Charging pile service guiding method and system based on digital twinning

Abstract

The invention provides a charging pile service guiding method and system based on digital twinning, and relates to the technical field of intelligent charging management.The method comprises the steps that a charging pile parking vehicle dynamic matching graph is constructed by collecting charging pile position topology, vehicle real-time positions and charging pile historical utilization rate data so as to evaluate the usability of charging piles in real time; generating a guiding strategy based on the matching graph, and distributing charging pile resources; calculating an optimal guiding path according to the vehicle position and the distributed charging pile, synchronously sending the optimal guiding path to the charging pile and the user terminal, and guiding the vehicle to run to a target position; when the vehicle enters the preset range of the charging pile to be parked, the position of a charging interface and the optimal parking posture are recognized by fusing the vehicle contour point cloud and the parking track image data; and a fine guiding instruction is generated and sent to the user terminal and the charging pile to assist the vehicle to accurately park and complete service guiding. The charging pile utilization efficiency and the user service experience are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent charging management technology, and in particular to a charging pile service guidance method and system based on digital twins. Background Technology

[0002] With the rapid development of electric vehicles, the demand for charging stations in scenarios such as large parking lots and highway service areas is increasing. Users need to quickly find available charging stations and complete precise parking and charging, while charging station resources also need to be rationally allocated and efficiently utilized. This places higher demands on the intelligence level of service guidance technology.

[0003] Current solutions utilize geomagnetic sensors to detect parking space occupancy, combine this with vehicle location data obtained from user mobile devices, recommend the nearest available charging stations to users based on predetermined allocation rules, and provide two-dimensional planar navigation paths to guide vehicles to their target areas.

[0004] The solution has limitations in terms of precise vehicle positioning and charging pile matching accuracy, and insufficient support for charging interface orientation recognition and parking posture adjustment. This can easily lead to inconvenience and reduced efficiency during actual charging docking. In addition, the resource allocation strategy needs to be more adaptable to dynamically changing environments. Summary of the Invention

[0005] This application provides a charging pile service guidance method and system based on digital twins to solve the problems of low utilization efficiency of charging piles and poor user service experience in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a charging pile service guidance method based on digital twins, comprising:

[0007] Collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area;

[0008] Based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data, a dynamic matching map of charging pile parking spaces and vehicles is constructed to assess the availability of charging piles in real time.

[0009] Based on the dynamic matching map of charging pile parking spaces and vehicles, combined with the digital twin model, a charging pile guidance strategy is generated. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, which is used to allocate charging pile resources.

[0010] Based on the vehicle's real-time location data and the available charging piles, a path planning algorithm is used to calculate the optimal guidance path from the vehicle's current location to the available charging piles. The optimal guidance path information is then sent to the charging piles and the user's mobile terminal to guide the vehicle to the available charging piles.

[0011] When the vehicle drives to the preset range of the charging pile and prepares to park, the vehicle contour point cloud data and parking trajectory image data are acquired, and the vehicle contour point cloud data and the parking trajectory image data are subjected to multi-sensor data fusion processing to identify the vehicle charging interface position and the optimal parking posture.

[0012] Based on the location of the vehicle's charging port and the optimal parking posture, a precise guidance instruction is generated to assist the vehicle in parking the charging station. The precise guidance instruction is then sent to the user's mobile terminal and the charging station to provide service guidance.

[0013] Optionally, the step of performing path characteristic analysis on each candidate path to generate a path length value and a turning complexity value, and calculating the comprehensive cost of each candidate path based on the path length value and the turning complexity value, combined with a deep Q-network model, includes:

[0014] For each candidate path, perform path geometric feature analysis to extract the path node sequence and connecting edge attributes. Based on the path node sequence and connecting edge attributes, calculate the actual travel distance of each connecting edge. By accumulating the actual travel distances of each connecting edge, generate the path length value.

[0015] For each candidate path, perform turning feature analysis, count the number of turns in the candidate path, and calculate the angle change for each turn. Then, perform a weighted calculation on the number of turns and the angle change to obtain the turning complexity value.

[0016] The path length value and the turning complexity value are normalized respectively, and the normalized path length value and the normalized turning complexity value are combined into a standardized path feature vector.

[0017] The standardized path feature vector is input into a deep Q-network model, and path space features are extracted through the multi-layer convolutional network of the deep Q-network model.

[0018] The path space features are evaluated using the fully connected layers of the deep Q-network model, and the comprehensive cost value of each candidate path is output.

[0019] Optionally, the step of generating a charging pile guidance strategy based on the dynamic vehicle matching map of the charging pile parking space and combined with a digital twin model includes:

[0020] The dynamic matching map of the charging pile parking space vehicles is input into the digital twin model. The graph data processing module of the digital twin model processes the dynamic matching map of the charging pile parking space vehicles to obtain charging pile nodes with real-time load coefficients and temporary connection edges with connection weights.

[0021] The load assessment module of the digital twin model is used to screen charging pile nodes whose real-time load coefficient is less than a preset coefficient threshold to form a candidate charging pile set.

[0022] The digital twin model's weight filtering module selects charging pile nodes with connection weights greater than a preset weight threshold from the candidate charging pile set, and determines them as target charging piles.

[0023] The spatial analysis module of the digital twin model extracts the spatial coordinates of the parking space node and the position coordinates of the dynamic vehicle node corresponding to the target charging pile, and calculates the path reachability parameters between the position coordinates of the dynamic vehicle node and the spatial coordinates of the parking space node.

[0024] The decision generation module of the digital twin model generates a dynamic ranking result of the matching priority from the vehicle node to the target charging pile based on the path reachability parameters.

[0025] The optimal charging pile allocation scheme is determined by the strategy output module of the digital twin model based on the matching priority ranking result, and a charging pile guidance strategy is formed based on the optimal charging pile allocation scheme.

[0026] Optionally, the step of constructing a dynamic matching map of charging pile parking spaces based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data includes:

[0027] Establish charging pile nodes containing the location coordinates of charging piles and berth nodes containing the spatial coordinates of berths;

[0028] Based on the charging pile location topology data, a fixed connection edge is established between the charging pile node and the berth node;

[0029] The real-time vehicle location data is converted into dynamic vehicle nodes with timestamps, and the location coordinates of the dynamic vehicle nodes are updated in real time.

[0030] Based on the historical usage data of the charging piles, the real-time load factor of each charging pile node is calculated;

[0031] Based on the spatial proximity relationship between the updated position coordinates of the dynamic vehicle node and the spatial coordinates of the berth node, a temporary connection edge is established between the dynamic vehicle node and the berth node.

[0032] Using the real-time load coefficient of the charging pile node, the temporary connection edge is weighted to obtain a temporary connection edge with connection weight.

[0033] A dynamic matching graph for charging piles, parking spaces, and vehicles is constructed by integrating charging pile nodes with real-time load coefficients, parking space nodes, dynamic vehicle nodes, fixed connection edges, and temporary connection edges with connection weights.

[0034] Optionally, the step of performing multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging port location and optimal parking posture includes:

[0035] The vehicle contour point cloud data is used to perform three-dimensional structure reconstruction to generate a three-dimensional model of the vehicle surface.

[0036] Motion trajectory extraction is performed on the parking trajectory image data to obtain a vehicle pose change sequence;

[0037] The three-dimensional model of the vehicle surface and the vehicle pose change sequence are input into a multimodal fusion neural network for spatiotemporal correlation to generate a three-dimensional pose sequence during vehicle motion.

