Charging path planning method for intelligent mobile charging vehicle

Abstract

The invention belongs to the technical field of path planning, and particularly relates to a charging path planning method for an intelligent mobile charging vehicle, and the method comprises the steps: collecting the three-dimensional data information of a charging scene through a camera and a laser radar, and constructing a three-dimensional space model; image acquisition of a charging scene is carried out through a camera and a laser radar, and a plane navigation map is established; after the mobile charging vehicle receives a charging instruction, performing global path planning based on an improved PSO algorithm to obtain an optimal path; real-time image acquisition is carried out through a camera and a laser radar, local path planning is carried out through an improved DWA algorithm, obstacles are avoided in time, and whether a target point is reached or not is verified; and the charging manipulator grabs the charging wire to perform charging operation. According to the invention, the optimization of global and local path planning is realized, the accuracy and efficiency of path planning are improved, and the charging vehicle can safely and quickly reach the target charging position.

Description

Technical Field

[0001] The present invention belongs to the technical field of path planning, and in particular relates to a charging path planning method for an intelligent mobile charging vehicle. Background Art

[0002] With the continuous progress of new high-energy secondary battery technology, especially lithium-ion power battery technology, the development and promotion of electric vehicle technology represented by electric bicycles and electric vehicles has received increasing attention. The use of electric vehicles is increasing, but due to the impact of the endurance of electric vehicles and the construction of charging facilities, the convenience of electric vehicle charging is one of the current pain points, such as difficulty in charging at highway service areas during holidays and difficulty in charging at daily community parking lots.

[0003] In order to solve this problem, mobile charging vehicles came into being. Mobile charging vehicles can charge themselves and move to electric vehicles that need to be charged when needed to charge the vehicles. Therefore, path planning is required for them. The existing mobile charging vehicle path planning methods and their shortcomings mainly include: Inflexible charging time: The traditional charging strategy is usually a full charging strategy, that is, the nearest charging station is visited and fully charged only when the battery power cannot last to the next customer point. This strategy will lead to inflexible vehicle charging time and difficulty in meeting customer time windows and other issues that electric vehicle users are more concerned about.

[0004] Traffic congestion and grid load issues: Large-scale charging of electric vehicles during peak charging hours may cause local traffic congestion near charging stations, resulting in problems such as users waiting for charging and large fluctuations in grid load.

[0005] The construction speed of charging facilities cannot keep up with the growth rate of electric vehicles: With the rapid development of electric vehicles, the construction speed of supporting charging facilities is far behind its growth rate. The mismatch of charging capacity has become a bottleneck restricting the development of shared electric vehicles.

[0006] Limitations of path planning methods: Although some of the most commonly used route planning methods at home and abroad, such as genetic algorithms, simulated annealing algorithms, particle swarm algorithms, etc., are widely used, they have limitations in efficiency and optimization degree.

[0007] Data transmission reliability and security issues: In the charging pile network access solution, although the wired method has reliable data transmission and simple design, it is complex to lay out, has poor scalability, high construction cost, poor flexibility, and low communication capacity. Although the wireless method does not require complex wiring, has strong flexibility, and low construction cost, it requires expensive fees to use the mobile data services of mobile operators, and the security of data and the reliability of the network are limited by mobile operators. Summary of the invention

[0008] In view of the above deficiencies in the prior art, the purpose of the present invention is to provide a charging path planning method for an intelligent mobile charging vehicle, which realizes the optimization of global and local path planning, improves the accuracy and efficiency of path planning, and enables the charging vehicle to reach the target charging position safely and quickly.

[0009] To achieve the above objectives, the present invention provides a charging path planning method for an intelligent mobile charging vehicle. For a mobile charging vehicle equipped with a camera, a display screen, a charging manipulator, a charging cable, and a laser radar, a charging path planning is performed, and the steps include: S1. In the charging scene, the mobile charging vehicle collects three-dimensional data information of the charging scene through cameras and laser radars to build a three-dimensional space model; S2. Capture images of the charging scene through cameras and lidar, establish a coordinate system, mark the current environment and the location of obstacles, and create a plane navigation map; S3. When the mobile charging vehicle receives the charging instruction, it performs global path planning based on the improved PSO algorithm to obtain the optimal path, and the mobile charging vehicle moves according to the optimal path; S4, during the movement, real-time image acquisition is performed through the camera and the laser radar, and the acquired real-time images are spliced ​​and the perspective is corrected to obtain a real-time monitoring plane image; S5. Based on the real-time monitoring plane image, the improved DWA algorithm is used for local path planning, obstacles are avoided in time, the next motion control input is generated through the dynamic window method, the mobile charging vehicle status is updated, and it is verified whether the target point is reached, until the distance between the position of the mobile charging vehicle and the target point does not exceed the set range; S6. Grab the charging cable with the charging manipulator to perform charging operations.

