Route planning method and system for new energy automobile charging pile

Abstract

The invention provides a route planning method and system for a new energy vehicle charging pile, and relates to the field of new energy vehicle charging, and the method comprises the steps: obtaining a vehicle state and charging pile information, and constructing a load state map; calculating a path energy consumption value and a waiting time prediction value, and constructing a value field; reversely calculating a charging position boundary and a time boundary, screening candidate charging piles and decomposing charging tasks; selecting a target charging pile based on the value weight to construct a charging cooperation group; and distributing charging sub-tasks according to the charging power and the value weight, and generating a charging route scheme. The charging waiting time is effectively shortened, the energy utilization efficiency is optimized, and the charging experience is improved.

Description

Technical Field

[0001] This invention relates to new energy vehicle charging technology, and more particularly to a route planning method and system for new energy vehicle charging piles. Background Technology

[0002] With the rapid development of the new energy vehicle industry, the construction and utilization efficiency of charging infrastructure are receiving increasing attention. Charging route planning for new energy vehicles, as a key technology for addressing "range anxiety," is of great significance for improving the user charging experience and optimizing the utilization of charging resources. Currently, new energy vehicle charging route planning mainly relies on charging station location information provided by navigation systems and simple path planning algorithms, allowing users to select suitable charging stations based on their vehicle's current battery level and destination.

[0003] Traditional charging route planning methods primarily rely on static information, such as the geographical location of charging stations and charging power. While this approach can meet basic charging needs, it has several shortcomings. With the development of charging technology and the diversification of user demands, existing technologies can no longer fully meet the actual needs of new energy vehicle users. Current charging route planning methods lack effective prediction and assessment of the real-time load status of charging stations. Most systems cannot accurately predict the waiting time for users upon arrival at charging stations, potentially leading to queues and severely impacting the user's charging experience and travel efficiency.

[0004] Current technologies lack the ability to collaboratively plan multiple charging stations. In long-distance travel or complex routes, multiple charging sessions may be required to complete the entire journey. However, existing methods struggle to effectively decompose the total charging demand into multiple sub-tasks and allocate them reasonably to different charging stations, thus failing to achieve globally optimal charging route planning. Summary of the Invention

[0005] This invention provides a route planning method and system for new energy vehicle charging piles, which can solve the problems in the prior art.

[0006] A first aspect of this invention provides a route planning method for new energy vehicle charging piles, comprising: Obtain vehicle status information and charging station information; For each charging pile in the charging pile information, extract the time distribution characteristics of vehicle arrival at the charging pile and the charging duration distribution characteristics from historical charging data to construct a load status map; Based on the load status map and vehicle status information, the path energy consumption and waiting time prediction of vehicles arriving at each charging station are calculated, and each charging station is assigned a value weight at different times to construct a value field. Based on charging pile information and path energy consumption value, the charging location boundary and charging time boundary are calculated in reverse with vehicle status information as the benchmark. Candidate charging piles are selected according to the charging location boundary, the total charging amount required by the vehicle is calculated, and the total charging amount is decomposed into multiple charging sub-tasks according to the charging power of the candidate charging piles. Based on the value field, target charging piles that meet the preset weight threshold within the charging time boundary are selected from the candidate charging piles to form a charging collaboration group; The charging sub-tasks are allocated to the charging coordination group according to the charging power and value weight of the target charging piles. A charging route plan is generated based on the allocation results of the target charging piles and charging sub-tasks in the charging coordination group.

[0007] For each charging station in the charging station information, the time distribution characteristics of vehicle arrival at the charging station and the charging duration distribution characteristics are extracted from historical charging data to construct a load status map, including: The historical charging data of charging piles is divided into time slices according to time. The rate of change and acceleration of change of the number of vehicles arriving at charging piles in adjacent time slices are calculated. An arrival time-series fluctuation curve is constructed based on the rate of change and acceleration of change. Identify abrupt fluctuation intervals from the arrival time-series fluctuation curve, aggregate time slices into multiple arrival time periods, perform smoothing on the number of vehicles arriving at charging stations within each arrival time period, and generate an arrival smoothing sequence and arrival time period division points; Based on the arrival smooth sequence and the arrival time period division point, the distribution density and dispersion of vehicle charging time within the arrival time period are calculated, a charging time feature curve is constructed, and the arrival smooth sequence and the charging time feature curve are combined to generate a feature descriptor. The information redundancy of the time period feature descriptors is calculated, the time period weight factor is determined based on the information redundancy, and the feature descriptors are fused using the time period weight factor to construct the load status map.

[0008] Based on the load status map and vehicle status information, the path energy consumption and predicted waiting time for vehicles to reach each charging station are calculated. A value weight is assigned to each charging station at different times, constructing a value field that includes: The load status map is converted into a spatiotemporal coordinate sequence to obtain real-time road speed and road congestion status, and to calculate road travel time and congestion index. The driving path from the vehicle location to the charging station is determined based on the travel time and congestion index. Combined with the remaining battery power and target charging amount in the vehicle status information, the path energy consumption value and estimated arrival time of the driving path are calculated, and an energy consumption distribution sequence and arrival time sequence are generated. Based on the energy consumption distribution sequence and arrival time sequence, the charging load value at the corresponding time is extracted from the load state map, and the dynamic ratio of the charging load value to the power supply of the charging pile is calculated. A game payoff function is constructed by combining the dynamic ratio and the charging demand from the vehicle status information. The equilibrium solution of the game payoff function is solved to obtain the number of charging queues. The predicted waiting time is calculated based on the number of queues and the duration of a single charging session. A joint probability distribution is constructed by the energy consumption distribution sequence, arrival time sequence, and waiting time prediction. The mutual information values ​​between variables in the joint probability distribution are extracted to generate a mutual information matrix. The value weight of the charging pile is calculated based on the mutual information matrix, and the value weight is mapped to the time dimension to construct a value field.

[0009] A game theory payoff function is constructed using the dynamic ratio and vehicle status information to determine charging demand. The equilibrium solution of this payoff function is then obtained to determine the number of charging queues. Based on the number of queues and the duration of a single charging session, the predicted waiting time is calculated, including: The dynamic ratio is divided into time segments to obtain the time segment revenue value. The charging demand is extracted from the vehicle status information to obtain the time segment demand value. The time segment revenue value and the time segment demand value are combined in the time dimension to form a revenue mapping matrix. The player's revenue vector is extracted from the revenue mapping matrix to generate the game revenue distribution. Based on the game payoff distribution, a game strategy space is constructed, the strategy selection probability is calculated in the game strategy space, a strategy probability matrix is ​​generated, the equilibrium strategy vector is extracted from the strategy probability matrix, and the optimal response sequence is constructed. The optimal response sequence is expanded into a response probability matrix in the time dimension. The probability distribution value in the response probability matrix is ​​calculated. The game equilibrium point is determined based on the probability distribution value. The game equilibrium point is mapped onto the time segment to obtain the vehicle number distribution within the time period. The vehicle number distribution is multiplied by the single charging time to obtain the waiting time prediction value.

[0010] Based on charging pile information and route energy consumption values, and using vehicle status information as a benchmark, the charging location boundary and charging time boundary are calculated in reverse. Candidate charging piles are selected based on the charging location boundary, and the total charging amount required by the vehicle is calculated. Based on the charging power of the candidate charging piles, the total charging amount is decomposed into multiple charging sub-tasks, including: The remaining battery power and target battery power in the vehicle status information are calculated to obtain the battery power difference. The route energy consumption value and road congestion level are converted into energy consumption compensation coefficients. The battery power difference is adjusted to obtain the actual battery power demand. The upper limit of the driving distance is calculated based on the actual battery power demand to generate the charging location boundary. Calculate the time difference between the expected arrival time and the current time in the vehicle status information, convert the path energy consumption value and the number of charging pile queues into a time compensation coefficient, adjust the time difference according to the time compensation coefficient to obtain the actual time window, calculate the upper limit of driving time based on the actual time window, and generate the charging time boundary. Based on the charging location boundary and charging time boundary, the geographical location and service time period are extracted from the charging pile information. The density value of the geographical location in the spatial distribution and the occupancy rate of the service time period in the time dimension are calculated. Based on the density value and occupancy rate, a list of candidate charging piles that meet the charging location boundary and charging time boundary is selected. The total charging amount is calculated from the target battery level and remaining battery level in the vehicle status information. The charging power and total charging amount in the candidate charging pile list are combined to generate a task allocation matrix. The charging amount interval is divided according to the task allocation matrix, and a charging sub-task list is output.