[0038] In the three-dimensional model of the vehicle surface, the characteristic areas of the charging interface are identified to determine the location of the vehicle charging interface;

[0039] Based on the three-dimensional posture sequence, the relative positional relationship between the vehicle and the charging pile is analyzed. Based on the relative positional relationship, the target docking posture of the charging interface corresponding to the charging pile socket is calculated using a deep reinforcement learning model.

[0040] The optimal parking posture is determined based on the location of the vehicle charging port and the target docking posture.

[0041] Optionally, the step of generating precise guidance instructions based on the vehicle's charging port location and optimal parking posture to assist the vehicle in parking the charging station space, and sending the precise guidance instructions to the user's mobile terminal and the charging station, includes:

[0042] Calculate the horizontal angle and vertical distance between the vehicle charging interface and the charging pile interface, and generate a lateral correction command based on the horizontal angle and the vertical distance;

[0043] Based on the optimal parking posture, the vehicle's forward direction and rotation angle that need to be adjusted are determined, and longitudinal control commands are generated according to the forward direction and rotation angle.

[0044] By combining the lateral correction command and the longitudinal control command, a fine guidance command is formed;

[0045] Based on the precise guidance instructions, visual prompts and device control signals are generated. The visual prompts are sent to the user's mobile terminal for graphical display, and the device control signals are sent to the charging pile to control the charging pile to prepare for charging service.

[0046] Secondly, this application provides a charging pile service guidance system based on digital twins, comprising:

[0047] The data acquisition module is used to collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area.

[0048] The module is used to construct a dynamic matching map of charging pile parking spaces and vehicles based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data, so as to evaluate the availability of charging piles in real time.

[0049] The generation module is used to generate a charging pile guidance strategy based on the dynamic matching map of the charging pile parking space and the digital twin model. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, which is used to allocate charging pile resources.

[0050] The calculation module is used to calculate the optimal guidance path from the vehicle's current location to the available charging pile based on the vehicle's real-time location data and the available charging pile using a path planning algorithm, and send the optimal guidance path information to the charging pile and the user's mobile terminal respectively to guide the vehicle to the available charging pile.

[0051] The acquisition module is used to acquire vehicle contour point cloud data and parking trajectory image data when the vehicle drives to the preset range of the charging pile and prepares to park, and to perform multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging interface position and the optimal parking posture.

[0052] The sending module is used to generate fine guidance instructions based on the location of the vehicle's charging interface and the optimal parking posture to assist the vehicle in parking in the charging pile space, and send the fine guidance instructions to the user's mobile terminal and the charging pile to realize service guidance.

[0053] Thirdly, this application provides an electronic device, comprising:

[0054] Memory, used to store computer programs;

[0055] A processor, configured to execute the computer program to implement the steps of the digital twin-based charging pile service guidance method as described in the first aspect above.

[0056] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the digital twin-based charging pile service guidance method described in the first aspect above.

[0057] The technical solution provided in this application has the following beneficial effects:

[0058] This application establishes a complete charging service environment information database through multi-dimensional data collection, providing a data foundation for subsequent analysis and decision-making. It enables visualized management and real-time status monitoring of charging pile resources, providing an intuitive basis for optimized resource allocation. Intelligent decision-making facilitates the rational allocation of charging pile resources, improving resource utilization efficiency. It provides users with accurate driving guidance, reducing the time spent searching for charging piles. It accurately identifies vehicle characteristics and movement status, providing data support for precise parking. Finally, it achieves precise vehicle-to-charging-pile docking, enhancing the user charging experience.

[0059] Furthermore, this application also obtains the current location of the vehicle and the location of the target charging pile, establishes a path search area and extracts environmental data; then constructs a path topology map, calculates multiple candidate paths considering the distribution of obstacles; analyzes the length and turning complexity of each path, and calculates the comprehensive cost value in combination with an intelligent evaluation model; finally, selects the optimal path as the guiding path.

[0060] Furthermore, the solution can plan the optimal route based on real-time environmental data, taking into account both path length and ease of passage, effectively avoiding obstacle areas, providing safe and efficient navigation guidance, and improving the convenience of users finding charging stations and driving safety.

[0061] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 A flowchart illustrating a charging pile service guidance method based on digital twins, provided for an embodiment of this application;

[0064] Figure 2 A schematic diagram illustrating a specific implementation of a charging pile service guidance method based on digital twins, provided in this application embodiment;

[0065] Figure 3 This is a schematic diagram of a digital twin-based charging pile service guidance system provided in an embodiment of this application. Detailed Implementation

[0066] To address the problems of existing technologies, this application proposes a charging pile service guidance method based on digital twins. This method constructs a dynamic matching map of charging piles, parking spaces, and vehicles, simulating resource distribution and vehicle status in real time within a virtual environment. Based on multi-sensor fusion technology, it accurately identifies the vehicle's charging interface location and parking posture, generating phased guidance instructions from the driving path to the parking action. This solution achieves dynamic optimization of charging pile resource allocation and precise vehicle guidance, effectively improving the overall efficiency of charging services and the convenience of user operation, fundamentally solving the shortcomings of existing solutions in matching accuracy and process guidance.

[0067] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0068] The core of this application is to provide a charging pile service guidance method based on digital twins, and a flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0069] Step 101: Collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area.

[0070] In step 101, the charging pile location topology data represents the spatial connection relationship between the charging pile and the parking space. The vehicle refers to the vehicle waiting to be charged. This method collects its real-time location data, establishes nodes for it in a dynamic matching map, plans its path to the charging pile, and assists it in completing parking docking. The service object and goal of the entire process is to guide vehicles needing charging through the entire process from finding the charging pile to parking. The real-time vehicle location data is the vehicle's dynamic coordinates obtained through the positioning module. The historical usage rate data of the charging pile is a statistical analysis of the usage frequency of the charging pile within a historical time period.

[0071] In this embodiment of the application, the installation location and connection relationship of the charging piles are collected by a sensor network deployed in the parking lot. At the same time, the real-time location is obtained by receiving vehicle satellite positioning signals, and the historical usage records of the charging piles are retrieved from the server database. These data are integrated into a standardized format that the system can process.

[0072] For example, in a large parking lot system, the location coordinates of 8 charging piles and the information of 12 parking spaces they are connected to were collected, the real-time location coordinates of 3 electric vehicles were obtained, and the usage time data of each charging pile in the past 7 days were retrieved from the database. Among them, the historical usage rate of charging pile C4 was 30%.

[0073] Step 102: Based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data, construct a dynamic matching map of charging pile parking spaces for vehicles to assess the availability of charging piles in real time.

[0074] In step 102, the dynamic matching diagram of charging pile parking spaces and vehicles is a graph structure, in which the nodes include charging pile nodes representing the location of charging equipment, parking space nodes representing the location of parking space, and vehicle nodes representing the real-time location of vehicles to be charged. The edges include fixed connection edges representing the fixed affiliation between charging piles and parking spaces, and temporary connection edges representing the dynamic association between vehicles and parking spaces.

[0075] In this embodiment, a graph structure framework is established. Charging pile nodes and parking space nodes are created based on the charging pile location topology data, and fixed connection edges are established. Real-time vehicle location data is converted into vehicle nodes with timestamps. The load coefficient of the charging pile nodes is calculated based on historical usage data. Temporary connection edges are established based on the spatial distance between the vehicle nodes and the parking space nodes and assigned weights. Finally, all nodes and edges are integrated to construct a complete dynamic matching graph.