[0010] As a preferred embodiment of the present invention, a laser radar is arranged on the front side of the mobile charging vehicle, a display screen is arranged on the rear side of the mobile charging vehicle, a charging manipulator is installed on one side of the display screen, a camera is arranged on the lower side of the display screen, and a charging cord winding rod is also arranged below the camera. The charging manipulator is used to plug the charging cord wound on the charging cord winding rod into the charging pile or the vehicle to be charged during charging, and to recycle the charging cord after charging is completed.

[0011] As a preferred solution of the present invention, the mobile charging vehicle charges itself during low-peak electricity consumption and charges the vehicles that need to be charged during peak electricity consumption; When charging vehicles that need charging, the mobile charging vehicle will charge different vehicles continuously based on route planning, and users can make charging reservations based on their mobile phones or clients.

[0012] As a preferred solution of the present invention, in S3, the improved PSO algorithm is expressed as: (1); (2); In the formula, v i is the velocity of the ith particle, i=1, 2, , I, I is the total number of particles; c 1 、c 2 is the learning factor; rand() is a random number between 0 and 1; is the optimal position currently searched for the i-th particle; is the optimal position currently searched by the entire particle swarm; x i is the current position of the ith particle, and equation (2) indicates updating the particle position according to the new velocity; The speed update formula is: (3); In the formula, is the inertia factor; The inertia weight update strategy based on chaotic mapping makes the inertia weight change strongly nonlinearly with the number of iterations, which can be expressed as: (4); (5); In the formula, is the inertia factor of the ith particle; , They represent the upper and lower limits of the inertia factor respectively; k represents the current number of iterations, is the maximum number of iterations of the particle swarm; , Respectively represent the values ​​of chaotic mapping at the kth and k-1th iterations; is the parameter of the chaotic map; rand is a random number between 0 and 1.

[0013] As a preferred embodiment of the present invention, c 1 、c 2 The average value is 2. The value range is 3.6-4.0.

[0014] As a preferred solution of the present invention, in S4, the improved DWA algorithm uses a membership function and a range transformation method to normalize the eigenvalues ​​of the optimal path into fuzzy values, and constructs a membership matrix, which is expressed as: (6); In the formula, is an element in the membership matrix, which represents the fuzzy value of the mth path plan on the nth evaluation index after the range transformation method; represents the actual value of the mth path plan on the nth evaluation index; the original Convert to a value between 0 and 1 , so that all indicator values ​​are on the same scale, which is convenient for comprehensive evaluation.

[0015] As a preferred solution of the present invention, in the improved DWA algorithm, the entropy weight method is used to determine the weight vector, and for the membership matrix , the entropy of the nth evaluation index is: (7); (8); In the formula, h n is the entropy value of the nth evaluation index; is the membership probability of the mth evaluation object under the nth evaluation index, M is the number of evaluation objects, and the evaluation object is the path plan; lg is a logarithmic function, which represents the logarithm with base 10; h n After determination, the entropy weight of the nth evaluation index is: (9); In the formula, w n is the weight of the nth evaluation index; N is the number of evaluation indicators; According to w n and Get the performance vector of the optimal path solution , thereby calculating the optimal local path solution.

[0016] As a preferred solution of the present invention, a weather API is integrated in the mobile charging vehicle. When charging operations are performed in an open-air environment, rainfall, snowfall, temperature, wind speed, and visibility are obtained in real time. In combination with the visual information captured by the camera, image recognition technology is used to identify the real-time weather. In the path planning process of S3-S5, heuristic rules are introduced. According to the impact of weather conditions on driving safety, weather influencing factors are defined, and safety factors are assigned to different weather conditions. Heuristic rules are set according to weather influencing factors and safety factors.