[0011] Based on the value field, target charging piles that satisfy a preset weight threshold within the charging time boundary are selected from the candidate charging piles to form a charging collaboration group, including: The value weight sequence of candidate charging piles within the charging time boundary is obtained from the value field. The value weight sequence is mapped to the time dimension to form a value distribution curve. The value distribution curve is divided into intervals to generate a time period value matrix. The value overlap between charging piles is calculated based on the time period value matrix. The value overlap between charging piles is compared with a preset weight threshold to calculate the value matching degree. Based on the value matching degree, charging piles that meet the preset weight threshold requirements are selected as target charging piles, and a target charging pile value mapping map is generated. The overlapping area of ​​value between charging piles is calculated by the value mapping map of the target charging piles. The overlapping area is normalized to obtain the connectivity value. The connection matrix of the target charging piles is constructed based on the connectivity value. The maximum connected component is extracted from the connection matrix to form a charging collaboration group.

[0012] The charging sub-tasks are allocated to the charging coordination group according to the charging power and value weight of the target charging piles. Based on the allocation results of the target charging piles and charging sub-tasks in the charging coordination group, a charging route plan is generated, including: The charging power of the target charging pile is mapped according to the service period to obtain the charging power distribution sequence, and the value weight is mapped according to the service period to obtain the weight distribution sequence. Calculate the time-period supply capacity for the charging power distribution sequence, calculate the time-period service value for the weight distribution sequence, combine the time-period supply capacity and time-period service value to form the target charging pile service characteristics, and construct task characteristics by distributing the demand for charging sub-tasks according to the time-period distribution. The matching degree is calculated based on service characteristics and task characteristics, and a task allocation matrix is ​​constructed. The task allocation matrix is ​​divided into sub-blocks according to the charging coordination group. The similarity between the service features and task features within the sub-block is calculated to obtain the matching coefficient. The charging sub-tasks are sorted according to the matching coefficient to generate a task sequence. The charging sub-tasks in the task sequence are allocated to the charging coordination group according to the matching coefficient to form an allocation scheme. The task load value of the charging coordination group is calculated based on the charging coordination group allocation scheme. A charging service network is constructed according to the charging coordination group allocation scheme and the task load value of the charging coordination group. The path combination between the charging sub-task and the target charging pile is extracted in the charging service network. The distance cost and time cost of the path combination are calculated to obtain the path evaluation value. The path combination with the best path evaluation value is selected to generate a charging route scheme.

[0013] A second aspect of this invention provides a route planning system for new energy vehicle charging piles, comprising: The first unit is used to obtain vehicle status information and charging pile information; The second unit is used to extract the time distribution characteristics of vehicle arrival at the charging pile and the charging duration distribution characteristics from historical charging data for each charging pile in the charging pile information, and to construct a load status map. The third unit is used to calculate the path energy consumption value and waiting time prediction value of the vehicle to each charging pile based on the load status map and vehicle status information, and to assign value weight to each charging pile at different times to construct a value field. The fourth unit is used to calculate the charging location boundary and charging time boundary in reverse based on the charging pile information and the path energy consumption value, with the vehicle status information as the benchmark. It then filters candidate charging piles according to the charging location boundary, calculates the total charging amount required by the vehicle, and decomposes the total charging amount into multiple charging sub-tasks according to the charging power of the candidate charging piles. The fifth unit is used to select target charging piles from candidate charging piles based on the value field, and construct charging collaboration groups that meet the preset weight threshold within the charging time boundary. The sixth unit is used to allocate charging sub-tasks to the charging coordination group according to the charging power and value weight of the target charging piles, and generate charging route plans based on the allocation results of the target charging piles and charging sub-tasks in the charging coordination group.

[0014] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0016] In this embodiment, by constructing a charging pile load status map and value field, accurate prediction of charging pile usage is achieved, avoiding excessive waiting time for vehicles at charging piles and improving the time efficiency of the charging process. By reverse-calculating charging location boundaries and charging time boundaries, and combining this with the value weight of charging piles for selection, this invention can rationally plan charging routes, reduce vehicle energy consumption, extend driving range, and improve the user experience of new energy vehicles. This invention decomposes the total charging amount into multiple charging sub-tasks and allocates them to charging coordination groups, achieving optimized allocation of charging demand, making the charging process more flexible and efficient, while reducing the load pressure on individual charging piles, which is conducive to the balanced utilization of charging infrastructure and the coordinated operation of the entire new energy vehicle charging network. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the route planning method for new energy vehicle charging piles according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the charging game and waiting time prediction analysis under a dynamic electricity price environment according to an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0020] Figure 1 This is a flowchart illustrating the route planning method for new energy vehicle charging piles according to an embodiment of the present invention. Figure 1 As shown, the method includes: Obtain vehicle status information and charging station information; For each charging pile in the charging pile information, extract the time distribution characteristics of vehicle arrival at the charging pile and the charging duration distribution characteristics from historical charging data to construct a load status map; Based on the load status map and vehicle status information, the path energy consumption and waiting time prediction of vehicles arriving at each charging station are calculated, and each charging station is assigned a value weight at different times to construct a value field. Based on charging pile information and path energy consumption value, the charging location boundary and charging time boundary are calculated in reverse with vehicle status information as the benchmark. Candidate charging piles are selected according to the charging location boundary, the total charging amount required by the vehicle is calculated, and the total charging amount is decomposed into multiple charging sub-tasks according to the charging power of the candidate charging piles. Based on the value field, target charging piles that meet the preset weight threshold within the charging time boundary are selected from the candidate charging piles to form a charging collaboration group; The charging sub-tasks are allocated to the charging coordination group according to the charging power and value weight of the target charging piles. A charging route plan is generated based on the allocation results of the target charging piles and charging sub-tasks in the charging coordination group.

[0021] In one optional implementation, for each charging pile in the charging pile information, the time distribution characteristics of vehicle arrival at the charging pile and the charging duration distribution characteristics are extracted from historical charging data to construct a load status map, including: The historical charging data of charging piles is divided into time slices according to time. The rate of change and acceleration of change of the number of vehicles arriving at charging piles in adjacent time slices are calculated. An arrival time-series fluctuation curve is constructed based on the rate of change and acceleration of change. Identify abrupt fluctuation intervals from the arrival time-series fluctuation curve, aggregate time slices into multiple arrival time periods, perform smoothing on the number of vehicles arriving at charging stations within each arrival time period, and generate an arrival smoothing sequence and arrival time period division points; Based on the arrival smooth sequence and the arrival time period division point, the distribution density and dispersion of vehicle charging time within the arrival time period are calculated, a charging time feature curve is constructed, and the arrival smooth sequence and the charging time feature curve are combined to generate a feature descriptor. The information redundancy of the time period feature descriptors is calculated, the time period weight factor is determined based on the information redundancy, and the feature descriptors are fused using the time period weight factor to construct the load status map.

[0022] For a given charging station's historical charging data, it is divided into multiple time slices along the time dimension. For example, a 24-hour day can be divided into 96 time slices at 15-minute intervals, labeled t1, t2, ..., t96. For adjacent time slices, the rate of change and acceleration of change of the number of vehicles arriving at the charging station are calculated. The rate of change is represented by the difference in the number of vehicles arriving in the current time slice compared to the previous time slice. For example, if 5 vehicles arrive at time t2 and 8 vehicles arrive at time t3, then the rate of change at time t3 is 3 vehicles. The acceleration of change is represented by the change in the rate of change. For example, if the rate of change at time t2 is 2 vehicles and the rate of change at time t3 is 3 vehicles, then the acceleration of change at time t3 is 1 vehicle. By connecting the rate of change and acceleration values ​​of each time slice, arrival rate of change curves and acceleration of change curves are constructed respectively. These two curves together constitute the arrival time-series fluctuation curve.