[0076] For example, based on the data from step 101, a graph structure containing 8 charging pile nodes and 12 parking space nodes is established in the digital twin system. A fixed connection edge is established between charging pile C4 and parking spaces P5, P6, and P7. The real-time positions of the three vehicles are converted into vehicle nodes V1, V2, and V3. The load factor of C4 is calculated to be 0.3. A temporary connection edge is established based on the distance of 12.2 meters between V1 and P5 and assigned a weight of 0.63, thus completing the construction of the dynamic matching graph.

[0077] Step 103: Based on the dynamic matching map of the charging pile parking spaces and combined with the digital twin model, generate a charging pile guidance strategy. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, which is used to allocate charging pile resources.

[0078] In step 103, the charging pile guidance strategy is a resource allocation scheme generated by the system, which includes the matching relationship between available charging piles and vehicles, and is used to guide the allocation of charging services.

[0079] In this embodiment, a dynamic matching graph is loaded into the digital twin model, node and edge information is extracted by the graph data processing module, a load assessment module is used to screen low-load charging piles to form a candidate set, a weight screening module is used to select charging piles with high weight connections as target charging piles, a spatial analysis module is used to calculate path reachability parameters, and finally a decision generation module is used to generate a matching priority ranking and form a guidance strategy.

[0080] For example, from the dynamic matching graph, charging piles C4, C6, C7, and C8 with a load factor lower than 0.5 are identified to form a candidate set. C4 with a connection weight greater than 0.8 is selected as the target charging pile. The path reachability parameter from V1 to C4 is calculated to be 111.25. After generating the matching priority ranking, the guidance strategy for C4 to be assigned to V1 is determined.

[0081] Step 104: Based on the real-time vehicle location data and the available charging piles, calculate the optimal guidance path from the vehicle's current location to the available charging piles using a path planning algorithm, and send the optimal guidance path information to the charging piles and the user's mobile terminal respectively, so as to guide the vehicle to the available charging piles.

[0082] In step 104, the vehicle's current position refers to its instantaneous coordinates in physical space during path planning; this position is obtained through precise vehicle positioning results output from multi-sensor fusion processing. An available charging station refers to a charging station that is currently idle, whose spatial location matches the vehicle's parking posture, and capable of providing charging services; this is determined by integrating historical charging station usage data, dynamic vehicle matching maps of charging station parking spaces, and precise vehicle positioning results within a digital twin environment for status assessment and filtering. The optimal guidance path is the navigation route from the vehicle's current position to the target charging station, taking into account both path length and traffic complexity.

[0083] In this embodiment, the current location coordinates of the vehicle and the coordinates of the target charging pile are obtained, a path search area is established and channel network and obstacle data are extracted, a path topology map is constructed, and a path planning algorithm is used to calculate multiple candidate paths. The path length and turning complexity of each path are analyzed, and the comprehensive cost value is calculated by combining a deep Q-network model. The optimal path is selected and converted into navigation commands for transmission.

[0084] For example, the current location of V1 (10580) and the coordinates of the target charging pile C4 (135210) are obtained. After establishing the search area, 12 channels and 3 obstacles are extracted. A topology graph containing 86 nodes and 124 edges is constructed. Five candidate paths are calculated. Path 1 has a length of 285 meters and a turning complexity of 5.8. After evaluation by a deep Q network, the cost value is 0.68. Path 1 is selected as the optimal path and navigation instructions are sent.

[0085] Step 105: When the vehicle is driven to the preset range of the charging pile and is ready to park, acquire vehicle contour point cloud data and parking trajectory image data, and perform multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging interface position and the optimal parking posture.

[0086] In step 105, the preset range refers to a range with a radius of R centered on the charging pile. The vehicle contour point cloud data is a set of three-dimensional coordinates of the vehicle surface collected by LiDAR, and the parking trajectory image data is a sequence of vehicle motion images collected by the camera.

[0087] In this embodiment, point cloud data of the vehicle surface is obtained by LiDAR scanning and a three-dimensional model is reconstructed. The vehicle pose change is extracted by capturing a continuous image sequence by a camera. The three-dimensional model and the pose sequence are input into a multimodal fusion neural network for spatiotemporal correlation to generate a three-dimensional pose sequence. The charging interface area is identified in the model to determine the position. The relative positional relationship between the vehicle and the charging pile is analyzed. The target docking posture is calculated using a deep reinforcement learning model.

[0088] For example, when the vehicle enters the preset range of C4, the LiDAR collects 12,560 point cloud data to reconstruct a 3D model, the camera collects 30 frames of images to extract the pose sequence, and generates a 3D pose sequence through a multimodal fusion neural network. It identifies that the charging port is located on the left rear side, 0.85 meters above the ground and 0.35 meters from the tail. It analyzes that the relative position of the vehicle and C4 is 2.5 meters and 15 degrees apart. The deep reinforcement learning model outputs the target docking posture of moving 12 degrees to the right and 1.8 meters forward.

[0089] Step 106: Based on the location of the vehicle's charging interface and the optimal parking posture, generate a fine-grained guidance instruction to assist the vehicle in parking the charging station space, and send the fine-grained guidance instruction to the user's mobile terminal and the charging station to achieve service guidance.

[0090] In step 106, the fine guidance instructions are a set of operational instructions that guide the vehicle to complete precise parking, including lateral correction and longitudinal control instructions.

[0091] In this embodiment, a lateral correction command is generated by calculating the horizontal deflection angle and vertical distance from the charging pile based on the location of the charging interface. A longitudinal control command is generated by determining the forward direction and rotation angle based on the optimal parking posture. These commands are integrated to form a complete guidance command sequence, which is then converted into visual prompts and device control signals and sent to the user terminal and the charging pile, respectively.

[0092] For example, based on the location of the charging interface, a horizontal deviation angle of 12 degrees and a vertical distance of 0.02 meters are calculated. A lateral correction command of turning the steering wheel 12 degrees to the right is generated. A longitudinal control command of advancing 1.8 meters is generated based on the target docking posture. These are integrated into a three-step guidance sequence, converted into a graphic prompt and sent to the user's mobile phone. At the same time, a preparation command is sent to the charging pile C4.

[0093] This method achieves intelligent allocation of charging pile resources by constructing a dynamic matching map in a digital twin environment. Combined with multi-sensor data fusion technology, it provides precise guidance throughout the entire process from finding a charging pile to parking, effectively improving the utilization efficiency and service quality of charging piles and enhancing the user's charging experience.

[0094] To address the issues of existing charging guidance schemes having limited route planning and insufficient consideration of actual traffic convenience, some embodiments include step 104: calculating the optimal guidance path from the vehicle's current location to an available charging station using a path planning algorithm based on the vehicle's real-time location data and the available charging stations. Figure 2 As shown, it includes:

[0095] Step 201: Obtain the current coordinate point corresponding to the real-time location data of the vehicle and the target parking space coordinate point corresponding to the available charging pile.

[0096] In step 201, the current coordinate point is the specific location point corresponding to the vehicle's real-time location data, and the target parking space coordinate point is the specific parking location point corresponding to the available charging pile.

[0097] In this embodiment, the system extracts the vehicle's current precise location information from the real-time data stream, and simultaneously obtains the parking space location information of the allocated charging piles, providing accurate start and end coordinates for route planning.

[0098] Step 202: Establish a path search area with the current coordinate point as the starting point and the target berth coordinate point as the ending point, and extract channel network data and obstacle distribution data from the path search area.