[0017] As a preferred solution of the present invention, the open-air environment is an open-air parking lot, and the heuristic rules include: Rainfall heuristics: Avoid charging in low-lying areas or areas with poor drainage during rainfall; give priority to parking spaces with shelter or near buildings to reduce the interference of rain on the charging process; Snowfall heuristic: After a snowfall, choose a parking space that has been cleared of snow; High temperature heuristic: monitor battery temperature and take cooling measures when it exceeds the safety threshold; Low temperature heuristic: preheat the battery before charging to improve charging efficiency; Strong wind heuristic: Avoid charging in open areas during strong winds; Lightning heuristic: suspend charging operations during thunderstorms; Fog heuristic: In foggy weather, use the parking lot lighting system to ensure that the charging vehicle can safely navigate to the parking space; Parking lot capacity heuristic: Prioritize easy-to-access parking spaces during inclement weather conditions; User preference heuristics: Allow users to set sensitivity to weather conditions based on personal preferences, such as not charging in rainy or snowy weather.

[0018] As a preferred solution of the present invention, an adaptive learning algorithm is introduced in charging path planning to continuously optimize the path planning strategy based on historical charging data and user feedback. The specific process is as follows: Collect data on the operation of mobile charging vehicles under different conditions, including route selection, charging time, user satisfaction, energy consumption, and weather conditions; Preprocess the operation data and extract features after preprocessing, including path length, charging efficiency, waiting time, and the impact of weather changes on charging; Using machine learning technology, we screened out several features that ranked highly in terms of their impact on charging path planning, built a reinforcement learning model, trained the reinforcement learning model using the extracted features and historical charging data, and learned the optimal strategy by simulating the effects of different path planning strategies. Test the performance of the reinforcement learning model on an independent validation set, adjust the model parameters based on the validation results, optimize the model structure, and improve the decision-making quality of the reinforcement learning model.

[0019] The algorithm involved in the present invention can be executed by an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the above algorithm calculation is implemented by executing the software by the processor.

[0020] The beneficial effects of the present invention are: The present invention can realize charging itself during low-peak electricity consumption periods and provide charging services for electric vehicles during peak electricity consumption periods, thereby optimizing energy use and improving charging efficiency; by combining an improved particle swarm optimization (PSO) algorithm with a dynamic window algorithm (DWA), global and local path planning optimization is achieved, the accuracy and efficiency of path planning are improved, and the mobile charging vehicle can reach the target charging location safely and quickly.

[0021] The present invention provides two operation modes, manual and automatic, which enhances flexibility and applicability. In the automatic mode, the environmental data is collected by high-definition cameras and laser radars, combined with improved PSO and DWA algorithms, to achieve unmanned operation of the intelligent mobile charging vehicle, reduce manual intervention, and improve the automation level of the charging process. At the same time, users can remotely control and make reservations for charging services through mobile phone APP, further improving user experience and charging convenience. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 It is a front view structural schematic diagram of the mobile charging vehicle in the present invention; Figure 3 It is a right view structural schematic diagram of the mobile charging vehicle in the present invention; Figure 4 It is a structural schematic diagram of the charging manipulator in the present invention; Figure 5 It is a schematic diagram of the charging path planning of the present invention.

[0023] In the picture, 1. Camera; 2. Display screen; 3. Charging robot; 4. Charging cable winding rod; 5. LiDAR. DETAILED DESCRIPTION

[0024] The embodiments of the present invention are further described below in conjunction with the accompanying drawings: A charging path planning method for an intelligent mobile charging vehicle is provided. For a mobile charging vehicle provided with a camera 1, a display screen 2, a charging manipulator 3, a charging line, and a laser radar 5, a charging path planning is performed. Figure 1 As shown, the steps include: S1. In a charging scene, the mobile charging vehicle collects three-dimensional data information of the charging scene through the camera 1 and the laser radar 5 to construct a three-dimensional space model; S2. Capture images of the charging scene through the camera 1 and the laser radar 5, establish a coordinate system, mark the current environment and the location of obstacles, and establish a plane navigation map; S3. When the mobile charging vehicle receives the charging instruction, it performs global path planning based on the improved PSO algorithm to obtain the optimal path, and the mobile charging vehicle moves according to the optimal path; S4, during the movement, real-time image acquisition is performed through the camera 1 and the laser radar 5, and the acquired real-time images are spliced ​​and the perspective is corrected to obtain a real-time monitoring plane image; S5. Based on the real-time monitoring plane image, the improved DWA algorithm is used for local path planning, obstacles are avoided in time, the next motion control input is generated through the dynamic window method, the mobile charging vehicle status is updated, and it is verified whether the target point is reached, until the distance between the position of the mobile charging vehicle and the target point does not exceed the set range; S6. The charging manipulator 3 grabs the charging cable and performs charging operation.

