted, but the turning radius must be less than or equal to the turning radius requirement for road clearing vehicles. =2, allowing turning and the turning radius must be greater than the turning radius requirement for road clearing vehicles. =1, road segment with no turning requirement =0. The traffic gradient penalty can be the product of the gradient penalty coefficient and the travel distance cost. The gradient penalty coefficient can be determined based on the road gradient and the load capacity of the recovery vehicle (uphill load loss, downhill braking loss) in the traffic physics attributes. For example, the gradient penalty coefficient is 1 on flat roads, 1.2 for uphill gradients [5°, 10°], 1.5 for uphill gradients [10°, 20°], and 2 for uphill gradients greater than 20°. It is also set to 1.1 for downhill gradients [5°, 10°] and 1.3 for downhill gradients greater than 10°. If the recovery vehicle has a strong load capacity, its gradient penalty coefficient can be appropriately reduced. Then, the travel distance cost, traffic light waiting cost, turning penalty, and traffic gradient penalty are weighted using expert scoring or entropy weighting methods and summed to obtain the grid traffic cost function. ,in, For the cost of travel distance, For traffic light waiting costs, Penalty for turning while passing through, Slope penalty , , and The weights are respectively for travel distance cost, traffic light waiting cost, turning penalty, and gradient penalty.

[0034] The traffic congestion index can be calculated based on the average vehicle speed in the traffic dynamics attributes. The initial congestion index calculation formula is as follows: ,in, The average traffic speed, The road design speed refers to the maximum safe speed that can be maintained under good road traffic and weather conditions, limited only by the physical conditions of the road. When the initial congestion index is negative, it is set to 0 and used as the traffic congestion index; when the initial congestion index is positive, it is directly used as the traffic congestion index. Traffic redundancy is calculated based on real-time traffic flow in the traffic dynamics attributes, and its calculation formula is as follows: , For real-time traffic flow, Design traffic flow refers to the maximum traffic volume a road segment can safely and orderly handle per unit time under favorable weather conditions, limited only by physical road conditions. This can be obtained by retrieving preliminary design documents and construction drawings for the current road segment. The traffic risk coefficient of the clearing route grid is calculated based on the calculated congestion index and traffic redundancy level. The calculation formula is as follows: ,in, The traffic congestion index. For the degree of redundancy in traffic, and The weights for traffic congestion index and traffic redundancy can be determined using the entropy weight method or multi-level analysis method.

[0035] Traffic sensitivity is a quantitative indicator used to quantify the impact of road patrol vehicles on social order and public services within different clearing route grids. It is generally determined based on the social attributes of traffic. For example, when the clearing route grid includes core functional areas such as hospitals, schools, and fire stations, the communication sensitivity is set to 1. When the clearing route grid includes non-core functional areas such as ordinary residential areas and commercial areas with generally high pedestrian traffic, the communication sensitivity is set to 0.6. When the clearing route grid includes non-core functional areas such as suburbs with generally low pedestrian traffic, the communication sensitivity is set to 0.2. "General pedestrian traffic" refers to the average pedestrian traffic in the area under normal circumstances, such as non-holidays and when no large-scale events are held, which can be statistically analyzed using video surveillance deployed in each area. After calculating the grid traffic cost using the above grid traffic cost function, the grid traffic cost, traffic risk coefficient, and traffic sensitivity are normalized and then weighted and summed. The weights can be determined using expert scoring or entropy weighting methods to obtain the prohibition index: Grid trafficability = 1 - Prohibition index. The higher the grid accessibility, the more suitable the roads within the clearing route grid are for road clearing vehicles.

[0036] The passage of all road clearing route grids is verified based on the grid accessibility. If the grid accessibility of a road clearing route grid is greater than or equal to a preset accessibility threshold, the passage verification is considered successful; if the grid accessibility of a road clearing route grid is less than the preset accessibility threshold, the passage verification is considered unsuccessful. The accessibility threshold can be determined by multiplying the median grid accessibility of all road clearing route grids corresponding to all road clearing vehicles by 1.5 to obtain the accessibility threshold.

[0037] If the accessibility verification of a clearing route grid fails, the route changeability verification of the clearing route grid is completed based on the traffic multidimensional attributes and the accessibility verification results of its corresponding adjacent clearing route grids. If multiple adjacent clearing route grids exist, such as two adjacent clearing route grids passing the accessibility verification, and the road network topology of the adjacent clearing route grids and the clearing route grid can be smoothly connected (e.g., without constraints such as prohibited turns or road interruptions), and the physical attributes of the adjacent clearing route grids, such as road width and slope, meet the requirements for road clearing vehicle passage, then the route changeability verification of the clearing route grid is determined to have passed. This means that the route can be partially adjusted to bypass the clearing route grid and pass through adjacent clearing route grids, and the primary clearing route that has completed the partial route adjustment is included in the primary route set as a candidate clearing route. If all adjacent clearing route grids fail the access verification, or if adjacent clearing route grids cannot achieve topological connection with this clearing route grid, or if the physical attributes of adjacent clearing route grids do not meet the requirements for road clearing vehicle passage, then the route change verification of this clearing route grid is determined to fail. This means that primary clearing routes with overlapping areas with this clearing route grid cannot bypass the clearing route grid that failed the access verification by adjusting the route locally. Therefore, this primary clearing route can be directly removed from the primary route set.

[0038] After completing the above steps, the change verification and route optimization of all the clearance route grids that failed the passage verification are completed. The remaining primary clearance routes, including the partially optimized routes, are used as candidate clearance routes to form a candidate route set. Each route in the candidate route set has undergone cost accounting, passage verification and optimization adjustment to ensure basic feasibility.

[0039] In one embodiment, constructing a clearing objective function for all road clearing vehicles based on regional clearing routes, road clearing information, and clearing vehicle configuration information, and then selecting target clearing vehicles within the optimization game group based on the clearing objective function includes the following steps: Based on the road clearing information, the clearing needs of the target clearing area are broken down, and a clearing needs vector is constructed. For any road clearing vehicle in the optimization game group, the clearing function vector of the road clearing vehicle is constructed based on the clearing vehicle configuration information and the regional clearing route; The road clearing vehicle's clearance matching degree is calculated by combining the clearance demand vector and the clearance function vector; Based on the traffic multidimensional attributes of the traffic grid corresponding to the regional clearing route, the clearing cost information of the regional clearing route is calculated. Combined with the clearing cost information, clearing demand vector and clearing function vector, the clearing constraint set is defined, and the clearing closed convex set of the road clearing vehicle is constructed based on the clearing constraint set. The objective function for road clearing vehicles is constructed by combining information on clearing matching degree and clearing cost; By combining the closed convex set of obstacle removal and the obstacle removal objective function, the local optimal solution for all road clearing vehicles is calculated iteratively, and the target clearing vehicle in the optimization game group is selected based on the local optimal solution.

[0040] In this embodiment, road clearing information includes, but is not limited to, fault or accident type: such as minor scratches, severe collisions, vehicle breakdowns, road obstacles, etc.; specific location of the fault or accident: including latitude and longitude, road name, lane number, etc.; required clearing operation type: such as towing, hoisting, road cleaning, emergency dismantling, etc.; clearing operation time requirements: such as arriving on site within 30 minutes, etc. Next, the road clearing information is broken down into multiple dimensions to obtain detailed multi-dimensional clearing requirements. These requirements are then quantified to form an ordered numerical vector, which is the clearing requirement vector. Each element in the clearing requirement vector corresponds to a quantified value of a specific clearing requirement. For example, towing requirements: 0 = not needed, 1 = 2 tons and below, 2 = 2-5 tons, 3 = 5-10 tons, 4 = above 10 tons; hoisting requirements: 0 = not needed, 1 = 5 tons and below, 2 = 5-10 tons, 3 = above 10 tons; road surface cleaning requirements: 0 = not needed, 1 = 10 square meters and below, 2 = 10-30 square meters, 3 = above 30 square meters; emergency dismantling requirements: 0 = not needed, 1 = simple dismantling, 2 = complex dismantling; operation time requirements (time required to arrive): 0 = no urgent requirement (>30 minutes), 1 = moderately urgent (15-30 minutes), 2 = highly urgent (≤15 minutes). This vector is used to accurately describe the core requirements of the clearing task. For example, the obstacle clearing demand vector is [1, 0, 2, 1, 2], which means, in order, traction demand of 2 tons and below, no hoisting demand, road clearing demand of 10-30 square meters, and simple emergency dismantling demand that is highly urgent.

[0041] Next, the functional configurations of each road clearing vehicle within the optimized game group are decomposed to construct a clearing vehicle function vector. The elements of each clearing vehicle function vector correspond one-to-one with the elements of the clearing demand vector, and the quantification rules are consistent. Clearing vehicle configuration information includes, but is not limited to, clearing vehicle type: such as flatbed towing type, heavy-duty lifting type, comprehensive operation type, etc.; core functional parameters: such as maximum towing weight, maximum lifting height or weight, road clearing efficiency, etc.; current load status: whether it is empty or loaded; equipment condition: whether each functional module can work normally; and remaining energy: the remaining fuel or electric range. The clearing vehicle function vector includes towing function, lifting function, road clearing function, emergency dismantling function, and arrival time. The arrival time is calculated based on the area clearing route corresponding to each road clearing vehicle. After calculating the clearing vehicle function vector for each road clearing vehicle, the clearing matching degree between each clearing vehicle function vector and the clearing demand vector is calculated. The clearing matching degree can be obtained by calculating the cosine similarity between the clearing vehicle function vector and the clearing demand vector.