[0099] In step 202, the path search area is a spatial range defined by the current coordinate point and the target berth coordinate point. The channel network data is the road connection information that is passable within this area, and the obstacle distribution data is the location information of fixed or temporary obstacles that affect passage within this area.

[0100] In this embodiment of the application, the system automatically defines the search range based on the coordinates of the starting point and the ending point, extracts the topological relationship of all feasible roads by analyzing the electronic map data of the area, and identifies and marks the spatial distribution of various obstacles.

[0101] Step 203: Based on the channel network data, construct a path topology graph containing path nodes and connecting edges. In the path topology graph, based on the obstacle distribution data, use a path planning algorithm to calculate multiple candidate paths from the starting point to the ending point.

[0102] In step 203, the path topology graph is a network graph composed of path nodes and connecting edges. Path nodes represent road intersections or feature points, and connecting edges represent passable road segments between nodes.

[0103] In this embodiment, the system constructs a complete path network structure based on channel network data. Taking into account the distribution of obstacles, the system uses a path search algorithm to calculate multiple feasible routes from the starting point to the ending point in the topology map as candidate paths.

[0104] Step 204: Perform path characteristic analysis on each candidate path to generate path length and turning complexity values. Based on the path length and turning complexity values, and combined with the deep Q-network model, calculate the comprehensive cost of each candidate path.

[0105] In step 204, the path length value is a measure of the total length of the candidate path, the turning complexity value is a measure of the complexity of turning operations in the path, and the comprehensive cost value is a quantitative score of the overall quality of the path through an intelligent evaluation model.

[0106] In this embodiment, the system performs geometric feature analysis on each candidate path, calculates the total path length and turning complexity, inputs these feature parameters into a trained deep Q-network model, and outputs a comprehensive score for each path through the model's value evaluation network.

[0107] Step 205: Select the candidate path whose comprehensive cost value is less than the preset cost threshold as the optimal guiding path.

[0108] In step 205, the preset cost threshold is the path quality qualification standard set by the system.

[0109] In this embodiment of the application, the system compares the comprehensive cost of each candidate path with a preset threshold and selects the path with the best score as the final guidance route.

[0110] Here is a specific example:

[0111] After determining, based on the aforementioned embodiment, to assign charging pile C4 to vehicle V1 and establish a guidance task from V1's current location coordinates [105, 80] to the target parking space coordinates [135, 210], the system establishes a path search area with two points as the start and end points. From this area, channel network data of 12 channels and obstacle distribution data of 3 obstacles are extracted. Based on the channel network data, a path topology map containing 86 path nodes and 124 connecting edges is constructed. Considering the obstacle distribution, the A* path planning algorithm is used to calculate 5 candidate paths. When performing path characteristic analysis on the first candidate path, the actual length of its 18 connecting edges is accumulated to obtain a path length value of 285 meters. At the same time, 7 turning points in the path are identified and the turning angle value of each is calculated. The formula for turning complexity is equal to the number of turns multiplied by the weight coefficient k1 plus the average turning angle. The angle is multiplied by the weight coefficient k2, where k1 is 0.4 and k2 is 0.1. The seven turning angles are 90 degrees, 45 degrees, 90 degrees, 135 degrees, 90 degrees, 45 degrees, and 90 degrees, with an average turning angle of 85 degrees. Substituting these values ​​into the formula, the turning complexity is calculated as 7 × 0.4 + 85 × 0.1 = 2.8 + 8.5 = 11.3. The path length of 285 meters and the turning complexity of 11.3 are normalized and combined into a standardized path feature vector [0.72, 0.65]. This vector is then input into a pre-trained deep Q-network model, resulting in a comprehensive cost value of 0.68. The cost values ​​of the remaining four paths are calculated in the same way: 0.71, 0.75, 0.82, and 0.79, respectively. The first candidate path with a cost value of 0.68, which is less than the preset cost threshold of 0.70, is selected as the optimal guiding path, thus completing the path planning from the vehicle's current location to charging station C4.

[0112] In this embodiment of the application, the route planning scheme can generate an optimal route that takes into account both path length and accessibility based on actual environmental characteristics, effectively avoids obstacle areas, provides safe and efficient navigation guidance, and improves the convenience of users finding charging stations and driving comfort.

[0113] To further improve the accuracy and practicality of path evaluation, in some embodiments, step 204 involves: performing path characteristic analysis on each candidate path to generate a path length value and a turning complexity value; and calculating the comprehensive cost of each candidate path based on the path length value and the turning complexity value, combined with a deep Q-network model, including:

[0114] Step 301: Perform path geometric feature analysis on each candidate path to extract the path node sequence and connecting edge attributes, and calculate the actual travel distance of each connecting edge based on the path node sequence and connecting edge attributes. Generate the path length value by accumulating the actual travel distances of each connecting edge.

[0115] In step 301, the path node sequence is an ordered arrangement of the path nodes contained in the candidate path, and the connection edge attribute is the actual travel distance parameter of the connection edge.

[0116] In this embodiment of the application, the system extracts the path node sequence that constitutes the path according to the passage order of the candidate path, obtains the actual passage distance attribute of the connecting edge between adjacent nodes in the sequence, and accumulates the actual passage distance of these connecting edges in sequence to obtain the path length value that reflects the total length of the path.

[0117] Step 302: Perform turning feature analysis on each candidate path, count the number of turns in the candidate path, calculate the angle change for each turn, and perform a weighted calculation on the number of turns and the angle change to obtain the turning complexity value.

[0118] In step 302, the number of turns is the number of times the vehicle needs to change its direction of travel in the candidate path, and the angle change is the direction angle value that the vehicle needs to adjust for each turn.

[0119] In this embodiment of the application, the system identifies the turning point position by analyzing the directional changes of adjacent connecting edges in the path node sequence, counts the number of turning points to obtain the number of turns, calculates the directional angle between adjacent connecting edges at each turning point as the angle change, and performs a weighted summation of the number of turns and the angle change to obtain the turning complexity value.

[0120] Step 303: Normalize the path length value and the turning complexity value respectively, and combine the normalized path length value and the normalized turning complexity value into a standardized path feature vector.

[0121] In step 303, the standardized path feature vector is a multidimensional data representation composed of normalized path features.

[0122] In this embodiment, the system converts the path length value and the turning complexity value into values ​​in the range of zero to one using a normalization algorithm, and combines these two normalized values ​​to form a standardized path feature vector containing the core features of the path.

[0123] Step 304: Input the standardized path feature vector into the deep Q-network model, and extract path spatial features through the multi-layer convolutional network of the deep Q-network model.

[0124] In step 304, the path spatial features are high-dimensional feature representations that reflect the geometric shape and spatial distribution patterns of the path.

[0125] In this embodiment, the system inputs the standardized path feature vector into a deep Q-network model, and performs nonlinear transformation and feature extraction on the input features through a multi-layer convolutional network in the model to obtain path space features that better reflect the essential characteristics of the path.

[0126] Step 305: Evaluate the value of the path space features through the fully connected layer of the deep Q-network model, and output the comprehensive cost value of each candidate path.

[0127] In this embodiment, the system inputs path space features into the fully connected layer of the deep Q-network model, and outputs a cost value reflecting the overall quality of the path through nonlinear calculation and weight analysis of the fully connected layer.