[0023] Based on the arrival time-series fluctuation curve, fluctuation abrupt change intervals are identified. A fluctuation abrupt change interval refers to a set of time slices where the rate of change or acceleration exceeds a preset threshold. The rate of change threshold is set to 5 vehicles / time slice, and the acceleration threshold is set to 3 vehicles / time slice. When the rate of change in a time slice exceeds 5 vehicles or the acceleration exceeds 3 vehicles, that time slice is marked as a fluctuation abrupt change point. Adjacent fluctuation abrupt change points can be merged to form fluctuation abrupt change intervals. Taking a certain charging pile as an example, analysis of its historical data reveals that the rate of change consistently exceeds the threshold between t10 and t15, forming a fluctuation abrupt change interval. Similar situations exist between t35 and t40 and between t60 and t65, thus determining that this charging pile has three fluctuation abrupt change intervals per day.

[0024] Based on the fluctuation and abrupt change intervals, time slices are aggregated into multiple arrival time periods. The start point of each fluctuation and abrupt change interval serves as the end point of the previous time period, and the end point serves as the start point of the next time period. For the example above, four arrival time periods can be defined: t1 to t9, t16 to t34, t41 to t59, and t66 to t96. A smoothing process is performed on the number of vehicles arriving at charging stations within each arrival time period to generate a smoothed arrival sequence. The smoothing process can use a moving average method; for example, the smoothed value for each moment is obtained by averaging the number of arriving vehicles at each time point within the t1 to t9 time period with the number of arriving vehicles in the two time slices before and after it. By connecting the smoothed sequences of each time period, a complete smoothed arrival sequence is formed. Simultaneously, the boundary points of each time period constitute a set of arrival time period division points, such as {t9, t16, t34, t41, t59, t66}.

[0025] Based on the arrival smoothing sequence and arrival time period division points, the distribution density and dispersion of vehicle charging duration within each arrival time period are calculated. Distribution density refers to the proportion of vehicles within a specific charging duration range, and dispersion refers to the standard deviation of the charging duration. For example, within the time period t1 to t9, the charging duration of all vehicles is statistically analyzed, and the proportion of vehicles in different duration ranges such as 0-1 hour, 1-2 hours, and 2-3 hours is calculated to obtain the charging duration distribution density. Taking a certain charging station as an example, within the time period t1 to t9, the proportion of vehicles charging for 0-1 hours is 20%, 1-2 hours is 50%, 2-3 hours is 25%, and more than 3 hours is 5%, with a standard deviation of 0.8 hours. Similar calculations are performed for each arrival time period to construct a charging duration feature curve. The arrival smoothing sequence is combined with the charging duration feature curve to generate a feature descriptor for each time period. The feature descriptor includes a smoothed value of the number of arriving vehicles, as well as the distribution density and dispersion values ​​for different charging duration ranges.

[0026] Information redundancy is calculated for time-period feature descriptors, and time-period weighting factors are determined based on this redundancy. Information redundancy represents the degree of repetitive or predictable information in a feature descriptor. Time periods with high redundancy contribute less to the representation of the overall load status and should be assigned lower weights; time periods with low redundancy contain more unique information and should be assigned higher weights. For time periods with obvious regularity (such as nighttime off-peak periods), their feature descriptors have higher information redundancy and can be assigned a weighting factor of 0.5; for time periods with drastic changes (such as morning and evening peak periods), their feature descriptors have lower information redundancy and can be assigned a weighting factor of 1.5. The feature descriptors are then fused using the time-period weighting factors to construct a load status map.

[0027] The load state map is represented in two-dimensional matrix form, with rows representing the time dimension and columns representing the charging duration dimension. The matrix element values ​​represent the load intensity at a specific time and for a specific charging duration. The load state map constructed using the above method can intuitively display the usage characteristics of charging piles in different time periods, providing a data foundation for charging station operation optimization and power load forecasting.

[0028] In this embodiment, by jointly modeling the arrival rhythm of vehicles and the characteristics of charging duration, the actual load variation of charging piles in different time periods can be accurately depicted. Fluctuation regions are identified using the rate of change and acceleration, improving the sensitivity of load time series analysis to sudden congestion or idle periods. A more stable arrival characteristic is constructed by smoothing sequences and time period division points, protecting the load description from interference from local anomalies. Integrating the charging duration distribution density with arrival volume characteristics comprehensively reflects the coupling relationship between load intensity and persistence.

[0029] In one optional implementation, based on the load state map and vehicle state information, the path energy consumption and predicted waiting time for vehicles to reach each charging station are calculated, and a value field is constructed by assigning a value weight to each charging station at different times, including: The load status map is converted into a spatiotemporal coordinate sequence to obtain real-time road speed and road congestion status, and to calculate road travel time and congestion index. The driving path from the vehicle location to the charging station is determined based on the travel time and congestion index. Combined with the remaining battery power and target charging amount in the vehicle status information, the path energy consumption value and estimated arrival time of the driving path are calculated, and an energy consumption distribution sequence and arrival time sequence are generated. Based on the energy consumption distribution sequence and arrival time sequence, the charging load value at the corresponding time is extracted from the load state map, and the dynamic ratio of the charging load value to the power supply of the charging pile is calculated. A game payoff function is constructed by combining the dynamic ratio and the charging demand from the vehicle status information. The equilibrium solution of the game payoff function is solved to obtain the number of charging queues. The predicted waiting time is calculated based on the number of queues and the duration of a single charging session. A joint probability distribution is constructed by the energy consumption distribution sequence, arrival time sequence, and waiting time prediction. The mutual information values ​​between variables in the joint probability distribution are extracted to generate a mutual information matrix. The value weight of the charging pile is calculated based on the mutual information matrix, and the value weight is mapped to the time dimension to construct a value field.

[0030] When constructing the value field of charging piles, the first step is to convert the load state map into a spatiotemporal coordinate sequence. During this conversion, the load intensity value for each time period in the load state map is associated with its corresponding geographical location information, forming a triplet sequence containing time, space, and load intensity. Simultaneously, real-time road speed and congestion data are acquired; this data can be obtained through traffic monitoring systems or map service interfaces. The road is segmented, and the travel time for each segment is obtained by dividing its length by the real-time speed. The congestion index can be calculated based on the ratio of real-time speed to the historical average speed; the lower the real-time speed, the higher the congestion index.

[0031] Based on travel time and congestion index, Dijkstra's algorithm is used to determine the optimal driving path from the vehicle's current location to each charging station. This algorithm treats the road network as a weighted graph, where the weight of each edge is determined by the weighted sum of travel time and congestion index. The path with the smallest weight is selected as the optimal driving path. Combining the remaining battery power, battery capacity, and energy consumption per unit distance data from the vehicle's status information, the energy consumption value of the selected path is calculated. The calculation method involves accumulating the length of each road segment on the path multiplied by the energy consumption per unit distance under that road condition, considering the impact of environmental factors such as slope and temperature on energy consumption. By accumulating the travel time of each segment of the path, the estimated time for the vehicle to reach the charging station from its current location is obtained. The above calculation is repeated for all reachable charging stations to generate an energy consumption distribution sequence and an arrival time sequence.

[0032] Taking an electric vehicle as an example, its current remaining battery power is 30%, the battery capacity is 60 kWh, and the energy consumption per unit distance is 15 kWh / 100 km. The vehicle plans to charge at three candidate charging stations. Based on the above calculations, the energy consumption for the route to charging station A is 5 kWh, with an estimated arrival time of 10:30; the energy consumption for the route to charging station B is 7 kWh, with an estimated arrival time of 10:40; and the energy consumption for the route to charging station C is 4 kWh, with an estimated arrival time of 10:45.

[0033] Based on the energy consumption distribution sequence and arrival time sequence, the charging load value at the corresponding time moment is extracted from the load state spectrum. The charging load value represents the usage intensity of the charging pile at a specific time. The dynamic ratio is obtained by dividing the charging load value by the power supply of the charging pile. This ratio reflects the saturation level of the charging pile at the expected arrival time; a higher ratio indicates a higher charging pile utilization rate and fewer available resources. The power supply of charging piles A, B, and C are 60 kW, 120 kW, and 90 kW, respectively. At the expected arrival time, their charging load values ​​extracted from the load state spectrum are 45 kW, 60 kW, and 36 kW, respectively, and the calculated dynamic ratios are 0.75, 0.5, and 0.4, respectively.