[0042] The grid traffic cost function can be modified by removing the turning penalty and gradient penalty, retaining only the distance cost and traffic light waiting cost, and then adding the travel time cost to obtain the grid traffic cost function for each clearing route grid. Then, the grid traffic cost functions corresponding to each road clearing vehicle's area clearing route are summed to obtain the route traffic cost function. The calculation methods for distance cost and traffic light waiting cost are the same as those for the grid traffic cost function. The total actual travel time, i.e., the arrival time, is obtained by multiplying the average speed of each road segment by the actual distance traveled on that segment. The total actual travel time is then normalized to obtain the travel time cost. The clearing cost is calculated based on the route traffic cost function, and the clearing cost, total actual distance, and total actual time are integrated into the clearing cost information. Next, based on the clearing cost, total actual travel distance, total actual travel time, clearing demand vector, and clearing function vector, clearing constraints are defined. All clearing constraints are then converted into mathematically closed convex sets. A closed convex set satisfies the property that the line connecting any two points remains within the set and the set contains all boundary points. The clearing constraints include: the quantized value of each element in the clearing function vector is greater than or equal to a certain percentage of the quantized value of the corresponding element in the clearing demand vector, for example, 80%; the total actual travel time is less than or equal to the operation time requirement in the clearing demand vector; and the total actual travel distance is less than or equal to a certain percentage of the remaining fuel or electric range, for example, 40%, to prevent road clearing vehicles from being unable to complete the clearing operation smoothly. After converting all clearing constraints into mathematical inequalities, their intersection is defined as the clearing closed convex set for each road clearing vehicle, i.e., the clearing decision variable for the clearing vehicle. Only when all attributes of a road clearing vehicle fall into this clearing closed convex set can it be considered as a target clearing vehicle.

[0043] Next, a clearing objective function is constructed with the goal of maximizing the clearing matching degree and minimizing the clearing cost. The clearing objective function is as follows: ,in, and These are the obstacle clearing matching scores. The weights of the route travel cost function can be determined using expert scoring or entropy weighting. k refers to the total number of clearing route grids that overlap with the area of ​​the regional clearing route. For the cost of travel distance, For traffic light waiting costs, For travel time cost, , and The weights for travel distance cost, traffic light waiting cost, and travel time cost can be determined using expert scoring or entropy weighting.

[0044] Next, by combining the closed convex set of obstacle removal and the obstacle removal objective function, the local obstacle removal optimal solution for all road obstacle removal vehicles is calculated iteratively. After global optimization of the local obstacle removal optimal solution, the global obstacle removal optimal solution for all road obstacle removal vehicles is calculated. Based on the global obstacle removal optimal solution, all road obstacle removal vehicles are prioritized and the road obstacle removal vehicle with the highest priority is selected as the target obstacle removal vehicle.

[0045] Through the above steps, not only can the road clearing vehicles required for the clearing operation be accurately matched, but the most cost-effective road clearing vehicles can also be selected as the target clearing vehicles from various dimensions such as cost and time, effectively ensuring the efficient completion of the clearing task.

[0046] In one embodiment, the local optimal solution for all road clearing vehicles is calculated iteratively by combining the clearing closed convex set and the clearing objective function, and the target clearing vehicle in the optimization game group is selected based on the local optimal solution, including the following steps: The obstacle clearing constraints of all road clearing vehicles in the optimization game group are verified by using the obstacle clearing closed convex set, and the obstacle clearing decision variables are constructed by combining the obstacle clearing constraint verification results and the obstacle clearing function vector. By combining the obstacle removal decision variables and the obstacle removal objective function, and using the gradient descent method to iteratively calculate the local obstacle removal optimal solution for all road clearing vehicles; Based on the optimal solution of all local obstacle removal, complete the Lagrange multiplier exchange of all adjacent road obstacle removal workshops in the optimization game group, and based on the Lagrange multiplier exchange results, complete the iterative update of the optimal solution of all road obstacle removal vehicles to obtain the global optimal solution of obstacle removal; Extract the optimal solution function value of all global optimal solutions for obstacle clearing; Based on the optimal solution function value, the priority of all road clearing vehicles is sorted to obtain the priority of all road clearing vehicles. Target clearing vehicles within the optimized game group are selected based on the priority of clearing vehicles.

[0047] In this embodiment, obstacle clearing decision variables are constructed based on the obstacle clearing function vector and the obstacle clearing closed convex set. These variables include participation capability (0 = cannot participate, 1 = can participate), service modules (including towing, hoisting, road clearing, and emergency dismantling functions, e.g., [1, 0, 2, 1]), and a time module (the time required to arrive). Obstacle clearing constraints include: the quantized value of each element in the obstacle clearing function vector is greater than or equal to a certain percentage of the quantized value of the corresponding element in the obstacle clearing demand vector, e.g., 80%; the total actual travel time is less than or equal to the operation time requirement in the obstacle clearing demand vector; and the total actual travel distance is less than or equal to a certain percentage of the remaining fuel or electric range, e.g., 40%, to prevent the road clearing vehicle from failing to complete the obstacle clearing operation smoothly. If the road clearing vehicle meets all the above obstacle clearing constraints, it is determined that it has passed all obstacle clearing constraints, and its participation capability is set to 1. Next, the partial derivatives of the obstacle removal objective function with respect to each term of the obstacle removal decision variable are calculated to obtain the gradient components. All gradient components are integrated to form a gradient vector. Then, the iteration step size is set, which can be set according to the convergence speed of the obstacle removal objective function, such as 0.01. The obstacle removal decision variable is updated along the negative direction of the subgradient according to the iteration step size. The obstacle removal objective function after updating the obstacle removal decision variable is calculated. After several consecutive iterations, if the rate of change of the obstacle removal objective function is less than the preset rate of change threshold, such as 1%, and the obstacle removal decision variable does not fluctuate significantly, and if the participation capability remains unchanged, the iteration is stopped. The current obstacle removal decision variable and the corresponding obstacle removal objective function value are taken as the local obstacle removal optimal solution for each road clearing vehicle.

[0048] Next, for the local optimal solution of each road clearing vehicle, the Lagrange multipliers corresponding to each clearing constraint are calculated based on the KKT (Kuhn-Tucker) conditions. The Lagrange multipliers are equal to the partial derivatives of the clearing objective function with respect to each clearing constraint. Only non-negative values ​​are retained. A negative multiplier indicates that the clearing constraint has no actual effect and is directly set to 0. The Lagrange multipliers corresponding to all clearing constraints are integrated and normalized to form a Lagrange multiplier vector. Next, Lagrange multiplier exchange is performed on all adjacent road clearing workshops. The main exchange content of Lagrange multiplier exchange is the Lagrange multiplier vector of each road clearing vehicle and the clearing objective function of the local clearing optimal solution. After each road clearing vehicle receives the Lagrange multiplier vector, it will perform integrity verification and value range verification. It will verify whether the length of the Lagrange multiplier vector is consistent with the number of clearing constraints and whether the value range is in the interval [0,1]. If the integrity verification or value range verification fails, it will be regarded as an invalid multiplier and discarded.

[0049] After each road clearing vehicle collects the Lagrange multiplier vectors of all adjacent road clearing vehicles, for each type of clearing constraint, the mean of the corresponding multiplier vectors of all adjacent road clearing vehicles is calculated to reflect the overall level of constraint tightness of the clearing constraints on the road clearing vehicle group. If the multiplier vector value of the road clearing vehicle is much larger than the calculated mean multiplier vector value, for example, if the multiplier vector value of the road clearing vehicle is 1.5 times or more of the mean multiplier vector value, it indicates that the constraint force of this type of road clearing condition is higher, and the cost of participating in the road clearing operation is also higher. It is necessary to lower the vector value of its road clearing decision variable, for example, adjust the service module from [1,0,1,0] to [1,0,0,0] to reduce the workload. If the multiplier vector value of the road clearing vehicle is much smaller than the calculated mean multiplier vector value, for example, if the mean multiplier vector value is 1.5 times or more of the multiplier vector value of the road clearing vehicle, then the vector value of the road clearing decision variable of the road clearing vehicle should be increased, for example, shorten the arrival time to 25 minutes to further reduce the objective function value. Repeat the above steps of multiplier exchange → decision adjustment → optimal solution update until the rate of change of the objective function value of each road clearing vehicle is less than the preset rate of change threshold, such as 1%, and then stop the above steps. Then, based on the final obstacle removal decision variables, they are re-substituted into the obstacle removal objective function, and the local obstacle removal optimal solution is iteratively updated using the gradient descent method described above, to obtain the global obstacle removal optimal solution.

[0050] Next, the optimal solution function value of the global optimal solution for clearing obstacles is extracted. The optimal solution function value refers to the value calculated through the obstacle clearing objective function. The optimal solution function values ​​are arranged in ascending order. Since the smaller the optimal solution function value, the higher the overall cost-effectiveness of dispatching the road clearing vehicle to the road clearing point to perform the clearing operation, the higher the ranking, the higher the priority of the road clearing vehicle. The road clearing vehicle with the highest priority is selected as the target road clearing vehicle.

[0051] By following the steps above, the road clearing vehicle with the highest cost performance and the ability to basically complete the clearing operation can be selected as the target clearing vehicle, providing an equipment foundation for the efficient execution of subsequent clearing tasks.

[0052] In one embodiment, performing multi-scale mode decomposition on the traffic information along the target clearance route in real time, and constructing a traffic fluctuation propagation map along the target clearance route based on the multi-scale mode decomposition results includes the following steps: The patrol grids of all areas overlapping the target clearance route are split into multiple area patrol sub-grids; For any patrol subgrid in a given area, calculate the time-series data of traffic density for the patrol subgrid based on the traffic information from clearing obstacles. An adaptive signal decomposition method is used to perform multi-scale mode decomposition on traffic density time series data to obtain multi-scale traffic signals; Multi-scale fluctuation characteristic parameters of multi-scale traffic signals are extracted, including multi-scale fluctuation amplitude, multi-scale fluctuation phase and multi-scale fluctuation frequency. All regional patrol subgrids are treated as traffic nodes, and all multi-scale fluctuation characteristic parameters are used as node attributes of the corresponding traffic nodes. The traffic complex impedance parameters of all traffic nodes are calculated based on the multi-scale fluctuation characteristic parameters; the coherence analysis algorithm is used to complete the coherence analysis of all multi-scale traffic signals, and the fluctuation coherence coefficients between all adjacent traffic nodes are obtained. Identify the timing of fluctuation events in multi-scale traffic signals based on multi-scale fluctuation characteristic parameters; By combining the wave coherence coefficient, traffic complex impedance parameters, and wave event time, a directed edge for propagation between all adjacent traffic nodes is constructed, resulting in a traffic wave propagation graph.