[0128] Here is a specific example:

[0129] Based on the five candidate paths obtained in the aforementioned embodiments, the system performs geometric feature analysis on the first candidate path, extracting the 18 path node sequences N1 to N18 and 17 connecting edge attributes. Based on the actual travel distance data in the connecting edge attributes, the path length of 285 meters is obtained by accumulating the actual travel distances of the 17 connecting edges. Next, the system performs turning feature analysis, identifying seven turning points by comparing the direction vectors of adjacent connecting edges. The number of turns is counted as seven, and the angle changes for each turn are calculated to be 90 degrees, 45 degrees, 90 degrees, 135 degrees, 90 degrees, 45 degrees, and 90 degrees respectively. The turning complexity is calculated using the formula: number of turns multiplied by weight coefficient k1 plus average turning angle multiplied by weight coefficient k2, where k1 is a dimensionless coefficient of 0.4 and k2 is a dimensionless coefficient of 0.1. The number of turns is seven, and the average turning angle is 85 degrees. Substituting the values ​​into the formula, the turning complexity value is calculated as 7 multiplied by 0.4 plus 85 multiplied by 0.1, which equals 2.8 plus 8.5, which equals 11.3. The path length value of 285 meters and the turning complexity value of 11.3 are normalized. Using the maximum-minimum normalization method, the path length value is converted to 0.72 in the range of 0 to 1, and the turning complexity value is converted to 0.65 in the range of 0 to 1. These two normalized values ​​are combined to form a standardized path feature vector [0.72, 0.65]. This vector is input into a deep Q-network model, and the path space features are extracted through the multi-layer convolutional network of the model. Then, the value of these features is evaluated through a fully connected layer. Finally, the comprehensive cost value of the candidate path is output as 0.68. The other four candidate paths are processed in the same way, and the comprehensive cost values ​​are 0.71, 0.75, 0.82 and 0.79, respectively, completing the comprehensive evaluation of all candidate paths.

[0130] In this embodiment of the application, the path evaluation scheme can comprehensively consider multi-dimensional features such as path length and turning complexity, accurately quantify the overall quality of the path through an intelligent evaluation model, provide a scientific basis for selecting the optimal guiding path, and effectively improve the rationality and practicality of path planning.

[0131] To further improve the intelligence level of charging pile resource allocation, in some embodiments, step 103: generating a charging pile guidance strategy based on the charging pile parking space vehicle dynamic matching map and combined with a digital twin model includes:

[0132] Step 401: Input the dynamic matching map of the charging pile parking space vehicle into the digital twin model. Process the dynamic matching map of the charging pile parking space vehicle through the graph data processing module of the digital twin model to obtain charging pile nodes with real-time load coefficients and temporary connection edges with connection weights.

[0133] In this embodiment, the system inputs the dynamic matching map of charging pile parking spaces and vehicles into the digital twin model, and uses the graph data processing module to parse the node and edge information in the graph, extracting the real-time load coefficient data attached to the charging pile node and the connection weight data attached to the temporary connection edge.

[0134] Step 402: Using the load assessment module of the digital twin model, select charging pile nodes whose real-time load coefficient is less than a preset coefficient threshold to form a candidate charging pile set.

[0135] In step 402, the preset coefficient threshold is the load level qualification standard set by the system, and the candidate charging pile set is the group of available charging piles that have been preliminarily screened.

[0136] In this embodiment of the application, the system compares the real-time load coefficient of each charging pile node with the preset coefficient threshold through the load evaluation module, and selects the charging pile nodes whose load levels meet the requirements to form a candidate set.

[0137] Step 403: Using the weight filtering module of the digital twin model, select charging pile nodes with connection weights greater than a preset weight threshold from the candidate charging pile set, and determine them as target charging piles.

[0138] In step 403, the preset weight threshold is the connection strength qualification standard set by the system, and the target charging pile is the final determined charging equipment suitable for serving the current vehicle.

[0139] In this embodiment of the application, the system uses a weight filtering module to check whether the connection weight of the temporary connection edge of each charging pile node in the candidate charging pile set reaches the preset weight threshold requirement, and selects the charging pile node that meets the conditions as the target charging pile.

[0140] Step 404: Using the spatial analysis module of the digital twin model, extract the spatial coordinates of the parking space node corresponding to the target charging pile and the position coordinates of the dynamic vehicle node, and calculate the path reachability parameters between the position coordinates of the dynamic vehicle node and the spatial coordinates of the parking space node.

[0141] In step 404, the path accessibility parameter is a comprehensive indicator reflecting the ease of access from the vehicle location to the charging station parking space.

[0142] In this embodiment, the system obtains the spatial coordinates of the parking space node corresponding to the target charging pile and the position coordinates of the associated dynamic vehicle node through the spatial analysis module, and calculates the path reachability parameters based on the spatial relationship between the two points and environmental characteristics.

[0143] Step 405: Based on the path reachability parameters, the decision generation module of the digital twin model generates a dynamic ranking result of the matching priority from the vehicle node to the target charging pile.

[0144] In step 405, the matching priority ranking result is the order of the matching priority between the vehicle and multiple candidate charging stations.

[0145] In this embodiment of the application, the system generates a dynamic ranking result of the matching priority from vehicle nodes to different target charging piles based on the path reachability parameters of each candidate charging pile through the decision generation module.

[0146] Step 406: Through the strategy output module of the digital twin model, determine the optimal charging pile allocation scheme according to the matching priority sorting result, and form a charging pile guidance strategy based on the optimal charging pile allocation scheme.

[0147] In step 406, the optimal charging pile allocation scheme is the best resource allocation scheme determined by the system.

[0148] In this embodiment of the application, the system selects the optimal matching pair according to the matching priority sorting result through the strategy output module to form a charging pile allocation scheme, and formulates a charging pile guidance strategy containing specific execution steps based on the scheme.

[0149] Here is a specific example:

[0150] After constructing a dynamic matching map of charging piles, parking spaces, and vehicles, comprising 8 charging pile nodes, 12 parking space nodes, and 3 dynamic vehicle nodes based on the aforementioned embodiments, the system inputs this dynamic matching map into a digital twin model. The graph data processing module processes the data to obtain the real-time load coefficients of each charging pile node (0.8, 0.6, 0.9, 0.3, 0.7, 0.4, 0.5, 0.2) and the connection weights of each temporary connection edge (0.7, 0.8, 0.6, 0.9, 0.5). The load evaluation module then filters charging pile nodes whose real-time load coefficients are less than a preset threshold of 0.5. The charging pile nodes were identified, resulting in a candidate charging pile set consisting of four charging piles numbered 4, 6, 7, and 8. A weighted filtering module selected charging pile nodes from this set whose connection weights were greater than a preset weight threshold of 0.8. Charging pile number 4 was determined as the target charging pile. The spatial analysis module extracted the spatial coordinates (135210) of the corresponding parking space node and the position coordinates (10580) of the dynamic vehicle node V1. The path reachability parameter between them was calculated using the formula: path reachability parameter equals straight-line distance divided by path complexity coefficient. The straight-line distance is calculated using the formula for distance between two points. The path complexity coefficient is determined to be 1.2 based on the number of turns and obstacles. The path reachability parameter is 133.5 divided by 1.2, which equals 111.25. The decision generation module generates a priority ranking result for matching dynamic vehicle nodes to each candidate charging pile based on this path reachability parameter. Among them, charging pile number 4 has the highest priority score of 8.5 points, charging pile number 6 has 7.2 points, charging pile number 7 has 6.8 points, and charging pile number 8 has 6.5 points. The strategy output module determines charging pile number 4 as the optimal charging pile allocation scheme based on this ranking result, forming a charging pile guidance strategy to allocate charging pile number 4 to dynamic vehicle node V1.