[0025] like Figure 2 and Figure 3 As shown, the laser radar 5 is arranged on the front side of the mobile charging vehicle (the front of the vehicle is the front side), the display screen 2 is arranged on the rear side of the mobile charging vehicle, the charging manipulator 3 is installed on one side of the display screen 2, and the camera 1 is arranged on the lower side of the display screen 2. A charging cable winding rod 4 is also arranged below the camera 1. The charging manipulator 3 is used to plug the charging cable wound on the charging cable winding rod 4 into the charging pile or the vehicle to be charged during charging, and recycle the charging cable after charging is completed. Figure 2 and Figure 3 In the figure, the charging line is not shown, and the charging manipulator 3 is simplified. The structural diagram of the complete charging manipulator 3 is shown in FIG. Figure 4 As shown, the charging robot 3 can be set using known technology.

[0026] The mobile charging vehicle charges itself during low-peak hours and charges vehicles that need charging during peak hours. When charging vehicles that need charging, the mobile charging vehicle charges different vehicles continuously based on route planning, and users can make reservations for charging based on their mobile phones or clients. The mobile charging vehicle can also be pushed directly to a designated location for charging as needed. When the mobile charging vehicle is charging itself, the route can be planned with the charging pile as the target point.

[0027] In S3, the improved PSO algorithm is expressed as: (1); (2); In the formula, v i is the velocity of the ith particle, i=1, 2, , I, I is the total number of particles; c 1 、c 2 is the learning factor; rand() is a random number between 0 and 1; is the optimal position currently searched for the i-th particle; is the optimal position currently searched by the entire particle swarm; x i is the current position of the ith particle, and equation (2) indicates updating the particle position according to the new velocity; The speed update formula is: (3); In the formula, is the inertia factor; The inertia weight update strategy based on chaotic mapping makes the inertia weight change strongly nonlinearly with the number of iterations, which can be expressed as: (4); (5); In the formula, is the inertia factor of the ith particle; , They represent the upper and lower limits of the inertia factor respectively; k represents the current number of iterations, is the maximum number of iterations of the particle swarm; , Respectively represent the values ​​of chaotic mapping at the kth and k-1th iterations; is the parameter of the chaotic mapping; rand is a random number between 0 and 1. The parameters of the chaotic mapping are variables used to control the behavior of the chaotic system, affecting the shape, size and dynamic characteristics of the chaotic attractor. 1 、c 2 The average value is 2. The value range is 3.6-4.0.

[0028] In S4, the improved DWA algorithm uses the membership function and the range transformation method to normalize the eigenvalues ​​of the optimal path into fuzzy values ​​and construct the membership matrix, which is expressed as: (6); In the formula, is an element in the membership matrix, which represents the fuzzy value of the mth path plan on the nth evaluation index after the range transformation method; represents the actual value of the mth path plan on the nth evaluation index; the original Convert to a value between 0 and 1 , so that all index values ​​are on the same scale, which is convenient for comprehensive evaluation. In this embodiment, the evaluation indexes are selected as distance evaluation index, speed evaluation index and distance obstacle evaluation index; , They respectively represent the maximum and minimum values ​​of the nth evaluation index in all path solutions.

[0029] In the improved DWA algorithm, the entropy weight method is used to determine the weight vector. , the entropy of the nth evaluation index is: (7); (8); In the formula, h n is the entropy value of the nth evaluation index; is the membership probability of the mth evaluation object under the nth evaluation index, M is the number of evaluation objects, and the evaluation object is the path plan; lg is a logarithmic function, which represents the logarithm with base 10; h n After determination, the entropy weight of the nth evaluation index is: (9); In the formula, w n is the weight of the nth evaluation index; N is the number of evaluation indicators; According to w n and Get the performance vector of the optimal path solution , thereby calculating the optimal local path solution.