[0034] A game-theoretic payoff function is constructed by combining the dynamic ratio with the charging demand from vehicle status information. This payoff function considers factors such as charging pile utilization, charging speed, and price, simulating the decision-making process of multiple users simultaneously choosing charging piles. Specifically, the payoff function consists of four parts: charging duration, waiting time, charging cost, and convenience, each assigned a different weight. The equilibrium solution of the game-theoretic payoff function is solved to obtain the expected queue size for each charging pile. Based on the queue size and the historical average charging duration per session, the predicted waiting time is calculated. The charging duration per session can be obtained from the charging duration distribution characteristics in the load status map. The expected queue sizes for charging piles A, B, and C are 2, 1, and 0, respectively, and the historical average charging duration per session is 40 minutes, 35 minutes, and 45 minutes, respectively. The calculated predicted waiting times are 80 minutes, 35 minutes, and 0 minutes, respectively.

[0035] Next, a joint probability distribution is constructed from three variables: energy consumption distribution sequence, arrival time sequence, and predicted waiting time. This joint probability distribution describes the relationships between these three variables and is learned from historical data using kernel density estimation. Mutual information values ​​are extracted from the joint probability distribution to generate a mutual information matrix. The mutual information value quantifies the degree of interdependence between variables; a higher value indicates a stronger dependency. The value weight of each charging station is calculated based on the mutual information matrix. This calculation method involves normalizing the eigenvalues ​​corresponding to the eigenvectors of the mutual information matrix. Larger eigenvalues ​​correspond to higher value weights for the corresponding charging stations. Finally, these value weights are mapped to the time dimension, assigning different value weights to each charging station at different times, thus constructing a value field.

[0036] The value weights of charging piles A, B, and C at the current moment are calculated using a mutual information matrix: 0.6, 0.85, and 0.95, respectively. Extending these weights to the time dimension reveals the changes in value weights at different times of the day. For example, the value weight of charging pile C drops to 0.7 during peak hours (17:00-19:00) and rises to 0.99 during off-peak hours (23:00-05:00 the next day). This value field intuitively reflects the advantages and disadvantages of different charging piles at different times, providing a basis for subsequent route planning decisions.

[0037] The charging pile value field constructed through this implementation method comprehensively considers multiple factors such as path energy consumption, traffic conditions, arrival time, and waiting time, providing comprehensive decision support for new energy vehicle charging route planning. The value field dynamically reflects the availability and cost-effectiveness of charging pile resources, enabling vehicles to select the most suitable charging pile for their own conditions, avoiding congestion points and reducing waiting time. Simultaneously, the construction process of the value field integrates load forecasting, path planning, and queuing theory, improving the accuracy and practicality of route planning. By capturing the dependencies between variables through mutual information analysis, the adaptability to complex scenarios is enhanced.

[0038] like Figure 2 The diagram illustrates the charging game and waiting time prediction analysis under the dynamic electricity price environment of this embodiment.

[0039] In one optional implementation, a game payoff function is constructed using the dynamic ratio and charging demand from vehicle status information. The equilibrium solution of the game payoff function is solved to obtain the charging queue size. Based on the queue size and the duration of a single charging session, a predicted waiting time is calculated, including: The dynamic ratio is divided into time segments to obtain the time segment revenue value. The charging demand is extracted from the vehicle status information to obtain the time segment demand value. The time segment revenue value and the time segment demand value are combined in the time dimension to form a revenue mapping matrix. The player's revenue vector is extracted from the revenue mapping matrix to generate the game revenue distribution. Based on the game payoff distribution, a game strategy space is constructed, the strategy selection probability is calculated in the game strategy space, a strategy probability matrix is ​​generated, the equilibrium strategy vector is extracted from the strategy probability matrix, and the optimal response sequence is constructed. The optimal response sequence is expanded into a response probability matrix in the time dimension. The probability distribution value in the response probability matrix is ​​calculated. The game equilibrium point is determined based on the probability distribution value. The game equilibrium point is mapped onto the time segment to obtain the vehicle number distribution within the time period. The vehicle number distribution is multiplied by the single charging time to obtain the waiting time prediction value.

[0040] In practice, the dynamic ratio is divided into time segments to obtain the revenue value for each time segment. The dynamic ratio represents the ratio of the load to the power supply of the charging pile at a specific moment, reflecting the scarcity of charging resources. Each time segment can be set to 15 minutes, with a day divided into 96 time segments. The dynamic ratio in each time segment is extracted to form a time series. For a certain charging pile on a weekday, the dynamic ratios for the four time segments from 8:00 AM to 9:00 AM are 0.65, 0.73, 0.82, and 0.78, respectively. Therefore, the revenue values ​​for these four time segments are also 0.65, 0.73, 0.82, and 0.78.

[0041] Charging demand is extracted from vehicle status information to obtain time-segment demand values. Charging demand includes factors such as the vehicle's target charging amount, required charging rate, and preferred charging time period. For vehicles currently requiring route planning, data such as current remaining battery power, target charging amount, and expected charging time can be extracted from their status information. When converting charging demand into time-segment demand values, the relationship between charging rate and time is considered. For example, an electric vehicle needs 40 kWh of charging, expects to use a 60 kW fast charger, and has an ideal charging time of 40 minutes, spanning three time segments. Decomposing this charging demand into each time segment, we find that the first two time segments each require 15 kWh, and the third time segment requires 10 kWh, corresponding to time-segment demand values ​​of 0.25, 0.25, and 0.167.

[0042] A revenue mapping matrix is ​​formed by combining the revenue and demand values ​​for each time segment along the time dimension. This matrix is ​​a two-dimensional table where rows represent different charging stations, columns represent different time segments, and matrix elements represent the comprehensive revenue of that charging station within that time segment. The comprehensive revenue is calculated as a weighted average of the revenue and demand values ​​for each time segment, with weights adjustable according to user preferences. Assuming weights of 0.6 and 0.4, the corresponding comprehensive revenues for the time segments are 0.65×0.6+0.25×0.4=0.49, 0.73×0.6+0.25×0.4=0.538, and 0.82×0.6+0.167×0.4=0.5588, respectively. The revenue vector for each player is extracted from the revenue mapping matrix to generate a game payoff distribution. Players include all vehicle users who might use the charging station, and the payoff vector represents the expected revenue under different strategy choices.

[0043] Based on the game payoff distribution, a game strategy space is constructed. This space contains all possible combinations of charging decisions, such as choosing different charging stations, different arrival times, or different charging amounts. In this embodiment, the strategy primarily refers to choosing which time segment to charge. Each player has multiple possible choices, forming a strategy set. The strategy selection probabilities are calculated in the game strategy space using a mixed-strategy Nash equilibrium solution method. This method is based on an iterative optimal response algorithm, starting from an arbitrary initial strategy distribution and gradually adjusting the strategy choices of each player until a stable state is reached. A strategy probability matrix is ​​generated, where each element represents the probability of each player choosing a different strategy. An equilibrium strategy vector is extracted from the strategy probability matrix, representing the optimal strategy choice for each player in the equilibrium state. An optimal response sequence is constructed, recording the strategy adjustment path to reach the equilibrium state.

[0044] The optimal response sequence is expanded into a response probability matrix over time. This matrix represents the probability distribution of each player's chosen response in different time segments. The probability distribution values ​​in the response probability matrix are calculated using Monte Carlo simulation, employing extensive random sampling to determine the frequency of various strategy combinations. Based on these probability distribution values, the game equilibrium point is determined. The equilibrium point represents a state where the probabilities of each player's strategy choice are stable, at which point no player has an incentive to unilaterally change their strategy. Mapping the equilibrium point onto time segments yields the expected distribution of vehicle numbers within each time segment. For example, in the time segments of the above example, the expected vehicle numbers calculated through game equilibrium are 5, 7, and 4, respectively.

[0045] Multiplying the vehicle quantity distribution by the single charging duration yields the predicted waiting time. The single charging duration can be obtained from historical charging data statistics or estimated based on the charging pile power and average charging amount. For a 60 kW fast charging pile, the average single charging time is 30 minutes. Considering the number of charging piles and the queuing service mechanism, the waiting time is calculated. Assuming the charging station has 3 charging piles of the same type, using a first-come, first-served principle, the predicted waiting times for the above time segments are 5 / 3 × 30 = 50 minutes, 7 / 3 × 30 = 70 minutes, and 4 / 3 × 30 = 40 minutes, respectively. This means that if a vehicle arrives at the charging station in the second time segment, it is expected to wait 70 minutes before starting charging.