[0053] In this embodiment, the patrol grids of all overlapping areas along the target clearance route are further refined into multiple regional patrol sub-grids. This ensures that fluctuation signals at each scale can be accurately captured, while simultaneously achieving precise spatial binding of traffic data, laying a spatial foundation for subsequent multi-scale analysis. Furthermore, the regional patrol grid refers to the basic grid unit initially divided based on the target area's road network map to achieve routine patrol coverage by road clearance vehicles; it is the smallest spatial management unit for clearance patrol tasks. The clearance traffic grid refers to the grid unit refined based on road network topology and traffic attributes for the target clearance area formed by integrating multiple regional patrol grids; it is used for clearance route planning and traffic attribute calibration. The clearance route grid refers to the clearance traffic grid that overlaps with the planned initial clearance route; it is the core unit for route trafficability calculation and verification. The aforementioned regional patrol sub-grids are refined grid units obtained by further subdividing the regional patrol grids, used for traffic density time-series data collection and precise extraction of traffic fluctuation characteristics. Its spatial inclusion relationship is as follows: regional patrol subgrid ∈ regional patrol grid, obstacle clearing route grid ∈ obstacle clearing traffic grid, obstacle clearing traffic grid ∈ target obstacle clearing area, and the target obstacle clearing area is formed by the integration of multiple regional patrol grids.

[0054] Next, based on the traffic information from the clearing operation, the average number of vehicles in each patrol sub-grid within each time window is calculated. This average number of vehicles is then divided by the actual travel distance within each patrol sub-grid to obtain the traffic density (unit: vehicles / km). The traffic densities from different time windows are then integrated and arranged chronologically to obtain the traffic density time-series data. This process is repeated to calculate the traffic density time-series data for all patrol sub-grids. Next, multi-scale modal decomposition is performed on the traffic density time-series data to obtain macroscopic, mesoscopic, and microscopic traffic fluctuation signals. These three different scale fluctuation signals are then time-stamped, and missing signals are filled in before being integrated into a multi-scale traffic signal.

[0055] Next, the multi-scale fluctuation characteristic parameters of the multi-scale traffic signal are extracted. These parameters refer to the multi-scale fluctuation amplitude, phase, and frequency of each of the three different-scale fluctuation signals. The multi-scale fluctuation amplitude refers to the maximum deviation of each of the three different-scale fluctuation signals from the mean, reflecting the intensity of the fluctuation. The larger the amplitude, the more severe the fluctuation, and the more significant the impact on the driving safety of the target clearing vehicle. For each scale fluctuation signal in the multi-scale traffic signal, the difference between the peak and trough values ​​in each cycle can be calculated, and this difference can be used as the fluctuation amplitude of the corresponding scale. Multi-scale wave phase refers to the offset of the position of each of the three wave signals at a certain moment relative to the reference moment (the wave's start moment). It reflects the propagation time sequence of the wave. The phase difference can determine the order in which the wave arrives at different nodes. The start moment of each wave signal can be used as the reference phase. By using Fourier transform to convert the time domain signal into a frequency domain signal, the phase value of the wave signal at each sampling moment can be calculated, thus obtaining the wave phase of each of the three wave signals at different scales. Multi-scale wave frequency refers to the number of times each of the three wave signals at different scales completes periodic changes per unit time. It reflects the periodic characteristics of the wave. For each wave signal at different scales, the dominant frequency of the signal can be determined by power spectral density analysis. That is, the frequency component with the highest power, which is the frequency corresponding to the dominant period of the wave, and this frequency can be used as the wave frequency of each of the three wave signals at different scales.

[0056] Next, all patrol subgrids in the region are treated as traffic nodes, and all multi-scale fluctuation characteristic parameters are used as the node attributes of the corresponding traffic nodes. Then, the traffic complex impedance parameters of all traffic nodes are calculated based on the multi-scale fluctuation characteristic parameters. Specifically, the multi-scale fluctuation characteristic parameters are first normalized to eliminate dimensional differences. For example, the mean-standard deviation standardization method can be used to map the multi-scale fluctuation characteristic parameters to the [0,1] interval to obtain standardized fluctuation characteristic parameters, which include standardized fluctuation amplitude, standardized fluctuation phase, and standardized fluctuation frequency.

[0057] The traffic complex impedance parameters for each traffic node consist of traffic complex impedance, magnitude, and argument. The traffic complex impedance is composed of a real part and an imaginary part. The real part is the damping coefficient, which characterizes the road segment's ability to absorb and dissipate disturbance energy. The larger the damping coefficient, the stronger the road segment's ability to dissipate disturbance energy. Its calculation is based on the attenuation rate of the microscale fluctuation signal, combined with corrections to the corresponding traffic physical attributes of the road segment. Specifically, this includes: first, extracting the standard microscale fluctuation amplitude corresponding to the traffic microscale fluctuation signal from the standardized fluctuation amplitude; identifying the attenuation interval after the traffic microscale fluctuation signal reaches its peak; calculating the fluctuation amplitude attenuation rate within this interval; and then correcting it according to the traffic physical attributes corresponding to the traffic node, such as the number of lanes and road width. The entropy weight method can be used to assign weights to each physical attribute and calculate the damping correction factor. The value range of the damping correction factor is [0.8, 1.2]. The more lanes and the wider the road, the larger the damping correction factor, indicating a better foundation for the road segment's dissipation of disturbance energy. Calculate the product of the wave amplitude decay rate and the comprehensive correction factor, normalize the product and map it to the [0,1] interval to obtain the final damping coefficient.

[0058] The core component of the imaginary part of the traffic complex impedance parameter is the inductive reactance parameter, which characterizes the ability of the road segment corresponding to the traffic node to store and accumulate disturbance energy, that is, the elasticity of the road segment under traffic pressure. The larger the inductive reactance parameter, the stronger the ability of the road segment to accumulate disturbance energy. Its calculation is based on the frequency response of the mesoscale fluctuation signal, combined with the correction of the road segment's traffic dynamic attributes. Specifically, the mesoscale fluctuation signal in the time domain is converted into a frequency domain signal using a fast Fourier transform, the amplitude-frequency characteristic curve is plotted, and the amplitude-frequency response value corresponding to the main frequency in the curve is extracted. The larger the amplitude-frequency response value, the more significant the response of the traffic node to the main frequency of the mesoscale fluctuation, and the stronger the energy accumulation ability. Next, the dynamic traffic attributes are corrected. The analytic hierarchy process (AHP) can be used to assign weights to the dynamic traffic attributes such as real-time traffic density, average speed, and traffic flow redundancy, and the inductive reactance correction factor is calculated. The inductive reactance correction factor has a value range of [0.7, 1.3]. The larger the real-time traffic density, the lower the average speed, and the greater the traffic flow redundancy, the larger the inductive reactance correction factor. The product of the inductive reactance correction factor and the amplitude-frequency response value is calculated, and the product is normalized to obtain the inductive reactance parameter.

[0059] The frequency factor of the imaginary part of the traffic complex impedance parameter is the characteristic frequency, representing the core oscillation frequency of traffic fluctuations. Its value matches the dominant frequency of the mesoscale fluctuation signal. The standard mesoscale fluctuation frequency corresponding to the mesoscale fluctuation signal can be extracted from the standardized fluctuation frequency. Then, the frequency with the highest power spectral density is extracted from the standard mesoscale fluctuation frequency as its mesoscale fluctuation dominant frequency. Finally, the multi-scale frequency correction coefficient is calculated by combining the mesoscale fluctuation dominant frequency and the standard macroscopic fluctuation frequency and standard micro fluctuation frequency from the standardized fluctuation frequency. ,in, and These are the standard macroeconomic fluctuation frequencies. and standard micro-wave frequency The correction weights can be determined based on measured traffic flow data and simulations, for example... and They are 0.2 and 0.8 respectively. The mesoscopic wave principal frequency is divided by the standard macroscopic wave frequency and the standard microscopic wave frequency to convert them into relative ratios and eliminate the difference in frequency dimensions across different scales. The product of the mesoscopic wave principal frequency and the multi-scale frequency correction coefficient is calculated and converted into an angular frequency to obtain the characteristic frequency. The damping coefficient, inductive reactance parameter, and characteristic frequency are combined to form a complex impedance, resulting in the traffic negative impedance. , ,in, The damping coefficient is... The imaginary unit, For characteristic frequencies, This is the inductive reactance parameter. Calculate the magnitude of the traffic negative impedance. and argument Calculation formula: , .

[0060] For each pair of adjacent traffic nodes, a coherence analysis algorithm is used to calculate the fluctuating coherence coefficient between them. Commonly used coherence analysis algorithms include Fourier transform-based coherence analysis, wavelet coherence analysis, and cross-spectral coherence analysis. Taking the Fourier transform-based coherence analysis algorithm as an example, the Fourier transform is used to convert the multi-scale traffic signal in the time domain into a frequency domain signal. Then, the cross-power spectral density and self-power spectral density of the two frequency domain signals at different scales are calculated. Next, the coherence coefficients of the two adjacent traffic nodes at different scales are calculated using the coherence formula, resulting in the fluctuating coherence coefficient. The fluctuating coherence coefficient includes macroscopic coherence coefficient, mesoscopic coherence coefficient, and microscopic coherence coefficient. The coherence formula is as follows:

[0061] in, The cross-power spectral density of two multi-scale traffic signals. and These are the self-power spectral densities of two multi-scale traffic signals, respectively.

[0062] Next, for the multi-scale traffic signal at each traffic node, if the fluctuation amplitude at a certain scale suddenly exceeds the corresponding amplitude threshold and the duration exceeds the corresponding time scale threshold (e.g., the time scale threshold for macro-scale fluctuation amplitude is 30 seconds, the time scale threshold for meso-scale fluctuation amplitude is 10 seconds, and the time scale threshold for micro-scale fluctuation amplitude is 5 seconds), then the fluctuation is determined to be an independent fluctuation event and assigned a unique event ID. The amplitude threshold can be determined by retrieving historical traffic density time series data under different traffic conditions. The average fluctuation signal amplitude under different traffic densities is calculated based on the historical traffic density time series data. Then, the average fluctuation signal amplitude with traffic density greater than different safe densities, such as 50 vehicles / km, 30 vehicles / km, and 10 vehicles / km, is used as the amplitude threshold.