[0030] In this embodiment, the electrical system and power system in the mobile charging car are implemented by the lithium battery thermal balance system device based on improved PID intelligent control in CN117219917B. Specifically, the electrical system part is mainly composed of temperature sensors, current sensors, pressure sensors, plc controllers, electronic switch boxes, and battery management systems. The power system part is mainly composed of lithium battery modules, fire extinguishing devices, water pumps, aluminum heat sinks, liquid circulation pipes between modules, and flame retardant balloons. In addition, the transmission system includes a drive motor, a motor, a reducer, and a controller, which are conventional settings in this field. The charging path planning method of this embodiment is mainly suitable for parking environments, especially flat indoor parking lots, but can also be used in other environments.

[0031] The mobile charging vehicle is integrated with a weather API. When charging in an open-air environment, it obtains rainfall, snowfall, temperature, wind speed, and visibility in real time. Combined with the visual information captured by camera 1, it uses image recognition technology to identify real-time weather. In the path planning process of S3-S5, heuristic rules are introduced. According to the impact of weather conditions on driving safety, weather influencing factors are defined, and safety factors are assigned to different weather conditions. Heuristic rules are set according to weather influencing factors and safety factors.

[0032] The open-air environment is an open-air parking lot, and the heuristic rules include: Rainfall heuristics: Avoid charging in low-lying areas or areas with poor drainage during rainfall; give priority to parking spaces with shelter or near buildings to reduce the interference of rain on the charging process; Snowfall heuristic: After a snowfall, choose a parking space that has been cleared of snow; High temperature heuristic: monitor battery temperature and take cooling measures when it exceeds the safety threshold; Low temperature heuristic: preheat the battery before charging to improve charging efficiency; Strong wind heuristic: Avoid charging in open areas during strong winds; Lightning heuristic: suspend charging operations during thunderstorms; Fog heuristic: In foggy weather, use the parking lot lighting system to ensure that the charging vehicle can safely navigate to the parking space; Parking lot capacity heuristic: Prioritize easy-to-access parking spaces during inclement weather conditions; User preference heuristics: Allow users to set sensitivity to weather conditions based on personal preferences, such as not charging in rainy or snowy weather.

[0033] In charging path planning, an adaptive learning algorithm is introduced to continuously optimize the path planning strategy based on historical charging data and user feedback. The specific process is as follows: Collect data on the operation of mobile charging vehicles under different conditions, including route selection, charging time, user satisfaction, energy consumption, and weather conditions; Preprocess the operation data and extract features after preprocessing, including path length, charging efficiency, waiting time, and the impact of weather changes on charging; Using machine learning technology, we screened out several features that ranked highly in terms of their impact on charging path planning, built a reinforcement learning model, trained the reinforcement learning model using the extracted features and historical charging data, and learned the optimal strategy by simulating the effects of different path planning strategies. Test the performance of the reinforcement learning model on an independent validation set, adjust the model parameters based on the validation results, optimize the model structure, and improve the decision-making quality of the reinforcement learning model.

[0034] An example of a complete reinforcement learning model for charging vehicle path planning is as follows: State space, including the current location and status of the charging vehicle (power level, health status, etc.), current time (considering peak and valley power consumption), the location and status of the target vehicle (whether it is charging, power demand, etc.), and environmental information (weather, traffic conditions, parking lot occupancy, etc.); Action space, including moving to a specific location (e.g., the location of the next target vehicle), adjusting charging strategy (e.g., adjusting charging speed based on battery status), avoiding obstacles or adverse weather areas; Reward function settings: Reach the target vehicle and start charging: +10; Complete charging: +20; Avoid obstacles or adverse weather: +5; Battery exhausted: -100; Path length: -1×path length (encourage short paths); User satisfaction: +1 × satisfaction score; Environmental model, simulating the dynamic changes of real scenes, including traffic flow, weather changes, parking lot status, etc.; Learning algorithms, using deep Q-networks or policy gradient methods to learn optimal policies; using experience replay and target networks to stabilize the learning process; Model training: training in a simulated environment to collect data on the interaction between the charging vehicle and the environment; updating the model parameters (the model in this section refers to the reinforcement learning model) through the reinforcement learning algorithm to maximize the cumulative reward; Model evaluation: evaluate model performance in a test environment to ensure that the model can generalize to unseen situations; monitor key indicators such as charging success rate, average charging time, user satisfaction, etc. Deployment and monitoring: Integrate the trained model into the path planning process of the smart mobile charging vehicle, monitor the model performance in real time, and collect feedback data for further optimization of the model; Continuous learning, the model also supports online learning, enabling it to continuously update strategies based on new data and environmental changes.