[0046] In this embodiment, dynamic prediction of charging station congestion is achieved by accurately modeling user decision-making behavior, overcoming the shortcomings of traditional methods that ignore users' strategic choices. Using game theory equilibrium analysis instead of simple queuing theory more accurately reflects the complex scenario of multiple users competing for limited resources. Fine-grained time dimension partitioning improves prediction accuracy, making charging route planning more targeted. The introduction of the response probability matrix enables the system to cope with the randomness and uncertainty of user behavior, enhancing the robustness of the prediction model. This not only improves the accuracy of waiting time prediction but also provides a reliable basis for the rational scheduling of charging pile resources and user travel decisions, effectively improving the charging experience for new energy vehicle users.

[0047] In one optional implementation, based on charging pile information and path energy consumption values, the charging location boundary and charging time boundary are calculated in reverse using vehicle status information as a reference. Candidate charging piles are selected according to the charging location boundary, the total charging amount required by the vehicle is calculated, and the total charging amount is decomposed into multiple charging sub-tasks according to the charging power of the candidate charging piles, including: The remaining battery power and target battery power in the vehicle status information are calculated to obtain the battery power difference. The route energy consumption value and road congestion level are converted into energy consumption compensation coefficients. The battery power difference is adjusted to obtain the actual battery power demand. The upper limit of the driving distance is calculated based on the actual battery power demand to generate the charging location boundary. Calculate the time difference between the expected arrival time and the current time in the vehicle status information, convert the path energy consumption value and the number of charging pile queues into a time compensation coefficient, adjust the time difference according to the time compensation coefficient to obtain the actual time window, calculate the upper limit of driving time based on the actual time window, and generate the charging time boundary. Based on the charging location boundary and charging time boundary, the geographical location and service time period are extracted from the charging pile information. The density value of the geographical location in the spatial distribution and the occupancy rate of the service time period in the time dimension are calculated. Based on the density value and occupancy rate, a list of candidate charging piles that meet the charging location boundary and charging time boundary is selected. The total charging amount is calculated from the target battery level and remaining battery level in the vehicle status information. The charging power and total charging amount in the candidate charging pile list are combined to generate a task allocation matrix. The charging amount interval is divided according to the task allocation matrix, and a charging sub-task list is output.

[0048] First, the remaining battery level and target battery level data are extracted from the vehicle status information. Remaining battery level refers to the current charge level of the electric vehicle's battery, usually expressed as a percentage or in kilowatt-hours (kWh); the target battery level refers to the battery charge level the user expects to achieve. When calculating the charge difference, the target battery level is subtracted from the remaining battery level to obtain the theoretically required additional charge. For an electric vehicle with a battery capacity of 60 kWh, a current remaining charge of 15 kWh (25%), and a target charge of 54 kWh (90%), the charge difference is 39 kWh.

[0049] The path energy consumption value and road congestion level affect the actual charging demand, therefore, it needs to be converted into an energy consumption compensation coefficient to adjust the energy difference. The path energy consumption value refers to the estimated energy consumption of a vehicle from its current location to the charging station, calculated using a vehicle driving model. Road congestion level can be obtained through traffic flow data, generally divided into four levels: smooth flow, light congestion, moderate congestion, and severe congestion. The energy consumption compensation coefficient is calculated by adding a base coefficient (set to 1) to an additional coefficient corresponding to the congestion level. The additional coefficient is 0 for smooth flow, 0.05 for light congestion, 0.1 for moderate congestion, and 0.2 for severe congestion. If the road is in a moderate congestion state, the energy consumption compensation coefficient is 1.1, indicating that the actual energy consumption will increase by 10% compared to normal. Multiplying the path energy consumption value by the energy consumption compensation coefficient and adding the energy difference yields the actual energy demand. If the energy consumption of the route is 5 kWh, the road is in a state of mild congestion, and the energy consumption compensation coefficient is 1.05, then the actual electricity demand is 5 × 1.05 + 39 = 44.25 kWh.

[0050] The maximum driving distance is calculated based on actual battery demand. The maximum driving distance refers to the farthest distance a vehicle can travel after a full charge. The calculation method is to subtract the actual battery demand from the vehicle's battery capacity when fully charged, and then divide by the vehicle's energy consumption per unit distance. For example, if the vehicle's full charge capacity is 60 kWh, the energy consumption per unit distance is 15 kWh / 100 km, and the actual battery demand is 44.25 kWh, then the maximum driving distance is (60-44.25) / 15×100=105 km. This means that the charging location boundary is a circular area with a radius of 105 km centered on the vehicle's current location. Any charging station outside this boundary is unsuitable as a candidate charging station because the vehicle, even after being fully charged, will not be able to reach its destination once it reaches these charging stations.

[0051] The expected arrival time and current time are extracted from the vehicle status information, and the time difference is calculated. The expected arrival time refers to the time when the user expects to arrive at the destination, and the current time refers to the time when the charging route planning began. The time difference represents the total time the user has available for driving and charging. For example, if the current time is 8:00 AM and the expected arrival time is 2:00 PM, the time difference is 6 hours.

[0052] The path energy consumption value and the number of charging piles in the queue are converted into a time compensation coefficient to adjust the time difference. A higher path energy consumption value indicates a longer path to the charging pile or higher energy consumption, requiring more travel time; a larger number of charging piles in the queue indicates a longer waiting time for charging. The time compensation coefficient is calculated by subtracting the discount coefficients corresponding to energy consumption and queue number from the base coefficient (set to 1). The energy consumption discount coefficient is the path energy consumption value divided by the vehicle's full battery capacity, then multiplied by 0.5; the queue discount coefficient is the queue number multiplied by 0.1. If the path energy consumption value is 5 kWh, the vehicle's full battery capacity is 60 kWh, and the number of charging piles in the queue is 2, then the time compensation coefficient is 1 - (5 / 60 × 0.5) - (2 × 0.1) = 0.7583. The time difference is adjusted based on the time compensation coefficient to obtain the actual time window, calculated by multiplying the time difference by the time compensation coefficient. If the time difference is 6 hours and the time compensation coefficient is 0.7583, then the actual time window is 6 × 0.7583 = 4.55 hours.

[0053] The maximum driving time is calculated based on the actual time window. The maximum driving time refers to the longest time a user can use for driving. It is calculated by subtracting the minimum charging time from the actual time window. The minimum charging time is related to the charging power and the amount of electricity needed, and can generally be estimated based on the fastest charging rate. If the actual electricity demand is 44.25 kWh and the fastest charging power is 120 kW, then the minimum charging time is 44.25 / 120 = 0.37 hours. The maximum driving time is 4.55 - 0.37 = 4.18 hours. The charging time boundary means that the vehicle must arrive at a charging station within 4.18 hours after the current time to begin charging; otherwise, it will not be able to complete charging and reach its destination before the expected arrival time.

[0054] Based on charging location boundaries and charging time boundaries, geographical location and service period data are extracted from the charging pile information. Geographical location includes the latitude and longitude coordinates of the charging pile, and service period includes the operating hours of the charging pile. The density value of the geographical location in spatial distribution is calculated, i.e., the number of charging piles per unit area within the charging location boundary. The occupancy rate of the service period in the time dimension is calculated, i.e., the reservation and occupancy status of the charging pile within the charging time boundary. A higher density value indicates a denser concentration of charging piles in the area, providing users with more choices; a higher occupancy rate indicates a higher utilization rate of the charging pile, potentially requiring queuing. Based on the density value and occupancy rate, a filtering threshold is set, retaining charging piles with high density values ​​and low occupancy rates to form a candidate charging pile list.

[0055] The total charging amount is calculated from the target and remaining battery levels in the vehicle status information, i.e., the actual amount of electricity needed to charge. If the target battery level is 54 kWh and the remaining battery level is 15 kWh, the total charging amount is 39 kWh. The charging power data from the candidate charging station list is combined with the total charging amount to generate a task allocation matrix. The rows of the task allocation matrix represent different candidate charging stations, and the columns represent different charging amount allocation schemes. If there are three candidate charging stations with power ratings of 60 kW, 90 kW, and 120 kW, and a total charging amount of 39 kWh, multiple allocation schemes can be considered, such as charging all charging at one charging station or charging at multiple charging stations. The charging amount is divided into intervals based on the task allocation matrix, generating a list of charging sub-tasks. A charging sub-task refers to the task of charging a specific amount of electricity at a specific charging station. For example, the total charging amount of 39 kWh can be decomposed into two sub-tasks: charging 15 kWh at a 60 kW charging station and charging 24 kWh at a 120 kW charging station.