[0063] Next, for each independent fluctuation event, its fluctuation start time, fluctuation peak time, and fluctuation decay time are extracted. The fluctuation start time refers to the moment when the fluctuation amplitude first exceeds the corresponding amplitude threshold. The fluctuation peak time refers to the moment when the fluctuation amplitude reaches its maximum value. The fluctuation decay time refers to the moment when the fluctuation amplitude decreases to less than or equal to the corresponding amplitude threshold and persists for multiple sampling moments, such as three sampling moments. Then, similarity calculations are performed on the fluctuation times occurring in each group of adjacent traffic nodes. This can be done by calculating the frequency difference and amplitude difference in the multi-scale fluctuation characteristic parameters of the two events. If both the frequency difference and amplitude difference are less than the corresponding difference threshold (e.g., 10%), the two events are considered the same fluctuation event. If either the frequency difference or amplitude difference is greater than or equal to the corresponding difference threshold (e.g., 10%), the two events are considered different fluctuation events.

[0064] For adjacent traffic node pairs, if the fluctuation coherence coefficient at any scale between them is greater than or equal to a preset coherence coefficient threshold, such as 0.7, and they share the same fluctuation event, then a propagation directed edge is constructed between them. The direction of the propagation directed edge is determined based on the peak fluctuation time in the same fluctuation event, pointing from the traffic node with the smaller peak fluctuation time to the traffic node with the larger peak fluctuation time. The traffic complex impedance parameter is used as a traffic node attribute. Then, the fluctuation propagation coefficient of all propagation directed edges is calculated based on the traffic complex impedance parameter. That is, the fluctuation propagation coefficient of the propagation directed edge is calculated based on the traffic negative impedance magnitude corresponding to the adjacent traffic nodes connected by the propagation directed edge, and this is used as the edge attribute of the propagation directed edge. The formula for calculating the fluctuation propagation coefficient is as follows: , and These represent the negative impedance magnitudes of adjacent traffic nodes. The wave propagation efficiency of adjacent traffic nodes is determined by the node with the higher impedance, which represents the upper limit of the wave propagation bottleneck. Nodes with lower impedance represent the actual passability of wave propagation. A larger ratio of the negative impedance magnitudes of adjacent traffic nodes indicates that the impedances of the two nodes are closer, the bottleneck effect is weaker, and the wave propagation efficiency is higher. A smaller ratio indicates that the obstruction effect of the bottleneck node is more significant, and the wave propagation efficiency is lower. Through the above steps, a traffic wave propagation map can be constructed, which can intuitively present the propagation path, propagation direction, correlation strength, and wave intensity of multi-scale traffic waves in the overlapping area of ​​the target clearing route. This provides core technical support for the subsequent generation of patrol prompts for clearing vehicles and intelligent scheduling.

[0065] In one embodiment, the process of performing multi-scale mode decomposition on traffic density time-series data using an adaptive signal decomposition method to obtain multi-scale traffic signals includes the following steps: Traffic density time-series data is standardized to obtain standard density time-series data. The variational mode decomposition algorithm is used to perform macroscale decomposition of standard density time series data to obtain traffic macro-fluctuation signals and traffic macro-residual terms; The wavelet transform algorithm is used to perform mesoscale decomposition of the macroscopic residual term of traffic, and the mesoscopic fluctuation signal and mesoscopic residual term of traffic are obtained. The traffic mesoscopic residual term is decomposed at a microscale using the empirical mode decomposition algorithm to obtain the traffic micro-fluctuation signal; By integrating macroscopic, mesoscopic, and microscopic traffic fluctuation signals, multi-scale traffic signals are obtained.

[0066] In this embodiment, the traffic density time-series data of each patrol subgrid is first standardized using the mean-standard deviation standardization method to obtain standard density time-series data. Then, the variational mode decomposition algorithm is used to perform macroscale decomposition on the standard density time-series data. The variational mode decomposition algorithm (VMD) has the advantages of anti-mode aliasing and adaptive decomposition. This algorithm separates the macroscale fluctuation signal, namely the traffic macroscale fluctuation signal and the traffic macroscale residual term. The Variational Mode Decomposition (VMD) algorithm can be set to three decomposition levels to ensure the separation of Intrinsic Mode Components (IMFs) corresponding to macroscopic fluctuations, while avoiding signal distortion caused by over-decomposition. The penalty factor can be set to 2000-2500, and the noise tolerance to 0.001. Standard density time-series data is input into the VMD algorithm. Through iterative computation, the original signal is decomposed into three IMFs (IMF1, IMF2, and IMF3) and one residual component. Then, the IMF with the highest energy percentage (i.e., the energy percentage of each IMF to the total energy of all IMFs) is selected as the macroscopic traffic fluctuation signal. This signal mainly reflects the propagation trend and intensity changes of macroscopic congestion and is the core fluctuation affecting the efficiency of target clearing vehicles reaching road clearing points. The residual component is used as the macroscopic traffic residual term.

[0067] Next, the traffic macroscopic residual term is used as input to the wavelet transform algorithm. The db4 wavelet basis can be selected, as it possesses good time-frequency localization characteristics, adapts to the non-stationary characteristics of traffic fluctuation signals, and can accurately extract the local features of mesoscopic fluctuations. The decomposition level can be set to 4 levels, which yields signal components with different resolutions. The traffic macroscopic residual term is input into the wavelet transform algorithm for 4-level wavelet decomposition, resulting in 4 high-frequency components (cA1-cA4) and 1 low-frequency residual component (cD4, i.e., the traffic mesoscopic residual term). The high-frequency components of the 2nd and 3rd levels correspond to the frequency range of the mesoscopic fluctuations. After wavelet reconstruction of the high-frequency components of the 2nd and 3rd levels, the traffic mesoscopic fluctuation signal is obtained. The traffic mesoscopic fluctuation signal mainly reflects the propagation law of mesoscopic disturbances such as the target clearing vehicle entering the traffic flow and local lane congestion, and is the core fluctuation affecting the local safety of the clearing vehicle during its operation. Next, the decomposed low-frequency residual components are subjected to empirical mode decomposition (EMD). The EMD algorithm iteratively selects extreme points, fits the envelope, and calculates the residual, decomposing the low-frequency residual components into multiple intrinsic mode functions (IMFs) and one final residual component. Then, the first two layers of high-frequency IMF components are discarded, as these are considered equipment noise and random disturbances with no real traffic fluctuation significance. The remaining IMF components are merged and reconstructed to obtain the traffic micro-fluctuation signal. This signal primarily reflects speed oscillations during vehicle following and local density fluctuations caused by minor lane changes, representing the core fluctuations affecting the close-range driving safety of the target clearing vehicle.

[0068] Next, using the time axis of the standard density time series data as a reference, the macroscopic, mesoscopic, and microscopic traffic fluctuation signals are time-aligned to ensure that the sampling time and time span of the three signals are completely consistent. Each timestamp corresponds to the fluctuation value of three different scale fluctuation signals (i.e., macroscopic, mesoscopic, and microscopic traffic fluctuation signals). Then, the fluctuation values ​​of the three scales for each timestamp are verified one by one. If there is a gap in the fluctuation value of a certain scale (an abnormal gap generated during the decomposition process), linear interpolation of the fluctuation values ​​of adjacent timestamps at that scale is used to fill it in, ensuring that the integrated signal has no time discontinuities. Finally, the aligned and filled fluctuation signals of the three different scales are integrated according to the structure of timestamp-macroscopic fluctuation value-mesoscopic fluctuation value-microscopic fluctuation value to form a multi-scale traffic signal. This multi-scale traffic signal contains both the independent variation law of each scale fluctuation and reflects the collaborative variation characteristics of fluctuations at different scales, and can completely characterize the traffic fluctuation state within the regional patrol subgrid.

[0069] In one embodiment, reference is made to Figure 3 Generating clearing patrol prompts for target clearing vehicles based on traffic fluctuation propagation maps includes the following steps: S301. Calculate the route stability of the target obstacle removal route based on traffic micro-fluctuation signals and using the Lyapunov exponent calculation algorithm. S302. Calculate the route coupling risk value of the target obstacle removal route based on the time series data of route stability and traffic density; S303. Mark the key points of traffic fluctuations in the traffic fluctuation propagation diagram based on the route coupling risk value; S304. Based on the traffic wave propagation map, complete the phase consistency analysis of all key points of traffic waves, and locate the traffic wave resonance zone of the target clearance route according to the phase consistency analysis results. S305. Perform resonance criticality verification of the traffic fluctuation resonance zone based on multi-scale traffic signals; S306. If the traffic wave resonance zone fails the resonance critical verification, the proportion of wave energy of all traffic nodes in the traffic wave resonance zone shall be calculated based on the multi-scale wave characteristic parameters. S307. Combine the wave energy ratio and the resonance zone coordinates of the traffic wave resonance zone to complete the phase dissolution analysis of the traffic wave resonance zone, and generate a clearing patrol prompt item for the target clearing vehicle based on the phase dissolution analysis results.

[0070] In this embodiment, the moving average method or other sequence denoising methods are used to denoise the traffic micro-fluctuation signal, eliminating random noise interference. Then, the autocorrelation coefficient of the denoised traffic micro-fluctuation signal is calculated. The formula for calculating the autocorrelation coefficient is as follows: Where k is the time delay step, k=1,2,3,...,K, and K is the maximum delay step. Traffic micro-fluctuation signals The mean, Let be the covariance of the traffic micro-fluctuation signal over time delay steps k.