[0035] An example of path planning is: using the grid method to create an environmental map (plane navigation map), dividing the working environment into units, and representing them with equal-sized blocks. The selection of the grid size is a very important factor affecting the performance of the planning algorithm. The smaller the grid, the clearer the environmental information represented by the grid map, the greater the storage pressure, and the greater the interference signal, which reduces the path planning speed accordingly. When describing environmental information, the area where the obstacle is located appears black in the grid map and is marked as 1 in the map matrix. The freely traversable area appears white in the grid map and is marked as 0 in the map matrix. The purpose of path planning is to find an optimal traversable path in the established environmental map, such as Figure 5 As shown, according to the method of this embodiment, the lower left corner is the starting point (the starting position of the mobile charging vehicle), the upper right corner is the target point, and after path planning, the optimal path obtained is the path marked with a red line.

Claims

1. A charging path planning method for an intelligent mobile charging vehicle, characterized in that: For a mobile charging vehicle equipped with a camera, display screen, charging manipulator, charging cable, and laser radar, the charging path planning is performed, and the steps include: S1. In the charging scene, the mobile charging vehicle collects three-dimensional data information of the charging scene through cameras and laser radars to build a three-dimensional space model; S2. Capture images of the charging scene through cameras and lidar, establish a coordinate system, mark the current environment and the location of obstacles, and create a plane navigation map; S3. When the mobile charging vehicle receives the charging instruction, it performs global path planning based on the improved PSO algorithm to obtain the optimal path, and the mobile charging vehicle moves according to the optimal path; S4. During the movement, real-time image acquisition is performed through the camera and the laser radar, and the acquired real-time images are spliced ​​and the perspective is corrected to obtain a real-time monitoring plane image; S5. Based on the real-time monitoring plane image, the improved DWA algorithm is used for local path planning, obstacles are avoided in time, the next motion control input is generated through the dynamic window method, the mobile charging vehicle status is updated, and it is verified whether the target point is reached, until the distance between the position of the mobile charging vehicle and the target point does not exceed the set range; S6. Grab the charging cable through the charging manipulator to perform charging operation.

2. According to claim 1, a charging path planning method for an intelligent mobile charging vehicle is characterized in that: The laser radar is set on the front side of the mobile charging vehicle, the display screen is set on the rear side of the mobile charging vehicle, the charging robot is installed on one side of the display screen, the camera is set on the lower side of the display screen, and a charging cable winding rod is also set below the camera. The charging robot is used to plug the charging cable wound on the charging cable winding rod into the charging pile or the vehicle to be charged during charging, and recycle the charging cable after charging is completed.

3. The charging path planning method for an intelligent mobile charging vehicle according to claim 2, characterized in that: The mobile charging vehicle charges itself during low-peak hours and charges the vehicles that need charging during peak hours. When charging vehicles that need charging, the mobile charging vehicle will charge different vehicles continuously based on route planning, and users can make charging reservations based on their mobile phones or clients.

4. The charging path planning method for an intelligent mobile charging vehicle according to claim 1, characterized in that: In S3, the improved PSO algorithm is expressed as: （1）； （2）； In the formula, v i is the velocity of the ith particle, i=1, 2, , I, I is the total number of particles; c1, c2 are learning factors; rand() is a random number between 0 and 1; is the optimal position currently searched for the i-th particle; is the optimal position currently searched by the entire particle swarm; x i is the current position of the ith particle, and equation (2) indicates updating the particle position according to the new velocity; The speed update formula is: （3）； In the formula, is the inertia factor; The inertia weight update strategy based on chaotic mapping makes the inertia weight change strongly nonlinearly with the number of iterations, which can be expressed as: （4）； （5）； In the formula, is the inertia factor of the ith particle; , They represent the upper and lower limits of the inertia factor respectively; k represents the current number of iterations, is the maximum number of iterations of the particle swarm; , Respectively represent the values ​​of chaotic mapping at the kth and k-1th iterations; is the parameter of the chaotic map; rand is a random number between 0 and 1.