[0056] In this embodiment, by accurately calculating the charging location boundaries and charging time boundaries, the key constraint problem in electric vehicle charging route planning is effectively solved. The introduction of energy consumption compensation coefficients and time compensation coefficients enables the system to more accurately assess the impact of actual driving conditions on battery capacity and time, improving the practicality of the planning. The candidate charging pile screening mechanism based on geographical location density values ​​and service period occupancy rates considers both charging convenience and waiting time, balancing user experience and charging efficiency. The charging task decomposition technology breaks through the limitations of traditional single-point charging, achieving optimized matching of charging time and travel arrangements through a multi-point collaborative charging strategy, providing users with more flexible and diverse charging options, and significantly improving the convenience of using new energy vehicles and travel efficiency.

[0057] In one optional implementation, selecting target charging piles from candidate charging piles that satisfy a preset weight threshold within the charging time boundary, based on the value field, to construct a charging collaboration group includes: The value weight sequence of candidate charging piles within the charging time boundary is obtained from the value field. The value weight sequence is mapped to the time dimension to form a value distribution curve. The value distribution curve is divided into intervals to generate a time period value matrix. The value overlap between charging piles is calculated based on the time period value matrix. The value overlap between charging piles is compared with a preset weight threshold to calculate the value matching degree. Based on the value matching degree, charging piles that meet the preset weight threshold requirements are selected as target charging piles, and a target charging pile value mapping map is generated. The overlapping area of ​​value between charging piles is calculated by the value mapping map of the target charging piles. The overlapping area is normalized to obtain the connectivity value. The connection matrix of the target charging piles is constructed based on the connectivity value. The maximum connected component is extracted from the connection matrix to form a charging collaboration group.

[0058] A value field refers to a dataset describing the distribution of charging pile service value across time and space, encompassing a comprehensive evaluation of information such as charging pile location, service hours, charging power, and idle status. When obtaining the value weight sequence of candidate charging piles within the charging time boundary from the value field, it is necessary to extract the value weight data of each candidate charging pile at different time points. Value weight measures the service value of a charging pile to a user at a specific moment, calculated by weighting factors such as charging pile power, location convenience, waiting time, and service quality. Value weight values ​​typically range from 0 to 1, with higher values ​​indicating higher value to the user. For example, for a candidate charging pile, the possible value weight sequence within the charging time boundary is [0.65, 0.72, 0.78, 0.81, 0.85, 0.82, 0.79, 0.73], representing the change in value weight of the charging pile over eight consecutive time units.

[0059] The value weight sequence is mapped to the time dimension to form a value distribution curve. This involves using the value weight value at each time point as the ordinate and time as the abscissa, plotting a continuously changing curve. For the aforementioned value weight sequence, eight time points can be marked on the time axis, from t1 to t8, with the corresponding value weight values ​​as the ordinate. This value distribution curve visually demonstrates the changing trend of the charging pile's value over time.

[0060] The value distribution curve is divided into intervals to generate a time-period value matrix. The purpose of interval division is to discretize the continuous value distribution curve, facilitating the calculation and comparison of the value of different charging piles within the same time period. Interval division methods can include equal-length division or key-time-point division. Equal-length division divides the entire charging time boundary into several equal time periods; key-time-point division is based on inflection points of value change or characteristic points of user activity. Taking equal-length division as an example, the 8 time points are divided into 4 time periods, with each time period containing 2 time points. The rows of the time-period value matrix represent different charging piles, the columns represent different time periods, and the matrix element values ​​are the average value weight of the charging pile within that time period. For the aforementioned candidate charging piles, the average value weights for the 4 time periods are (0.65+0.72) / 2=0.685, (0.78+0.81) / 2=0.795, (0.85+0.82) / 2=0.835, and (0.79+0.73) / 2=0.76, respectively.

[0061] The value overlap between charging piles is calculated based on the time-period value matrix. Value overlap refers to the similarity of the value weights of two charging piles within the same time period, used to measure whether two charging piles can form an effective charging cooperation relationship. The calculation method is to sum the absolute values ​​of the value weight differences between the two charging piles in each time period, and then subtract 1 from the ratio of this sum to the number of time periods. If the time-period values ​​of charging piles A and B are [0.685, 0.795, 0.835, 0.76] and [0.71, 0.82, 0.79, 0.73] respectively, then the sum of the absolute values ​​of their value differences is |0.685-0.71|+|0.795-0.82|+|0.835-0.79|+|0.76-0.73|=0.025+0.025+0.045+0.03=0.125. The value overlap is 1 - (0.125 / 4) = 0.96875, indicating that the two charging piles have a 96.875% similarity in value distribution.

[0062] The value overlap between charging piles is compared with a preset weight threshold to calculate the value matching degree. The preset weight threshold is a minimum value overlap standard pre-set by the system based on user needs and service quality requirements, typically set between 0.8 and 0.95. The value matching degree is calculated by dividing the value overlap by the preset weight threshold. If the preset weight threshold is 0.9, then the value matching degree of the two charging piles is 0.96875 / 0.9 = 1.076, which is greater than 1, indicating that the preset weight threshold requirement is met. Charging piles that meet the preset weight threshold requirement are selected as target charging piles based on the value matching degree, i.e., charging piles with a value matching degree greater than or equal to 1. In this way, a subset of target charging piles whose value distribution characteristics meet user needs are selected from all candidate charging piles.

[0063] A target charging pile value mapping map is generated, which visualizes the value distribution curves of target charging piles in the same coordinate system, facilitating the analysis of value overlap relationships between charging piles. The value overlap area between charging piles is calculated using the target charging pile value mapping map. The value overlap area refers to the area jointly covered by the value distribution curves of two charging piles in the coordinate system, used to quantify the feasibility of collaborative charging between charging piles. The calculation method involves integrating the two value distribution curves on the time axis to obtain their intersection area. The value overlap area is then normalized to obtain a connectivity value. The normalization method is to divide the value overlap area by the arithmetic mean of the areas under the respective value distribution curves of the two charging piles.

[0064] A connection matrix for the target charging stations is constructed based on the connectivity values. This connection matrix is ​​a symmetric matrix where rows and columns represent target charging stations, and element values ​​are the connectivity values ​​between corresponding charging stations. If the target charging station set contains four charging stations, numbered 1 to 4, with connectivity values ​​as follows: 0.92 between charging stations 1 and 2, 0.85 between 1 and 3, 0.78 between 1 and 4, 0.88 between 2 and 3, 0.81 between 2 and 4, and 0.90 between 3 and 4, then the diagonal elements of the connection matrix are 1 (representing the connectivity between a charging station and itself), and the off-diagonal elements are the corresponding connectivity values.

[0065] Charging collaboration groups are formed by extracting the maximum connected component from the connection matrix. A connected component refers to the set of interconnected charging piles in the connection matrix, and the maximum connected component is the connected component with the most elements. The extraction method uses a depth-first search or breadth-first search algorithm from graph theory. Charging pile pairs with a connectivity degree greater than a set threshold (e.g., 0.8) are considered connected nodes, constructing an undirected graph, and then searching for the maximum connected subgraph within the graph. For the above connection matrix, if the connection threshold is set to 0.8, charging piles 1, 2, 3, and 4 can form one charging collaboration group; if the connection threshold is increased to 0.85, two charging collaboration groups are formed: charging piles 1, 2, and 3 form one group, and charging piles 3 and 4 form another group. A suitable connection threshold is selected according to actual needs to determine the final charging collaboration group structure.

[0066] In this embodiment, the value field-based charging collaboration group construction method optimizes the allocation of charging pile resources, significantly improving the accuracy and rationality of electric vehicle charging route planning. The time mapping technology of value weight sequences enables the system to accurately capture the dynamic changes in the service value of charging piles, providing users with time-sensitive charging decision support. The introduction of value overlap and value matching degree provides objective quantitative standards for charging pile selection, reducing the uncertainty caused by subjective judgment. The charging pile value mapping diagram intuitively displays the collaborative relationships between different charging piles, facilitating user understanding and selection of the optimal charging strategy. Connectivity analysis and maximum connected component extraction techniques realize the structured expression of the charging pile network, providing a theoretical basis for multi-point charging route planning.