[0071] Next, the delay steps corresponding to when the autocorrelation coefficient is less than or equal to a preset coefficient threshold (e.g., 0.5) are used as the embedding dimension. Then, the mutual information value of the traffic micro-fluctuation signal is calculated. The spurious nearest neighbor method is then used to assist in the verification of the embedding dimension. Specifically, a phase space is first constructed based on the initially determined embedding dimension m. Then, the distance change of the nearest neighbor of any trajectory point in the phase space within the phase space is calculated in the (m+1)-dimensional phase space. Next, the proportion of the distance change to the distance between the trajectory point and its nearest neighbor in the m-dimensional phase space is calculated to obtain the distance change rate. If the distance change rate is less than or equal to a preset change rate threshold (e.g., 10%), it indicates that the embedding dimension m is sufficient and no adjustment is needed. If the distance change rate is greater than the preset change rate threshold (e.g., 10%), the embedding dimension m is incremented by 1, and the auxiliary verification steps are repeated until the embedding dimension is sufficient.

[0072] Next, the mutual information of the traffic micro-fluctuation signals is calculated. The formula for calculating mutual information is as follows: ,in, It is the time delay step. =1,2,3,...,T, where T is the maximum delay step. for The marginal probability density of traffic micro-fluctuation signals, i.e., the frequency of a certain value occurring. for and The joint probability density is the frequency at which two values ​​occur simultaneously.

[0073] By gradually increasing Repeatedly calculate the mutual information of traffic micro-fluctuation signals to obtain a mutual information sequence. Traverse the mutual information sequence and find the delay step corresponding to the first time the mutual information in the sequence reaches a minimum value (global minimum). , indicating at this time and The information independence is the strongest and the redundancy is the lowest, making it the most suitable time delay for phase space reconstruction. Therefore, the delay steps at this point are directly used. As a final time delay.

[0074] Then, based on the calculated embedding dimension m and time delay Phase space reconstruction is performed, converting traffic micro-fluctuation signals into trajectory points in an m-dimensional phase space. Each trajectory point corresponds to a time-state of the nonlinear system. All trajectory points are integrated to generate a phase space trajectory matrix, where each row corresponds to a trajectory point and each column corresponds to a dimension of the phase space. Then, for each trajectory point in the phase space trajectory matrix, the Euclidean distance formula is used to calculate its Euclidean distance to other adjacent trajectory points. The adjacent trajectory point with the smallest Euclidean distance is selected as the nearest neighbor trajectory point. Then, based on the time delay... Define the evolution time step j (j=1,2,...) between the trajectory point and its nearest neighbor trajectory point. For each trajectory point, calculate its Euclidean distance Dj to its nearest neighbor trajectory point after j evolution steps. For each evolution step size j, calculate the mean D(j) of Dj for all trajectory points, forming a distance evolution sequence. Next, perform a logarithmic transformation on the distance evolution sequence, and then use the least squares method to perform linear fitting on the logarithmically transformed distance evolution sequence to obtain the fitted line equation. Calculate the slope of this fitted line equation to obtain the maximum Lyapunov exponent, and use the calculated maximum Lyapunov exponent as the route stability of the target obstacle removal route.

[0075] The average traffic density of the target clearing route is calculated based on time-series traffic density data, i.e., the average density within the analysis period. Then, the average traffic density and route stability are normalized separately and then weighted and summed to obtain the route coupling risk value of the target clearing route. The weights can be determined using the analytic hierarchy process (AHP) or the entropy weight method. Next, all traffic nodes in the traffic fluctuation propagation map are traversed, and traffic nodes with a route coupling risk value greater than a preset risk threshold, such as 0.6, are marked as key points of traffic fluctuation.

[0076] The steps for completing the resonance critical verification of the traffic fluctuation resonance zone based on multi-scale traffic signals include: For any traffic fluctuation key point in the traffic fluctuation propagation map, the multi-scale phase difference of the traffic fluctuation key point is calculated based on the multi-scale fluctuation phase. Phase consistency verification of all key points of traffic fluctuations is completed based on multi-scale phase differences, and phase-consistent nodes are selected based on the phase consistency verification results. Spatial clustering algorithms are used to perform spatial clustering of all phase-consistent nodes, resulting in multiple phase-consistent regions; For any phase-consistent region, the total node disturbance energy of all traffic fluctuation key points is calculated based on the multi-scale fluctuation amplitude, and high-energy nodes are selected from all traffic fluctuation key points based on the total node disturbance energy. Calculate the spatial geometric center of all high-energy nodes, and mark the traffic resonance center of the phase-consistent region based on the spatial geometric center; Starting from the traffic resonance center, the nodes are continuously expanded along the directed edge of the traffic wave propagation graph, and the expanded nodes in the traffic wave propagation graph are marked. The node expansion stops when the ratio of the total disturbance energy of the expanded nodes to the total disturbance energy of the nodes at the traffic resonance center is less than a preset ratio threshold, and the traffic wave resonance zone is output according to the node expansion nodes.

[0077] Specifically, for any key point in the traffic wave propagation map, the multi-scale phase difference of the key point is calculated based on multi-scale phase, including macro-micro phase difference, macro-meso phase difference, and micro-meso phase difference. If at least two of these phase differences are less than or equal to a preset phase difference threshold, such as 10°, the key point is determined to be a phase-consistent node, meaning that traffic waves at different scales are in phase synchronization, satisfying the core condition of constructive interference. All phase-consistent nodes are extracted, and adjacent phase-consistent nodes are aggregated into phase-consistent regions using a spatial clustering algorithm, such as density clustering (DBSCAN). Isolated phase-consistent nodes are then removed, resulting in multiple phase-consistent regions. Next, based on the multi-scale amplitude, the multi-scale disturbance energy of all key points of traffic fluctuations is calculated. Multi-scale disturbance energy = square of the normalized multi-scale amplitude result × (1 - damping coefficient). This is because a larger damping coefficient results in faster local dissipation of disturbance energy, weaker propagation and persistence, and a smaller impact range on surrounding traffic. Multi-scale disturbance energy includes macroscopic, mesoscopic, and microscopic disturbance energy. For phase-consistent nodes, the sum of their multi-scale disturbance energies is calculated to obtain the total node disturbance energy. The total node disturbance energy of all phase-consistent nodes is sorted in descending order, and the top 5% of phase-consistent nodes with the highest energy within the phase-consistent region are extracted as high-energy nodes. Furthermore, fluctuation energy is the total temporal energy of the traffic fluctuation signal itself, an inherent signal attribute of the fluctuation, reflecting the magnitude of the fluctuation's vibrational energy. Disturbance energy in this segment refers to the effective propagation impact energy of the fluctuation at traffic nodes, representing the actual risk attribute of the fluctuation to traffic flow and clearing vehicle movement, reflecting the propagation impact of the fluctuation on the outside. The two have similar but not identical meanings.

[0078] Calculate the spatial geometric center of all high-energy nodes, i.e., their average coordinates. If the phase-consistent node corresponding to this geometric center is also a high-energy node, it is directly marked as the traffic resonance center. If the phase-consistent node corresponding to this geometric center is not a high-energy node, the high-energy node closest to this geometric center is selected as the traffic resonance center. Starting from the resonance center, expand outward along the direction of the traffic network topology, i.e., the directed edge of the traffic wave propagation map. Calculate the ratio of the total node disturbance energy of each expanding node to the total node disturbance energy of the traffic resonance center. When this ratio is less than a preset ratio threshold, such as 0.5, stop expanding outward, forming the initial radiation range. Then, extract the impassable areas, such as construction sections, road interruptions, etc., as well as the areas where non-phase-consistent nodes are concentrated, to obtain the traffic wave resonance zone. In addition, if a high-impedance node is encountered during the outward expansion, the expansion in this direction terminates at the node before the high-impedance node. This is because the high-impedance node will hinder wave propagation and does not need to be included in the resonance zone. A high-impedance node is a node whose complex impedance modulus is greater than or equal to 1.5 times the complex impedance modulus of the traffic resonance center.

[0079] The sum of the multi-scale disturbance energies of all traffic nodes within the traffic fluctuation resonance zone is calculated to obtain the total disturbance energy. This total disturbance energy is then divided by the total number of directed edges propagating within the traffic fluctuation resonance zone and normalized to the [0,2] interval to obtain the resonance energy of the traffic fluctuation resonance zone. If the resonance energy is less than or equal to a preset resonance critical threshold, or if the resonance energy is greater than the preset resonance critical threshold but the duration is less than or equal to a preset time threshold (e.g., 60s), then the subsequent obstacle clearing patrol prompt generation step is not required. If the resonance energy is greater than the preset resonance critical threshold and the duration is greater than the preset time threshold (e.g., 60s), then the traffic fluctuation resonance zone is determined to have failed the resonance criticality check, triggering the active dynamics mitigation mode and initiating the subsequent obstacle clearing patrol prompt generation step. The resonance critical threshold can be calculated by retrieving the historical traffic fluctuation data of the road network in the target area for the past 6 months. During stable operating periods when the average speed of the road segment is ≥60% of the road design speed, there are no records of accidents, construction, vehicle breakdowns, or other obstacles, and the fluctuation range of traffic density time series data is ≤10%, the reference resonance energy can be calculated using the same steps. Then, 1.3-1.5 times the reference resonance energy is used as the resonance critical threshold.

[0080] Next, for each key point of traffic fluctuation, the multi-scale fluctuation characteristic parameters are input into the fluctuation energy calculation formula to calculate the fluctuation energy under macroscopic, mesoscopic, and microscopic fluctuations, thus obtaining the multi-scale fluctuation energy. Multi-scale fluctuation energy refers to the fluctuation energy under three different scale fluctuation signals. The fluctuation energy calculation formula is as follows: ,in, Let be the multi-scale fluctuation amplitude in the multi-scale fluctuation characteristic parameters corresponding to traffic node m. Let be the multi-scale fluctuation frequency in the multi-scale fluctuation characteristic parameters corresponding to traffic node m. The multi-scale fluctuation phase in the multi-scale fluctuation characteristic parameters corresponding to traffic node m, where t is the time point within the preset time window (t0 to T).

[0081] After calculating the wave energy under wave signals of different scales, the percentage of wave energy under wave signals of different scales to the sum of wave energy under all wave signals of all scales is calculated to obtain the wave energy ratio.