5. The charging path planning method of a smart mobile charging vehicle according to claim 4, characterized in that: c1 and c2 are both 2. The value range is 3.6-4.

0.

6. The charging path planning method for an intelligent mobile charging vehicle according to claim 1, characterized in that: In S4, the improved DWA algorithm uses the membership function and the range transformation method to normalize the eigenvalues ​​of the optimal path into fuzzy values, and constructs a membership matrix, which is expressed as: （6）； In the formula, is an element in the membership matrix, which represents the fuzzy value of the mth path plan on the nth evaluation index after the range transformation method; represents the actual value of the mth path plan on the nth evaluation index; the original Convert to a value between 0 and 1 , so that all indicator values ​​are on the same scale, which is convenient for comprehensive evaluation.

7. A charging path planning method for an intelligent mobile charging vehicle according to claim 6, characterized in that: In the improved DWA algorithm, the entropy weight method is used to determine the weight vector. , the entropy of the nth evaluation index is: （7）； （8）； In the formula, h n is the entropy value of the nth evaluation index; is the membership probability of the mth evaluation object under the nth evaluation index, M is the number of evaluation objects, and the evaluation object is the path plan; lg is a logarithmic function, which represents the logarithm with base 10; h n After determination, the entropy weight of the nth evaluation index is: （9）； In the formula, w n is the weight of the nth evaluation index; N is the number of evaluation indicators; According to w n and Get the performance vector of the optimal path solution , thereby calculating the optimal local path solution.

8. The charging path planning method for an intelligent mobile charging vehicle according to claim 1, characterized in that: The mobile charging vehicle is integrated with a weather API. When charging in an open-air environment, it obtains rainfall, snowfall, temperature, wind speed, and visibility in real time. It combines the visual information captured by the camera and uses image recognition technology to identify the real-time weather. In the path planning process of S3-S5, heuristic rules are introduced. According to the impact of weather conditions on driving safety, weather influencing factors are defined, and safety factors are assigned to different weather conditions. Heuristic rules are set based on weather influencing factors and safety factors.

9. A charging path planning method for an intelligent mobile charging vehicle according to claim 8, characterized in that: The open-air environment is an open-air parking lot, and the heuristic rules include: Rainfall heuristics: Avoid charging in low-lying areas or areas with poor drainage during rainfall; give priority to parking spaces with shelter or near buildings to reduce the interference of rain on the charging process; Snowfall heuristic: After a snowfall, choose a parking space that has been cleared of snow; High temperature heuristic: monitor battery temperature and take cooling measures when it exceeds the safety threshold; Low temperature heuristic: preheat the battery before charging to improve charging efficiency; Strong wind heuristic: Avoid charging in open areas during strong winds; Lightning heuristic: suspend charging operations during thunderstorms; Fog heuristic: In foggy weather, use the parking lot lighting system to ensure that the charging vehicle can safely navigate to the parking space; Parking lot capacity heuristic: Prioritize easy-to-access parking spaces during inclement weather conditions; User preference heuristics: Allow users to set sensitivity to weather conditions based on personal preferences, such as not charging in rainy or snowy weather.

10. The charging path planning method of a smart mobile charging vehicle according to claim 1, characterized in that: In charging path planning, an adaptive learning algorithm is introduced to continuously optimize the path planning strategy based on historical charging data and user feedback. The specific process is as follows: Collect data on the operation of mobile charging vehicles under different conditions, including route selection, charging time, user satisfaction, energy consumption, and weather conditions; Preprocess the operation data and extract features after preprocessing, including path length, charging efficiency, waiting time, and the impact of weather changes on charging; Using machine learning technology, we screened out several features that ranked highly in terms of their impact on charging path planning, built a reinforcement learning model, trained the reinforcement learning model using the extracted features and historical charging data, and learned the optimal strategy by simulating the effects of different path planning strategies. Test the performance of the reinforcement learning model on an independent validation set, adjust the model parameters based on the validation results, optimize the model structure, and improve the decision-making quality of the reinforcement learning model.

Images