[0067] In one optional implementation, the charging sub-tasks are allocated to the charging coordination group according to the charging power and value weight of the target charging piles. The charging route plan is generated based on the allocation results of the target charging piles and charging sub-tasks within the charging coordination group, including: The charging power of the target charging pile is mapped according to the service period to obtain the charging power distribution sequence, and the value weight is mapped according to the service period to obtain the weight distribution sequence. Calculate the time-period supply capacity for the charging power distribution sequence, calculate the time-period service value for the weight distribution sequence, combine the time-period supply capacity and time-period service value to form the target charging pile service characteristics, and construct task characteristics by distributing the demand for charging sub-tasks according to the time-period distribution. The matching degree is calculated based on service characteristics and task characteristics, and a task allocation matrix is ​​constructed. The task allocation matrix is ​​divided into sub-blocks according to the charging coordination group. The similarity between the service features and task features within the sub-block is calculated to obtain the matching coefficient. The charging sub-tasks are sorted according to the matching coefficient to generate a task sequence. The charging sub-tasks in the task sequence are allocated to the charging coordination group according to the matching coefficient to form an allocation scheme. The task load value of the charging coordination group is calculated based on the charging coordination group allocation scheme. A charging service network is constructed according to the charging coordination group allocation scheme and the task load value of the charging coordination group. The path combination between the charging sub-task and the target charging pile is extracted in the charging service network. The distance cost and time cost of the path combination are calculated to obtain the path evaluation value. The path combination with the best path evaluation value is selected to generate a charging route scheme.

[0068] When mapping the charging power of a target charging pile to a charging power distribution sequence according to service periods, it is necessary to obtain the actual available charging power data of each target charging pile in different service periods. Charging power refers to the electrical energy that a charging pile can provide per unit time, usually measured in kilowatts (kW). A service period refers to the time interval during which a charging pile can provide services, which can be divided into hours. The charging power distribution sequence represents the changes in the charging capacity of a charging pile in different time periods. If a target charging pile has hourly charging power of 90 kW, 90 kW, 120 kW, 120 kW, 150 kW, and 150 kW respectively during a continuous 6-hour service period, then its charging power distribution sequence is [90, 90, 120, 120, 150, 150]. Reasons for changes in charging power over time include grid load adjustments, switching of charging pile operating modes, or dynamic power allocation strategies.

[0069] The value weights are mapped according to service periods to obtain a weight distribution sequence. Value weight is a comprehensive indicator measuring the service value of a charging pile, calculated by weighting factors such as charging efficiency, location convenience, and service quality, typically ranging from 0 to 1. The weight distribution sequence describes the dynamic changes in the value of a charging pile over time. The value weight of the aforementioned target charging pile during a 6-hour service period might be [0.75, 0.78, 0.82, 0.85, 0.83, 0.80], indicating a trend of its service value first increasing and then slightly decreasing.

[0070] The charging power distribution sequence is used to calculate the time-period supply capacity, i.e., the maximum amount of electricity that a charging pile can provide in each time period. The calculation method is to multiply the charging power by the time period length, with the unit being kilowatt-hours (kWh). If the time period length is 1 hour, the supply capacity of the target charging pile in the six time periods is 90, 90, 120, 120, 150, and 150 kWh, respectively. The weighted distribution sequence is used to calculate the time-period service value, i.e., considering value weights, to evaluate the service quality level of the charging pile in each time period. The calculation method is to multiply the supply capacity by the value weight of the corresponding time period. The service values ​​of the target charging pile in the six time periods are 90 × 0.75 = 67.5, 90 × 0.78 = 70.2, 120 × 0.82 = 98.4, 120 × 0.85 = 102, 150 × 0.83 = 124.5, and 150 × 0.80 = 120, respectively.

[0071] The service characteristics of a target charging pile are formed by combining time-period supply capacity and time-period service value. Service characteristics are multi-dimensional data structures describing the service capacity and value of a charging pile, comprising a time-period supply capacity sequence and a time-period service value sequence. The service characteristics of the aforementioned target charging pile can be represented as a combination of these two sequences: a supply capacity sequence [90, 90, 120, 120, 150, 150] and a service value sequence [67.5, 70.2, 98.4, 102, 124.5, 120]. The demand for charging sub-tasks is then distributed according to time periods to construct task characteristics. A charging sub-task refers to a charging demand that needs to be completed within a specific time frame. Task characteristics include two key attributes: the required electricity volume and the expected completion time. If a charging sub-task requires a total electricity volume of 250 kWh, is expected to be completed within 6 hours, and prefers charging in the later part of the time period, its time period distribution might be [20, 30, 40, 50, 60, 50] kWh.

[0072] The matching degree is calculated based on service characteristics and task characteristics, which assesses the suitability of charging pile service capabilities to charging task demands. The matching degree is calculated by weighting the ratio of supply capacity to demand in each time period, with the weight being the proportion of service value in the corresponding time period to the total service value. If the supply-demand ratios of the target charging piles and charging sub-tasks in the six time periods are 90 / 20=4.5, 90 / 30=3, 120 / 40=3, 120 / 50=2.4, 150 / 60=2.5, and 150 / 50=3 respectively, and the service value proportions in each time period are 67.5 / 582.6=0.116, 70.2 / 582.6=0.12, 98.4 / 582.6=0.169, 102 / 582.6=0.175, 124.5 / 582.6=0.214, and 120 / 582.6=0.206, then the matching degree is 4.5×0.116+3×0.12+3×0.169+2.4×0.175+2.5×0.214+3×0.206=2.96. The higher the matching degree value, the better the charging pile can meet the charging needs.

[0073] Construct a task allocation matrix, which establishes the matching relationship matrix between charging sub-tasks and target charging stations. The rows of the task allocation matrix represent different charging sub-tasks, the columns represent different target charging stations, and the matrix element values ​​are the corresponding matching degrees. If there are 3 charging sub-tasks and 4 target charging stations, the calculated matching degree values ​​are as follows: Task 1: 2.96 with charging station A, 2.58 with B, 3.12 with C, and 2.85 with D; Task 2: 2.73 with A, 3.05 with B, 2.64 with C, and 2.91 with D; Task 3: 2.82 with A, 2.76 with B, 2.95 with C, and 3.18 with D. Therefore, the task allocation matrix is ​​a 3x4 matrix, with each element representing the aforementioned matching degree.

[0074] The task allocation matrix is ​​divided into sub-blocks according to charging coordination groups. A charging coordination group refers to a set of charging piles with a coordination relationship, which can be determined by factors such as the proximity of the charging piles and the overlap of their service times. If the four target charging piles are divided into two charging coordination groups: charging piles A and B in one group, and C and D in the other group, then the task allocation matrix can be divided into two sub-blocks: one 3x2 sub-block, which includes the matching degree of tasks 1, 2, and 3 with charging piles A and B; and another 3x2 sub-block, which includes the matching degree of tasks 1, 2, and 3 with charging piles C and D. The similarity between service features and task features within each sub-block is calculated to obtain the matching coefficient. The similarity calculation method is the cosine similarity between the service feature vector and the task feature vector, with a value ranging from 0 to 1, where a larger value indicates a higher matching degree.

[0075] The charging sub-tasks are sorted according to their matching coefficients to generate a task sequence. The sorting rule is to arrange them from highest to lowest matching coefficient. If the matching coefficients are the same, they are sorted according to task priority. For the matching degree data above, if the calculated matching coefficients are: Task 1 has a matching coefficient of 0.85 with Collaboration Group 1 (A and B) and 0.92 with Collaboration Group 2 (C and D); Task 2 has a matching coefficient of 0.94 with Collaboration Group 1 and 0.83 with Collaboration Group 2; Task 3 has a matching coefficient of 0.87 with Collaboration Group 1 and 0.9 with Collaboration Group 2, then the task sequence may be: Task 2 assigned to Collaboration Group 1, Task 1 assigned to Collaboration Group 2, and Task 3 assigned to Collaboration Group 2. The charging sub-tasks in the task sequence are then assigned to charging collaboration groups according to their matching coefficients to form an allocation scheme. The allocation principle is to prioritize tasks with high matching coefficients, while also considering the capacity constraints and time window limitations of the charging piles.