[0082] The phase resolution analysis of the traffic wave resonance zone is completed by combining the wave energy ratio and the resonance zone coordinates of the traffic wave resonance zone. Based on the phase resolution analysis results, the following steps are taken to generate a clearing patrol prompt for the target clearing vehicle: The target clearing vehicle is modeled as a point source disturbance, and the traffic fluctuations of the point source disturbance are defined as virtual interference waves; The interference destructive analysis of virtual interferometric waves is completed by combining the wave energy ratio and multi-scale wave characteristic parameters. Based on the interference destructive analysis results, the interference parameters of the reverse interferometric wave in the traffic wave resonance region are determined. The reverse interferometric wave interference parameters include virtual wave phase, virtual wave amplitude and virtual wave frequency. Extract the driving performance parameters of the target recovery vehicle based on the recovery vehicle configuration information; The basic boundary for clearing obstacles of the target clearing vehicle is constructed by combining driving performance parameters and obstacle clearing demand vector; The traffic wave resonance segment of the target obstacle removal route is extracted based on the resonance zone coordinates of the traffic wave resonance zone. The traffic speed control parameters for the traffic fluctuation resonance segment are determined based on the interference parameters of the reverse interference wave and the boundary of the obstacle clearance base. The traffic speed control parameters include the speed control cycle sequence, the speed fluctuation boundary parameters, and the speed control node sequence. Traffic lane change prompts are generated based on the traffic complex impedance parameters of all traffic nodes within the traffic fluctuation resonance segment and the pre-acquired road traffic rules. Traffic speed control parameters and lane change prompts are integrated into obstacle clearing and patrol prompts.

[0083] Specifically, when the target recovery vehicle travels at a constant speed, vehicles behind it will form a local queue due to the difference between the target recovery vehicle's speed and the average road speed. The queue length changes regularly with the target recovery vehicle's travel time. When the target recovery vehicle adjusts its speed or leaves the road segment, the queue will gradually dissipate, forming a complete queue-dissipation cycle. The periodic changes in traffic density and speed within this cycle will form a controllable traffic disturbance wave propagating along the road network. This wave is formally defined as a virtual interference wave, and the target recovery vehicle is modeled as a movable point source disturbance body.

[0084] The node with the highest total disturbance energy within the traffic fluctuation resonance zone and the corresponding phase is selected as the benchmark node. Based on the fluctuation energy ratio, the fluctuation signal with the highest fluctuation energy ratio among different scale fluctuation signals is selected as the dominant fluctuation signal of the benchmark node. The phase of the multi-scale fluctuation at the corresponding scale in the multi-scale fluctuation characteristic parameters is selected as the dominant signal phase of this dominant fluctuation signal. Based on the physical principle of linear wave interference cancellation, maximum amplitude offset and energy cancellation can only be achieved when the phase difference between two fluctuations of the same frequency is π. Therefore, a fixed phase offset of π is superimposed on the dominant signal phase to obtain the anti-phase target phase. Simultaneously, based on the 2π periodicity of the phase, the calculated anti-phase target phase is normalized to the standard phase interval [0, 2π], eliminating phase redundancy values ​​exceeding the period and ensuring that the phase value conforms to the propagation physics of traffic flow fluctuations.

[0085] The virtual interference wave generated by the target recovery vehicle propagates from its real-time location to the resonance center of the resonance zone. Due to the traffic characteristics of the road network along its propagation path, a fixed phase shift occurs. If this shift is not corrected, the virtual wave will deviate from the out-of-phase target when it reaches the resonance zone. First, the propagation path of the virtual wave needs to be determined, perfectly coinciding with the target recovery vehicle's route. The traffic impedance parameters, segment length, and average traffic speed of each traffic node along the propagation path are extracted. These parameters directly determine the propagation speed and phase change rate of the traffic wave in the corresponding segment. Then, the phase shift generated by the virtual wave propagating in each segment is calculated segment by segment. This shift is positively correlated with the segment length and negatively correlated with the traffic wave propagation speed. It is also corrected by the segment damping coefficient; the larger the damping coefficient, the more significant the phase attenuation of the wave, and the larger the corresponding phase shift. Finally, the phase shifts of all segments along the propagation path are summed to obtain the total phase shift of the virtual interference wave propagating from the recovery vehicle's location to the resonance center.

[0086] To ensure that the phase of the virtual phase wave is exactly equal to the phase of the anti-phase target when it reaches the resonance center of the resonance region, the initial output phase of the target clearing vehicle needs to be reversed based on the total phase offset. Subtracting the total phase offset from the anti-phase target phase yields the initial phase of the virtual wave that the target clearing vehicle needs to output at its current position, i.e., the virtual wave phase of the virtual interference wave generated by the target clearing vehicle at that position. Simultaneously, the virtual wave phase is normalized within the range [0, 2π] to ensure the validity of the phase value.

[0087] Next, the multi-scale fluctuation amplitude corresponding to the scale in the multi-scale fluctuation characteristic parameters is used as the dominant signal amplitude of each traffic node within the traffic fluctuation resonance zone. Then, the dominant signal amplitude is substituted into the multi-scale disturbance energy calculation formula, and the product of the square of the normalized multi-scale amplitude and (1 - damping coefficient) is calculated to obtain the disturbance energy of each traffic node. The disturbance energies of all traffic nodes within the traffic fluctuation resonance zone are accumulated to obtain the total dominant energy of the traffic fluctuation resonance zone. Since the disturbance energy is proportional to the square of the dominant signal amplitude of each traffic node, to achieve cancellation of the disturbance energy, it is necessary to calculate the effective amplitude when the virtual interference wave arrives at the traffic resonance center. Based on the total dominant energy of the traffic fluctuation resonance zone, combined with the propagation path energy attenuation rate of the virtual interference wave from the current position of the clearing vehicle to the resonance center (the propagation path energy attenuation rate is calculated by accumulating the damping coefficients of each node along the path), the target effective amplitude of the virtual interference wave arriving at the traffic fluctuation resonance zone is deduced. Next, the travel path of the target clearing vehicle from its current location to the traffic resonance center is extracted from the target clearing route. The traffic complex impedance parameters of each traffic node along this path are then extracted. The single-segment energy attenuation rate of the virtual interference wave at each traffic node is calculated segment by segment. The single-segment attenuation rate is positively correlated with the damping coefficient of the traffic node and the segment length, and negatively correlated with the average traffic speed. To ensure that the effective amplitude of the virtual interference wave reaches the traffic wave resonance zone is exactly equal to the target effective amplitude, the initial amplitude needs to be reversed based on the total energy attenuation rate. The difference between the target effective amplitude and 1 minus the total energy attenuation rate yields the initial amplitude that the target clearing vehicle needs to output at its current location, i.e., the virtual wave amplitude of the virtual interference wave generated by the target clearing vehicle at that location.

[0088] Next, nodes within the traffic wave resonance zone whose complex impedance modulus is less than 1.5 times that of the traffic resonance center are marked as valid traffic nodes. For each valid traffic node, weights are assigned to its multi-scale wave frequencies according to the wave energy proportion; the higher the wave energy proportion, the higher the weight assigned. Then, the weighted sum of its scale wave frequencies is obtained to obtain the effective wave frequency of each valid traffic node. The mean of the effective wave frequencies of all valid traffic nodes is calculated, and then the calculated mean of the effective wave frequencies is used as the virtual wave frequency of the virtual interference wave. This step ensures that the virtual interference wave can simultaneously cancel out waves at multiple scales, avoiding the problem of a single frequency only canceling out a single-scale wave while other scale waves continue to amplify. By integrating the virtual wave phase, virtual wave amplitude, and virtual wave frequency of the virtual interference wave, the reverse interference wave interference parameters of the traffic wave resonance zone are obtained.

[0089] From the configuration information of the recovery vehicle, extract the target recovery vehicle's driving performance parameters, including maximum speed, minimum stable speed, maximum acceleration, maximum deceleration, wheelbase, turning radius, and optimal lane width adaptation range. These parameters serve as the physical performance constraints of the target recovery vehicle, and traffic speed control parameters must not exceed these limits. Based on the longest permissible arrival time within the recovery demand vector, define the time constraint boundary for recovery operations, and determine the upper limit of the total travel time for the target recovery vehicle. All traffic speed control parameters must ensure that the total travel time does not exceed this upper limit, while also allowing time for attitude adjustment along the recovery operation path. Integrate the time constraint boundary for recovery operations and the physical performance constraints of the target recovery vehicle to obtain the basic recovery boundary for the target recovery vehicle.

[0090] Based on the proportion of wave energy, the wave signal with the highest proportion of wave energy among wave signals of different scales is taken as the dominant wave signal of the traffic node. The traffic wave resonance segment is divided into several continuous execution sub-segments according to the boundaries of the traffic nodes. Each execution sub-segment corresponds one-to-one with a traffic node, and each sub-segment is bound to the complex impedance parameters, wave phase, wave amplitude, wave frequency, and other attributes of the corresponding traffic node. Then, taking the current time of the target clearing vehicle as the initial time and the expected arrival time of the target clearing vehicle at the starting boundary of the traffic wave resonance segment as the execution start time, and using the wave period corresponding to the virtual wave frequency as the basic time unit, the total travel time within the traffic wave resonance segment is divided into several continuous control cycles. Each control cycle is bound to the travel time of the corresponding execution sub-segment, ensuring that the time unit of speed control is completely synchronized with the wave period. Among all execution sub-segments, the core sub-segment corresponding to the traffic resonance center is locked and used as the core reference unit for phase matching and speed calibration. The speed parameters of all preceding and following sub-segments are adjusted in conjunction with the phase timing requirements of the core sub-segment.

[0091] The reference fluctuation period T is calculated based on the virtual wave frequency ω, using the formula T = 2π / ω. This reference fluctuation period is the common oscillation period of the virtual interference wave and the dominant fluctuation signal, serving as the core time reference for speed control. Combining the length of each execution sub-segment and the road design speed, the reference travel time for the target clearing vehicle through a single sub-segment is calculated. The reference fluctuation period is then matched with the sub-segment's reference travel time, essentially merging the reference travel times to ensure consistency with the reference fluctuation period. If the sub-segment's travel time and the reference fluctuation period cannot be perfectly matched as an integer, the reference average travel speed of the sub-segment is slightly adjusted to ensure that the sub-segment's travel time matches the fluctuation period as an integer multiple without exceeding the basic clearing boundaries, thus ensuring complete synchronization between the speed control cycle and the fluctuation period. After matching, a continuous speed control cycle sequence for the entire road segment is generated based on the cycle matching results. This sequence defines the start and end times, the covered spatial road segments, and the corresponding dominant fluctuation signal's fluctuation phase interval for each control cycle, thus defining the time and spatial boundaries for subsequent basic speed generation.