[0151] In this embodiment of the application, the charging pile guidance strategy generation scheme can comprehensively consider the charging pile load status, spatial location relationship and path access conditions, and realize the intelligent allocation of charging pile resources through multi-module collaborative processing, effectively improving the efficiency and quality of charging services.

[0152] To further improve the accuracy and real-time performance of charging pile resource management, in some embodiments, step 102: constructing a dynamic matching map of charging pile parking spaces based on the charging pile location topology data, the real-time vehicle location data, and the historical usage rate data of the charging piles includes:

[0153] Step 501: Establish charging pile nodes containing the location coordinates of charging piles and berth nodes containing the spatial coordinates of berths.

[0154] In step 501, the charging pile node is a point element representing the spatial location of the charging equipment, and the parking space node is a point element representing the parking space location.

[0155] In this embodiment, the system creates charging pile nodes in a digital twin environment based on the actual installation location of the charging piles and creates berth nodes based on the actual spatial distribution of the berths, thus establishing a basic spatial location framework.

[0156] Step 502: Based on the charging pile location topology data, establish a fixed connection edge between the charging pile node and the berth node.

[0157] In step 502, the fixed connection edge is a connection line that indicates the fixed affiliation between the charging pile and the berth.

[0158] In this embodiment of the application, the system establishes a fixed connection edge between the corresponding charging pile node and the berth node based on the correspondence between charging piles and berths recorded in the charging pile location topology data.

[0159] Step 503: Convert the real-time vehicle location data into dynamic vehicle nodes with timestamps, and update the location coordinates of the dynamic vehicle nodes in real time.

[0160] In step 503, the dynamic vehicle node is a point element representing the real-time location of the vehicle to be charged, and the timestamp is a marker that records the time of location acquisition.

[0161] In this embodiment of the application, the system continuously receives real-time vehicle location data, converts this data into dynamic vehicle nodes with time stamps, and updates the coordinate information of these nodes in real time in the digital twin environment.

[0162] Step 504: Calculate the real-time load factor of each charging pile node based on the historical usage data of the charging pile.

[0163] In step 504, the real-time load factor is a dynamic parameter that reflects the current operating pressure of the charging pile.

[0164] In this embodiment of the application, the system calculates and analyzes the real-time load status index of each charging pile node based on the historical usage data of the charging pile.

[0165] Step 505: Based on the spatial proximity relationship between the updated position coordinates of the dynamic vehicle node and the spatial coordinates of the berth node, establish a temporary connection edge between the dynamic vehicle node and the berth node.

[0166] In step 505, the spatial proximity relationship is the degree of spatial proximity between the dynamic vehicle node and the parking space node in the digital twin environment. This relationship is formed by calculating the Euclidean distance between the real-time position coordinates of the dynamic vehicle node and the fixed spatial coordinates of the parking space node, and determining whether the distance is less than a set proximity threshold.

[0167] In this embodiment, the system establishes a temporary connection edge when the distance reaches a set condition, based on the distance relationship between the updated dynamic vehicle node position coordinates and the parking space node spatial coordinates.

[0168] Step 506: Using the real-time load coefficient of the charging pile node, assign weights to the temporary connection edge to obtain a temporary connection edge with connection weights.

[0169] In step 506, the connection weight is a quantified value of the association strength represented by the temporary connection edge.

[0170] In this embodiment of the application, the system uses the real-time load coefficient of the charging pile node to assign corresponding weight values ​​to the temporary connection edge according to the set calculation rules.

[0171] Step 507: Integrate the charging pile nodes with real-time load coefficients, the parking space nodes, the dynamic vehicle nodes, the fixed connection edges, and the temporary connection edges with connection weights to construct a dynamic matching graph for charging piles, parking spaces, and vehicles.

[0172] In this embodiment, the system integrates charging pile nodes with real-time load coefficients, parking space nodes, dynamic vehicle nodes, fixed connection edges, and temporary connection edges with connection weights to form a complete dynamic matching graph.

[0173] In this embodiment of the application, the dynamic matching graph construction scheme can reflect the dynamic changes in the status of charging pile resources and vehicle location in real time, establish an accurate relationship between resources and demand, and provide a reliable data foundation for subsequent intelligent guidance.

[0174] To further improve the accuracy and automation of vehicle charging docking, in some embodiments, step 105: performing multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging interface location and optimal parking posture, includes:

[0175] Step 601: Perform three-dimensional structure reconstruction on the vehicle contour point cloud data to generate a three-dimensional model of the vehicle surface.

[0176] In step 601, the three-dimensional model of the vehicle surface is a three-dimensional digital representation of the vehicle's external contour obtained by reconstructing point cloud data.

[0177] In this embodiment of the application, the system processes the vehicle contour point cloud data collected by the lidar and generates a three-dimensional model that accurately reflects the external shape of the vehicle through a surface reconstruction algorithm.

[0178] Step 602: Extract the motion trajectory from the parking trajectory image data to obtain the vehicle pose change sequence.

[0179] In step 602, the vehicle pose change sequence is a continuous record of the vehicle position and attitude changes over time, extracted from the image data.

[0180] In this embodiment, the system analyzes the parking trajectory image sequence captured by the camera and extracts the vehicle's pose change data during the parking process using a visual tracking algorithm to form a time series.

[0181] Step 603: Input the three-dimensional model of the vehicle surface and the vehicle pose change sequence into a multimodal fusion neural network for spatiotemporal correlation to generate a three-dimensional pose sequence during vehicle motion.

[0182] In step 603, the three-dimensional attitude sequence is a continuous state record of the vehicle's three-dimensional position and orientation during motion.

[0183] In this embodiment of the application, the system inputs the three-dimensional model of the vehicle surface and the vehicle pose change sequence into a multimodal fusion neural network, and generates a vehicle motion posture sequence containing complete three-dimensional information through the spatiotemporal feature extraction capability of the network.

[0184] Step 604: In the three-dimensional model of the vehicle surface, identify the feature area of ​​the charging interface to determine the location of the vehicle charging interface.

[0185] In this embodiment, the system uses a feature recognition algorithm to locate the feature area of ​​the charging interface in the three-dimensional model of the vehicle surface, and determines its precise position in the vehicle coordinate system.

[0186] Step 605: Based on the three-dimensional attitude sequence, analyze the relative positional relationship between the vehicle and the charging pile. Based on the relative positional relationship, use a deep reinforcement learning model to calculate the target docking attitude of the charging interface corresponding to the charging pile socket.

[0187] In step 605, the target docking posture is the ideal relative posture that needs to be achieved when the charging interface and the charging pile socket achieve perfect docking.

[0188] In this embodiment, the system analyzes the relative positional relationship between the vehicle and the charging pile based on a three-dimensional attitude sequence, and calculates the optimal docking attitude parameters using a deep reinforcement learning model. Specifically, based on the vehicle position coordinates, heading angle data extracted from the three-dimensional attitude sequence, and the charging pile position coordinates, the system calculates the relative distance and angle difference between them. These parameters, along with the vehicle's charging interface position, form a state vector which is input into the deep reinforcement learning model. This model evaluates the long-term benefits of different docking attitudes through its value network and outputs the optimal docking action sequence through its policy network. For example, when the system detects that the vehicle is located 2.5 meters southeast of the charging pile with a heading angle deviation of 15 degrees, the model calculates and outputs the target docking attitude of "first turning 12 degrees to the right and then moving forward in a straight line for 1.8 meters," ensuring that the charging interface is ultimately precisely aligned with the charging pile socket.