[0092] The average real-time traffic flow speed during the traffic fluctuation resonance segment is used as the initial base speed value. The initial base speed value is then corrected using the complex impedance parameters of the traffic nodes covered within the corresponding control period to obtain the damped speed reference value. Specifically, the average damping coefficient and average complex impedance magnitude of all traffic nodes within the control period are first calculated. The damping coefficient characterizes the road segment's ability to absorb and dissipate disturbance energy, while the complex impedance magnitude characterizes the road segment's overall ability to impede wave propagation. The damping coefficient percentage is calculated as: Damping coefficient percentage = Average damping coefficient / Average complex impedance magnitude. Based on this percentage, the initial base speed value is corrected once. A higher damping coefficient percentage indicates a stronger energy dissipation capability of the executing sub-segment, requiring a smaller damping speed correction. A lower percentage indicates a weaker energy dissipation capability of the executing sub-segment, requiring a stronger damping effect from the target clearing vehicle; therefore, the damped speed reference value needs to be appropriately lowered. Next, a second adjustment is made based on the parameters. The inductive reactance parameter characterizes the road segment's ability to store and accumulate disturbance energy. The larger the inductive reactance parameter, the easier it is for the road segment to amplify disturbance energy. Therefore, the damped speed reference value needs to be slightly lowered to strengthen the damping barrier effect. The smaller the inductive reactance parameter, the weaker the road segment's energy storage capacity. The damped speed reference value can be appropriately increased to ensure traffic efficiency. After two adjustments, the final damped speed reference value is obtained. The average damped speed reference value of all executed sub-segments within each speed control cycle is calculated to obtain the periodic damped speed. The periodic damped speeds of all speed control cycles are integrated to obtain the damped speed sequence.

[0093] For each speed control cycle, the virtual wave amplitude is normalized and mapped to the speed fluctuation amplitude, which is the maximum change in speed around the periodically damped speed. The mapping rule is: the larger the virtual wave amplitude, the larger the speed fluctuation amplitude, ensuring that the virtual wave energy is equivalent to the energy of the dominant fluctuation signal. For example, when the virtual wave amplitude is in the range [0, 0.3], the speed fluctuation amplitude = periodically damped speed × 0.1; when the virtual wave amplitude is in the range (0.3, 0.7], the speed fluctuation amplitude = periodically damped speed × 0.2; when the virtual wave amplitude is in the range (0.7, 1.0], the speed fluctuation amplitude = periodically damped speed × 0.3, ensuring that the mapping relationship fits the vibration damping energy matching requirements. Taking the periodically damped speed of each speed control cycle as the center, the initial speed fluctuation range is directly determined by combining the speed fluctuation amplitude calculated above: upper speed limit = periodically damped speed + speed fluctuation amplitude, lower speed limit = periodically damped speed + speed fluctuation amplitude. Damped speed – speed fluctuation amplitude. Based on the speed fluctuation amplitude and the duration of the corresponding speed control cycle, calculate the required acceleration and deceleration. Simultaneously, verify that the absolute values ​​of acceleration and deceleration do not exceed the maximum acceleration / deceleration limits of the clearing vehicle. If they do, slightly reduce the speed fluctuation amplitude, not exceeding 20% ​​of the original amplitude, and recalculate the upper and lower speed limits until the basic clearing boundary for acceleration / deceleration operations is reached, avoiding additional disturbances caused by sudden acceleration / deceleration. After completing all verifications, determine the final upper and lower limits of speed fluctuation for each speed control cycle, and together with the corresponding cycle damping speed, form complete speed fluctuation boundary parameters.

[0094] Extract the complete trajectory of the virtual wave phase from the starting phase to the ending phase for each speed control cycle. Align the phase change axis perfectly with the time axis of the speed control cycle, clarifying the specific times of key phase points such as phase 0, π / 2, π, and 3π / 2 on the time axis, and establishing a one-to-one phase-time mapping relationship. For each traffic node within the traffic wave resonance segment, extract its dominant signal phase and calibrate the π phase difference between it and the virtual wave phase, i.e., the inverse phase time node. Based on the phase-time mapping relationship and the inverse phase nodes, calibrate the speed change sequence within each speed control cycle. Specifically, when the virtual wave phase increases from 0 to π, the target clearing vehicle is scheduled to slightly decelerate from the upper speed limit to the lower speed limit; when the virtual wave phase increases from π to 2π, the target clearing vehicle is scheduled to accelerate from the lower speed limit to the upper speed limit. Key inverse phase nodes correspond to the turning points of speed changes, ensuring that the speed cycle changes are completely synchronized with the virtual wave phase changes and always remain inverse phase with the original disturbance wave. Arrange the inverse phase time nodes of all speed control cycles in chronological order to form a speed control node sequence.

[0095] Based on the target clearing route, all lanes in the traffic fluctuation resonance segment are extracted from the regional road network map. Lanes failing the grid throughput verification are then removed, including those closed for construction, under temporary traffic control, experiencing road interruptions, or exceeding the load / width range of the clearing vehicle. Next, lanes not conforming to traffic regulations are removed, including bus lanes, emergency lanes, and other lanes that the target clearing vehicle cannot occupy except in emergencies, forming a list of passable lanes. For all passable lanes in the list, the arithmetic mean of the damping coefficients of all corresponding traffic nodes is calculated to obtain the lane damping coefficient. Then, all passable lanes are sorted in descending order based on their lane damping coefficients; lanes with higher damping coefficients have higher priority and are included in the candidate list first. By integrating the priority of all passable lanes and the corresponding start and end positions of road segments, traffic lane change prompts are generated. For example, when the target tow truck is about to enter the traffic fluctuation resonance segment, the start and end positions of the highest priority passable lane are pushed to it through voice broadcast on the vehicle terminal and constant highlight on the screen, prompting the driver to drive in that passable lane.

[0096] By integrating traffic speed control parameters and lane change prompts, clearing patrol prompts are generated for the target clearing vehicle. When the target recovery vehicle reaches the boundary of the traffic fluctuation resonance segment, the vehicle terminal will broadcast a voice prompt and display a constantly highlighted screen to the driver, providing traffic lane change prompts and traffic speed control parameters. Simultaneously, the driver will be reminded that a lane change is permitted, for example, "Current position is 150 meters from the lane change starting point. The legal lane change zone is from K2+100 to K2+300 meters ahead. It is recommended to complete the lane change to the second middle lane within this zone." When the corresponding speed control node is reached, the driver will be reminded that speed adjustment is possible, and the corresponding speed fluctuation boundary parameters will be provided, for example, "Acceleration control node is 200 meters ahead. The pulse cycle acceleration segment is about to begin. Please prepare for smooth acceleration." When approaching the speed control node, "Acceleration control node has been reached. Please accelerate smoothly to 65 km / h within 10 seconds. Do not accelerate suddenly during acceleration; maintain a straight lane." When the target recovery vehicle's speed reaches the speed limit threshold, "Current speed is 66 km / h. The speed control zone upper limit has been reached. The deceleration adjustment node is about to begin. It is recommended to immediately and smoothly decelerate to the base speed of 60 km / h."

[0097] Because road traffic conditions change in real time, even though the above steps have matched the most suitable target vehicle for the clearing operation at the target clearing point using various data and planned the optimal clearing route for that vehicle, traffic conditions along the target clearing route are highly variable and difficult to predict directly based on historical traffic data. Therefore, it is necessary to generate clearing patrol prompts to allow the driver of the target clearing vehicle to understand the traffic conditions ahead in real time and make appropriate driving operations, maximizing the chances of the target clearing vehicle reaching the target clearing point smoothly to perform the clearing task. Furthermore, the clearing patrol prompts only serve as driving suggestions during the target clearing vehicle's journey, allowing the operator to know the road conditions ahead in advance and make the best driving operations to prevent traffic accidents that could affect the efficient completion of the clearing operation.

[0098] This application also provides a machine-readable storage medium storing instructions for causing a machine to execute the intelligent patrol control method for a clearing vehicle based on big data analysis according to any one of the preceding claims.

[0099] This application also provides an intelligent patrol and control system for road clearing vehicles based on big data analysis, including: The memory is configured to store instructions; and The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the intelligent patrol control method for clearing vehicles based on big data analysis according to any one of the preceding statements.

[0100] The processor can be a central processing unit (CPU). Of course, depending on the actual use, it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc., and this application does not limit it.

[0101] The memory can be an internal storage unit of a computer device, such as a hard disk or RAM, or an external storage device, such as a plug-in hard disk, smart memory card (SMC), secure digital card (SD), or flash memory card (FC) provided on the computer device. Furthermore, the memory can be a combination of internal storage units and external storage devices of a computer device. The memory is used to store computer programs and other programs and data required by the computer device. The memory can also be used to temporarily store data that has been output or will be output. This application does not limit this.

[0102] This application also provides a machine-readable storage medium storing instructions that cause the machine to execute the above-described intelligent patrol and control method for clearing vehicles based on big data analysis.

[0103] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0108] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0109] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0110] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0111] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for intelligent patrol and control of clearing vehicles based on big data analysis, characterized in that, The method includes the following steps: Obtain the regional road network map of the target area; The target area is divided into multiple patrol grids based on the regional road network map; When road clearing information is received from any road clearing point within any area patrol grid, the area patrol grid and its corresponding adjacent patrol grids are integrated into a target clearing area. All road clearing vehicles within the target clearing area are grouped into an optimal game group, and the configuration and location information of the clearing vehicles in the optimal game group are obtained. Real-time access to traffic information related to clearing obstacles within the target area; By combining the location information of clearing vehicles and clearing traffic information, regional clearing routes are planned for all road clearing vehicles within the optimal game group; Based on regional clearing routes, road clearing information, and clearing vehicle configuration information, a clearing objective function for all road clearing vehicles is constructed, and target clearing vehicles within the optimization game group are selected according to the clearing objective function. Mark the area clearing route corresponding to the target clearing vehicle as the target clearing route; As the target clearing vehicle proceeds to the road clearing point along the target clearing route, the traffic information along the target clearing route is decomposed in real time using multi-scale modal analysis, and a traffic fluctuation propagation map along the regional clearing route is constructed based on the multi-scale modal analysis results. Based on the traffic fluctuation propagation map, a clearing patrol prompt is generated for the target clearing vehicle.

2. The method according to claim 1, characterized in that, The process of combining the location information of clearing vehicles and clearing traffic information to plan regional clearing routes for all road clearing vehicles within the optimal game group includes the following steps: The target clearing area is divided into multiple clearing traffic grids based on the regional road network map; For any traffic clearing grid, the traffic multidimensional attributes of the traffic clearing grid are identified based on the regional road network map and traffic clearing information. These traffic multidimensional attributes include traffic topology attributes, traffic social attributes, traffic dynamic attributes, and traffic physical attributes. Based on the multidimensional attributes of traffic and the location information of the road clearing vehicle, multiple primary clearing routes are planned for the road clearing vehicle, and all primary clearing routes are integrated into a primary route set. For any clearing traffic grid that overlaps in area with any primary clearing route, the clearing traffic grid is marked as a clearing route grid; The grid accessibility of the clearing route grid is calculated based on the multi-dimensional traffic attributes of the clearing route grid. Based on the grid accessibility, all primary clearing routes in the primary route set are screened and optimized to obtain a candidate route set. If the total number of candidate clearance routes in the candidate route set is greater than the preset total threshold, then the clearance operation path segments of all candidate clearance routes in the candidate route set are extracted based on the road clearance points. Determine the obstacle removal driving posture parameters of the road clearing vehicle based on the regional road network map for the obstacle removal operation route; The clearance exit posture parameters of the road clearance vehicle are determined based on the pre-acquired clearance transportation destination and regional road network map; The attitude adjustment cost of the road clearing vehicle is calculated by combining the attitude parameters of the vehicle entering and exiting the clearing area, and the area clearing route of the road clearing vehicle is selected from the candidate route set based on the attitude adjustment cost.

3. The method according to claim 2, characterized in that, The process of calculating the grid accessibility of the clearing route grid based on its multi-dimensional traffic attributes, and then filtering and optimizing all primary clearing routes in the primary route set based on the grid accessibility to obtain a candidate route set includes the following steps: A grid traffic cost function for clearing routes is constructed by combining traffic topology and traffic physical properties. The grid traffic cost function includes travel distance cost, traffic light waiting cost, turning penalty, and gradient penalty. The traffic congestion index and traffic redundancy of the clearing route grid are calculated based on the traffic dynamic attributes, and the traffic risk coefficient of the clearing route grid is calculated by combining the traffic congestion index and traffic redundancy. The traffic sensitivity of the clearing route grid is determined based on the social attributes of traffic. For any clearing route grid, the grid accessibility of all clearing route grids is calculated by combining the grid access cost function, access risk coefficient and access sensitivity, and the accessibility verification of all clearing route grids is completed based on the grid accessibility. If the access verification of the clearing route grid fails, the route change verification of the clearing route grid is completed based on the traffic multidimensional attributes and the access verification results of the corresponding adjacent clearing route grids. If the route changeability check of the clearing route grid fails, then primary clearing routes that overlap with the clearing route grid in the primary route set will be removed. If the route changeability check of the obstacle removal route grid passes, then the primary obstacle removal routes that overlap with the obstacle removal route grid in the primary route set are locally optimized, and the primary obstacle removal routes that have completed local optimization are included in the primary route set to obtain the candidate route set.

4. The method according to claim 2, characterized in that, The process of constructing a clearing objective function for all road clearing vehicles based on regional clearing routes, road clearing information, and clearing vehicle configuration information, and then selecting target clearing vehicles within the optimization game group based on the clearing objective function, includes the following steps: Based on the road clearing information, the clearing needs of the target clearing area are broken down, and a clearing needs vector is constructed. For any road clearing vehicle in the optimization game group, the clearing function vector of the road clearing vehicle is constructed based on the clearing vehicle configuration information and the regional clearing route; The road clearing vehicle's clearance matching degree is calculated by combining the clearance demand vector and the clearance function vector; The clearance cost information of the regional clearance route is calculated based on the multi-dimensional traffic attributes of the clearance traffic grid corresponding to the regional clearance route. By combining the obstacle removal cost information, obstacle removal demand vector, and obstacle removal function vector, an obstacle removal constraint set is defined, and an obstacle removal closed convex set of the road clearing vehicle is constructed based on the obstacle removal constraint set; The objective function for road clearing vehicles is constructed by combining information on clearing matching degree and clearing cost; By combining the closed convex set of obstacle removal and the obstacle removal objective function, the local optimal solution for all road clearing vehicles is calculated iteratively, and the target clearing vehicle in the optimization game group is selected based on the local optimal solution.

5. The method according to claim 4, characterized in that, The process of iteratively calculating the local optimal solution for all road clearing vehicles by combining the closed convex set of clearing and the objective function of clearing, and then selecting the target clearing vehicle within the optimization game group based on the local optimal solution, includes the following steps: The obstacle clearing constraints of all road clearing vehicles in the optimization game group are verified by using the obstacle clearing closed convex set, and the obstacle clearing decision variables are constructed by combining the obstacle clearing constraint verification results and the obstacle clearing function vector. By combining the obstacle removal decision variables and the obstacle removal objective function, and using the gradient descent method to iteratively calculate the local obstacle removal optimal solution for all road clearing vehicles; Based on the optimal solution of all local obstacle removal, complete the Lagrange multiplier exchange of all adjacent road obstacle removal workshops in the optimization game group, and based on the Lagrange multiplier exchange results, complete the iterative update of the optimal solution of all road obstacle removal vehicles to obtain the global optimal solution of obstacle removal; Extract the optimal solution function value of all global optimal solutions for obstacle clearing; Based on the optimal solution function value, the priority of all road clearing vehicles is sorted to obtain the priority of all road clearing vehicles. Target clearing vehicles within the optimized game group are selected based on the priority of clearing vehicles.

6. The method according to claim 1, characterized in that, The process of performing multi-scale modal decomposition on the real-time traffic information along the target clearance route and constructing a traffic fluctuation propagation map along the target clearance route based on the multi-scale modal decomposition results includes the following steps: The patrol grids of all areas overlapping the target clearance route are split into multiple area patrol sub-grids; For any patrol subgrid in a given area, calculate the time-series data of traffic density for the patrol subgrid based on the traffic information from clearing obstacles. An adaptive signal decomposition method is used to perform multi-scale mode decomposition on traffic density time series data to obtain multi-scale traffic signals; Multi-scale fluctuation characteristic parameters of multi-scale traffic signals are extracted, including multi-scale fluctuation amplitude, multi-scale fluctuation phase and multi-scale fluctuation frequency. All regional patrol subgrids are treated as traffic nodes, and all multi-scale fluctuation characteristic parameters are used as node attributes of the corresponding traffic nodes. The traffic complex impedance parameters of all traffic nodes are calculated based on the multi-scale fluctuation characteristic parameters. The coherence analysis algorithm was used to complete the coherence analysis of all multi-scale traffic signals and obtain the fluctuation coherence coefficients between all adjacent traffic nodes. Identify the timing of fluctuation events in multi-scale traffic signals based on multi-scale fluctuation characteristic parameters; By combining the wave coherence coefficient, traffic complex impedance parameters, and wave event time, a directed edge for propagation between all adjacent traffic nodes is constructed, resulting in a traffic wave propagation graph.

7. The method according to claim 6, characterized in that, The method of using adaptive signal decomposition to perform multi-scale mode decomposition on traffic density time-series data to obtain multi-scale traffic signals includes the following steps: Traffic density time-series data is standardized to obtain standard density time-series data. The variational mode decomposition algorithm is used to perform macroscale decomposition of standard density time series data to obtain traffic macro-fluctuation signals and traffic macro-residual terms; The wavelet transform algorithm is used to perform mesoscale decomposition of the macroscopic residual term of traffic, and the mesoscopic fluctuation signal and mesoscopic residual term of traffic are obtained. The traffic mesoscopic residual term is decomposed at a microscale using the empirical mode decomposition algorithm to obtain the traffic micro-fluctuation signal; By integrating macroscopic, mesoscopic, and microscopic traffic fluctuation signals, multi-scale traffic signals are obtained.

8. The method according to claim 7, characterized in that, The process of generating a clearing patrol prompt for the target clearing vehicle based on the traffic fluctuation propagation map includes the following steps: The route stability of the target obstacle removal route is calculated based on traffic micro-fluctuation signals and using the Lyapunov exponent calculation algorithm. The route coupling risk value of the target obstacle removal route is calculated based on the time series data of route stability and traffic density. Based on the route coupling risk value, mark the key points of traffic fluctuations in the traffic fluctuation propagation diagram; Based on the traffic wave propagation map, phase consistency analysis of all key points of traffic waves was completed, and the traffic wave resonance zone of the target clearance route was located based on the phase consistency analysis results. Based on multi-scale traffic signals, the resonance criticality verification of the traffic fluctuation resonance zone is completed; If the traffic fluctuation resonance zone fails the resonance critical test, the proportion of fluctuation energy of all traffic nodes in the traffic fluctuation resonance zone is calculated based on the multi-scale fluctuation characteristic parameters. By combining the wave energy ratio and the resonance zone coordinates of the traffic wave resonance zone, phase resolution analysis of the traffic wave resonance zone is completed, and a clearing patrol prompt item is generated for the target clearing vehicle based on the phase resolution analysis results.

9. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to execute the intelligent patrol control method for clearing vehicles based on big data analysis according to any one of claims 1 to 8.

10. A smart patrol and control system for road clearing vehicles based on big data analysis, characterized in that: include: The memory is configured to store instructions; as well as The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the intelligent patrol and control method for clearing vehicles based on big data analysis according to any one of claims 1 to 8.

Images