[0189] Step 606: Determine the optimal parking posture based on the vehicle charging interface location and the target docking posture.

[0190] In this embodiment, the system calculates the optimal parking position and orientation angle of the vehicle based on the location of the vehicle charging interface and the target docking posture using geometric relationships.

[0191] In this embodiment, the multi-sensor data fusion processing scheme can accurately identify vehicle characteristics and motion state, calculate the optimal docking scheme through intelligent algorithms, realize the precise docking of the vehicle and the charging pile, and improve the convenience and success rate of charging operation.

[0192] To further improve the accuracy of charging and parking guidance and the user experience, in some embodiments, step 106: generating a fine-grained guidance instruction based on the vehicle's charging port location and optimal parking posture to assist the vehicle in parking the charging station, and sending the fine-grained guidance instruction to the user's mobile terminal and the charging station, includes:

[0193] Step 701: Calculate the horizontal deflection angle and vertical distance between the vehicle charging interface and the charging pile interface, and generate a lateral correction command based on the horizontal deflection angle and the vertical distance.

[0194] In step 701, the horizontal deviation angle is the angular deviation between the vehicle charging interface and the charging pile interface in the horizontal direction, the vertical distance is the height difference between the two in the vertical direction, and the lateral correction command is an operation guide used to adjust the horizontal deviation of the vehicle.

[0195] In this embodiment, the system calculates the horizontal angle deviation and vertical height difference between the vehicle charging interface and the charging pile interface through geometric relationships, and generates a lateral correction command that includes the steering wheel rotation direction and angle based on these parameters.

[0196] Step 702: Based on the optimal parking posture, determine the forward direction and rotation angle that the vehicle needs to adjust, and generate longitudinal control commands according to the forward direction and rotation angle.

[0197] In step 702, the forward direction is the direction in which the vehicle needs to move, the rotation angle is the amount of steering wheel rotation that the vehicle needs to adjust, and the longitudinal control command is the operation guide used to control the vehicle's forward and backward movement and speed.

[0198] In this embodiment of the application, the system determines the specific direction the vehicle needs to travel and the steering wheel rotation angle based on the optimal parking posture, and generates longitudinal control commands that include speed control and travel distance based on these parameters.

[0199] Step 703: Combine the lateral correction command and the longitudinal control command to form a fine guidance command.

[0200] In this embodiment, the system combines and arranges the lateral correction instructions and longitudinal control instructions in the order of execution to form a fine guidance instruction containing complete operation steps.

[0201] Step 704: Based on the fine-grained guidance instructions, generate visual prompts and device control signals, send the visual prompts to the user's mobile terminal for graphical display, and send the device control signals to the charging pile to control the charging pile to prepare for charging service.

[0202] In step 704, the visual prompts are converted into graphical interface display guidance content, and the device control signals are device control instructions sent to the charging pile. Controlling the charging pile to prepare for charging service refers to the system sending device control signals to the charging pile, triggering the charging pile to perform a series of preparatory tasks, including initiating a self-test program to check if the device status is normal, controlling the charging compartment door to open automatically, moving the charging gun to the pre-dock position, activating the payment QR code display screen, and updating the charging pile status to ready, so that it can immediately enter the charging process after the vehicle stops.

[0203] In this embodiment, the system converts the fine-grained guidance instructions into graphical prompts suitable for display on the mobile terminal, and generates control signals to control the charging pile's preparation status, which are then sent to the user's mobile terminal and the charging pile equipment, respectively.

[0204] In this embodiment of the application, the fine-grained guidance instruction generation scheme can provide clear and specific operation guidance, realize the precise docking of vehicles and charging piles, effectively reduce the difficulty of user operation, and improve the automation level of charging services and user experience.

[0205] Figure 3 A schematic diagram of a digital twin-based charging pile service guidance system provided in this application embodiment is shown below. Figure 3 As shown, the detailed implementation section describes:

[0206] The data acquisition module 31 is used to collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area.

[0207] The construction module 32 is used to construct a dynamic matching map of charging pile parking spaces and vehicles based on the charging pile location topology data, the vehicle real-time location data and the charging pile historical usage data, so as to evaluate the availability of charging piles in real time.

[0208] The generation module 33 is used to generate a charging pile guidance strategy based on the dynamic matching map of the charging pile parking space and the digital twin model. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, and is used to allocate charging pile resources.

[0209] The calculation module 34 is used to calculate the optimal guidance path from the vehicle's current location to the available charging pile based on the vehicle's real-time location data and the available charging pile using a path planning algorithm, and to send the optimal guidance path information to the charging pile and the user's mobile terminal respectively, so as to guide the vehicle to the available charging pile.

[0210] The acquisition module 35 is used to acquire vehicle contour point cloud data and parking trajectory image data when the vehicle drives to the preset range of the charging pile and prepares to park, and to perform multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging interface position and the optimal parking posture.

[0211] The sending module 36 is used to generate fine guidance instructions based on the location of the vehicle charging interface and the optimal parking posture to assist the vehicle in parking in the charging pile space, and send the fine guidance instructions to the user's mobile terminal and the charging pile to realize service guidance.

[0212] The digital twin-based charging pile service guidance system of this application embodiment is used to implement the aforementioned digital twin-based charging pile service guidance method. Therefore, the specific implementation of the digital twin-based charging pile service guidance system can be found in the embodiment section of the digital twin-based charging pile service guidance method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0213] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described digital twin-based charging pile service guidance methods.

[0214] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described digital twin-based charging pile service guidance methods.

[0215] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0216] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the digital twin-based charging pile service guidance method.

[0217] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0218] The foregoing has provided a detailed description of a digital twin-based charging pile service guidance method, system, electronic device, and storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A charging pile service guidance method based on digital twin, characterized in that, include: Collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area; Based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data, a dynamic matching map of charging pile parking spaces and vehicles is constructed to assess the availability of charging piles in real time. Based on the dynamic matching map of charging pile parking spaces and vehicles, combined with the digital twin model, a charging pile guidance strategy is generated. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, which is used to allocate charging pile resources. Based on the vehicle's real-time location data and the available charging piles, a path planning algorithm is used to calculate the optimal guidance path from the vehicle's current location to the available charging piles. The optimal guidance path information is then sent to the charging piles and the user's mobile terminal to guide the vehicle to the available charging piles. When the vehicle drives to the preset range of the charging pile and prepares to park, the vehicle contour point cloud data and parking trajectory image data are acquired, and the vehicle contour point cloud data and the parking trajectory image data are subjected to multi-sensor data fusion processing to identify the vehicle charging interface position and the optimal parking posture. Based on the location of the vehicle's charging port and the optimal parking posture, a precise guidance instruction is generated to assist the vehicle in parking the charging station. The precise guidance instruction is then sent to the user's mobile terminal and the charging station to provide service guidance.

2. The method according to claim 1, characterized in that, The step of calculating the optimal guidance path from the vehicle's current location to an available charging station using a path planning algorithm, based on the vehicle's real-time location data and the available charging stations, includes: Obtain the current coordinates of the vehicle's real-time location data and the coordinates of the target parking space corresponding to the available charging pile; Establish a path search area with the current coordinate point as the starting point and the target berth coordinate point as the ending point, and extract channel network data and obstacle distribution data from the path search area; Based on the channel network data, a path topology graph containing path nodes and connecting edges is constructed. In the path topology graph, based on the obstacle distribution data, a path planning algorithm is used to calculate multiple candidate paths from the starting point to the ending point. For each candidate path, perform path characteristic analysis to generate path length and turning complexity values. Based on the path length and turning complexity values, and combined with the deep Q-network model, calculate the comprehensive cost of each candidate path. Candidate paths whose overall cost is less than a preset cost threshold are selected as the optimal guiding path.

3. The method according to claim 2, characterized in that, The process involves analyzing the path characteristics of each candidate path to generate a path length value and a turning complexity value. Based on these values ​​and the turning complexity value, and using a deep Q-network model, the comprehensive cost of each candidate path is calculated, including: For each candidate path, perform path geometric feature analysis to extract the path node sequence and connecting edge attributes. Based on the path node sequence and connecting edge attributes, calculate the actual travel distance of each connecting edge. By accumulating the actual travel distances of each connecting edge, generate the path length value. For each candidate path, perform turning feature analysis, count the number of turns in the candidate path, and calculate the angle change for each turn. Then, perform a weighted calculation on the number of turns and the angle change to obtain the turning complexity value. The path length value and the turning complexity value are normalized respectively, and the normalized path length value and the normalized turning complexity value are combined into a standardized path feature vector. The standardized path feature vector is input into a deep Q-network model, and path space features are extracted through the multi-layer convolutional network of the deep Q-network model. The path space features are evaluated using the fully connected layers of the deep Q-network model, and the comprehensive cost value of each candidate path is output.

4. The method according to claim 1, characterized in that, The charging pile guidance strategy is generated based on the dynamic vehicle matching map of the charging pile parking space and combined with the digital twin model, including: The dynamic matching map of the charging pile parking space vehicles is input into the digital twin model. The graph data processing module of the digital twin model processes the dynamic matching map of the charging pile parking space vehicles to obtain charging pile nodes with real-time load coefficients and temporary connection edges with connection weights. The load assessment module of the digital twin model is used to screen charging pile nodes whose real-time load coefficient is less than a preset coefficient threshold to form a candidate charging pile set. The digital twin model's weight filtering module selects charging pile nodes with connection weights greater than a preset weight threshold from the candidate charging pile set, and determines them as target charging piles. The spatial analysis module of the digital twin model extracts the spatial coordinates of the parking space node and the position coordinates of the dynamic vehicle node corresponding to the target charging pile, and calculates the path reachability parameters between the position coordinates of the dynamic vehicle node and the spatial coordinates of the parking space node. The decision generation module of the digital twin model generates a dynamic ranking result of the matching priority from the vehicle node to the target charging pile based on the path reachability parameters. The optimal charging pile allocation scheme is determined by the strategy output module of the digital twin model based on the matching priority ranking result, and a charging pile guidance strategy is formed based on the optimal charging pile allocation scheme.

5. The method according to claim 1, characterized in that, The step of constructing a dynamic matching map of charging pile parking spaces based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data includes: Establish charging pile nodes containing the location coordinates of charging piles and berth nodes containing the spatial coordinates of berths; Based on the charging pile location topology data, a fixed connection edge is established between the charging pile node and the berth node; The real-time vehicle location data is converted into dynamic vehicle nodes with timestamps, and the location coordinates of the dynamic vehicle nodes are updated in real time. Based on the historical usage data of the charging piles, the real-time load factor of each charging pile node is calculated; Based on the spatial proximity relationship between the updated position coordinates of the dynamic vehicle node and the spatial coordinates of the berth node, a temporary connection edge is established between the dynamic vehicle node and the berth node. Using the real-time load coefficient of the charging pile node, the temporary connection edge is weighted to obtain a temporary connection edge with connection weight. A dynamic matching graph for charging piles, parking spaces, and vehicles is constructed by integrating charging pile nodes with real-time load coefficients, parking space nodes, dynamic vehicle nodes, fixed connection edges, and temporary connection edges with connection weights.

6. The method according to claim 1, characterized in that, The step of performing multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging port location and optimal parking posture includes: The vehicle contour point cloud data is used to perform three-dimensional structure reconstruction to generate a three-dimensional model of the vehicle surface. Motion trajectory extraction is performed on the parking trajectory image data to obtain a vehicle pose change sequence; The three-dimensional model of the vehicle surface and the vehicle pose change sequence are input into a multimodal fusion neural network for spatiotemporal correlation to generate a three-dimensional pose sequence during vehicle motion. In the three-dimensional model of the vehicle surface, the characteristic areas of the charging interface are identified to determine the location of the vehicle charging interface; Based on the three-dimensional posture sequence, the relative positional relationship between the vehicle and the charging pile is analyzed. Based on the relative positional relationship, the target docking posture of the charging interface corresponding to the charging pile socket is calculated using a deep reinforcement learning model. The optimal parking posture is determined based on the location of the vehicle charging port and the target docking posture.

7. The method according to claim 1, characterized in that, Based on the vehicle's charging port location and optimal parking posture, a refined guidance instruction is generated to assist the vehicle in parking the charging station space. This refined guidance instruction is then sent to the user's mobile terminal and the charging station, including: Calculate the horizontal angle and vertical distance between the vehicle charging interface and the charging pile interface, and generate a lateral correction command based on the horizontal angle and the vertical distance; Based on the optimal parking posture, the vehicle's forward direction and rotation angle that need to be adjusted are determined, and longitudinal control commands are generated according to the forward direction and rotation angle. By combining the lateral correction command and the longitudinal control command, a fine guidance command is formed; Based on the precise guidance instructions, visual prompts and device control signals are generated. The visual prompts are sent to the user's mobile terminal for graphical display, and the device control signals are sent to the charging pile to control the charging pile to prepare for charging service.

8. A charging pile service guidance system based on digital twins, characterized in that, include: The data acquisition module is used to collect topology data of charging pile locations, real-time vehicle location data, and historical usage data of charging piles within the target area. The module is used to construct a dynamic matching map of charging pile parking spaces and vehicles based on the charging pile location topology data, the vehicle real-time location data, and the charging pile historical usage rate data, so as to evaluate the availability of charging piles in real time. The generation module is used to generate a charging pile guidance strategy based on the dynamic matching map of the charging pile parking space and the digital twin model. The charging pile guidance strategy includes the correspondence between available charging piles and vehicles, which is used to allocate charging pile resources. The calculation module is used to calculate the optimal guidance path from the vehicle's current location to the available charging pile based on the vehicle's real-time location data and the available charging pile using a path planning algorithm, and send the optimal guidance path information to the charging pile and the user's mobile terminal respectively to guide the vehicle to the available charging pile. The acquisition module is used to acquire vehicle contour point cloud data and parking trajectory image data when the vehicle drives to the preset range of the charging pile and prepares to park, and to perform multi-sensor data fusion processing on the vehicle contour point cloud data and the parking trajectory image data to identify the vehicle charging interface position and the optimal parking posture. The sending module is used to generate fine guidance instructions based on the location of the vehicle's charging interface and the optimal parking posture to assist the vehicle in parking in the charging pile space, and send the fine guidance instructions to the user's mobile terminal and the charging pile to realize service guidance.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the digital twin-based charging pile service guidance method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the digital twin-based charging pile service guidance method as described in any one of claims 1 to 7.

Images