[0076] The task load value of each charging coordination group is calculated based on the charging coordination group allocation scheme, which evaluates the total charging task undertaken by each charging coordination group. The calculation method is to divide the sum of the charging volume allocated to all charging piles within the coordination group by the total supply capacity of the charging piles within the group. If the total charging task allocated to coordination group 1 is 300 kWh and its total supply capacity is 720 kWh, then the task load value is 300 / 720 = 0.417; if the total charging task allocated to coordination group 2 is 500 kWh and its total supply capacity is 900 kWh, then the task load value is 500 / 900 = 0.556. A charging service network is constructed based on the charging coordination group allocation scheme and the task load value of each group, establishing a relationship network between charging sub-tasks, target charging piles, and vehicle travel paths. Path combinations between charging sub-tasks and target charging piles are extracted from the charging service network, and the distance cost and time cost of the path combinations are calculated to obtain a path evaluation value. The path evaluation value comprehensively considers multiple factors such as distance, time, and charging efficiency to evaluate the merits of different path schemes. The optimal path combination with the best path evaluation value is selected to generate a charging route plan, which determines the final charging route planning result for the vehicle.

[0077] In this embodiment, by mapping charging power and value weights to the time dimension, the system can accurately grasp the dynamic changes in charging pile service capabilities, providing a time-sensitive decision-making basis for task allocation. The matching analysis technology between service features and task features achieves precise matching between charging demand and charging resources, improving the rationality of resource allocation. The sub-block partitioning method based on collaborative groups enables the system to optimize task allocation strategies at a finer granularity, improving charging efficiency and user experience. The matching coefficient sorting and task sequence generation mechanism ensures the priority allocation of critical tasks, reducing charging waiting time.

[0078] A second aspect of this invention provides a route planning system for new energy vehicle charging piles, the system comprising: The first unit is used to obtain vehicle status information and charging pile information; The second unit is used to extract the time distribution characteristics of vehicle arrival at the charging pile and the charging duration distribution characteristics from historical charging data for each charging pile in the charging pile information, and to construct a load status map. The third unit is used to calculate the path energy consumption value and waiting time prediction value of the vehicle to each charging pile based on the load status map and vehicle status information, and to assign value weight to each charging pile at different times to construct a value field. The fourth unit is used to calculate the charging location boundary and charging time boundary in reverse based on the charging pile information and the path energy consumption value, with the vehicle status information as the benchmark. It then filters candidate charging piles according to the charging location boundary, calculates the total charging amount required by the vehicle, and decomposes the total charging amount into multiple charging sub-tasks according to the charging power of the candidate charging piles. The fifth unit is used to select target charging piles from candidate charging piles based on the value field, and construct charging collaboration groups that meet the preset weight threshold within the charging time boundary. The sixth unit is used to allocate charging sub-tasks to the charging coordination group according to the charging power and value weight of the target charging piles, and generate charging route plans based on the allocation results of the target charging piles and charging sub-tasks in the charging coordination group.

[0079] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0080] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0081] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0082] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A route planning method for a new energy vehicle charging pile, characterized in that, The method comprises the following steps: acquiring vehicle state information and charging pile information; extracting time distribution characteristics and charging duration distribution characteristics of vehicles arriving at the charging piles from historical charging data for each charging pile in the charging pile information, and constructing a load state atlas; calculating path energy consumption values and waiting time prediction values of vehicles arriving at each charging pile for charging according to the load state atlas and the vehicle state information, assigning value weights to each charging pile at different time points, and constructing a value field; based on the charging pile information and the path energy consumption values, reversely calculating charging position boundaries and charging time boundaries based on the vehicle state information, screening candidate charging piles according to the charging position boundaries, calculating the total charging capacity required by the vehicle, and decomposing the total charging capacity into multiple charging subtasks according to the charging power of the candidate charging piles; selecting a target charging pile from the candidate charging piles according to the value field, and constructing a charging cooperative group when the value weight of the target charging pile within the charging time boundaries meets a preset weight threshold; allocating the charging subtasks to the charging cooperative group according to the charging power and the value weight of the target charging pile, and generating a charging route scheme according to the allocation results of the target charging pile and the charging subtasks of the charging cooperative group.

2. The method of claim 1, wherein, The method of constructing a load state atlas by extracting time distribution characteristics and charging duration distribution characteristics of vehicles arriving at the charging piles from historical charging data for each charging pile in the charging pile information comprises the following steps: dividing the historical charging data of the charging pile into time slices according to time, calculating the change rate and change acceleration of the number of vehicles arriving at the charging pile in adjacent time slices, and constructing an arrival time sequence fluctuation curve according to the change rate and change acceleration; identifying a fluctuation mutation interval from the arrival time sequence fluctuation curve, aggregating the time slices into multiple arrival time periods, and performing smoothing processing on the number of vehicles arriving at the charging pile in each arrival time period to generate an arrival smoothing sequence and an arrival time period division point; based on the arrival smoothing sequence and the arrival time period division point, calculating the distribution density and dispersion degree of the charging duration of the vehicles in the arrival time period, constructing a charging duration characteristic curve, and combining the arrival smoothing sequence and the charging duration characteristic curve to generate a feature descriptor; calculating the information redundancy of the time period feature descriptor, determining a time period weight factor according to the information redundancy, and performing fusion processing on the feature descriptor by using the time period weight factor to construct a load state atlas.

3. The method of claim 1, wherein, The method of calculating path energy consumption values and waiting time prediction values of vehicles arriving at each charging pile for charging according to the load state atlas and the vehicle state information, and assigning value weights to each charging pile at different time points to construct a value field comprises the following steps: converting the load state atlas into a space-time coordinate sequence, acquiring real-time road speed and road congestion status, calculating road travel time and congestion index; determining a driving path of the vehicle from the position to the charging pile according to the travel time and the congestion index, combining the residual power and the target charging power in the vehicle state information, calculating the path energy consumption value and the arrival prediction time of the driving path, and generating an energy consumption distribution sequence and an arrival time sequence; based on the energy consumption distribution sequence and the arrival time sequence, extracting charging load values at corresponding time points in the load state atlas, and calculating dynamic ratios of the charging load values to the charging pile power supply power; The dynamic ratio and the charging demand in the vehicle state information are used to construct a game payoff function, and the charging queue number is obtained by solving the equilibrium solution of the game payoff function; the waiting time prediction value is calculated according to the queue number and the single charging duration; The energy consumption distribution sequence, the arrival time sequence and the waiting time prediction value are used to construct a joint probability distribution, the mutual information value between variables in the joint probability distribution is extracted, a mutual information matrix is generated, the charging pile value weight is calculated according to the mutual information matrix, and the value field is constructed by mapping the value weight to the time dimension.

4. The method of claim 3, wherein, The dynamic ratio and the charging demand in the vehicle state information are used to construct a game payoff function, and the charging queue number is obtained by solving the equilibrium solution of the game payoff function; the waiting time prediction value is calculated according to the queue number and the single charging duration including: The time segment revenue value is obtained by dividing the dynamic ratio according to the time segment, the charging demand quantity is extracted from the vehicle state information to obtain the time segment demand value, the time segment revenue value and the time segment demand value are combined in the time dimension to form a revenue mapping matrix, and a game payoff distribution is generated from the revenue mapping matrix; Based on the game payoff distribution, a game strategy space is constructed, the strategy selection probability is calculated in the game strategy space, a strategy probability matrix is generated, an equilibrium strategy vector is extracted from the strategy probability matrix, and an optimal response sequence is constructed; The optimal response sequence is expanded into a response probability matrix in the time dimension, the probability distribution value in the response probability matrix is calculated, the game equilibrium point is determined based on the probability distribution value, and the game equilibrium point is mapped to the time segment to obtain the vehicle number distribution in the time period. The waiting time prediction value is obtained by multiplying the vehicle number distribution and the single charging duration.

5. The method of claim 1, wherein, Based on the charging pile information and the path energy consumption value, the charging position boundary and the charging time boundary are reversely calculated based on the vehicle state information, the candidate charging piles are screened according to the charging position boundary, the total charging quantity required by the vehicle is calculated, and the total charging quantity is decomposed into multiple charging tasks according to the charging power of the candidate charging piles including: The remaining electric quantity and the target electric quantity in the vehicle state information are used to calculate the electric quantity difference value, the path energy consumption value and the road congestion degree are converted into the energy consumption compensation coefficient, the electric quantity difference value is adjusted to obtain the actual electric quantity demand, the upper limit of the driving distance is calculated based on the actual electric quantity demand, and the charging position boundary is generated; The expected arrival time and the current time in the vehicle state information are used to calculate the time difference value, the path energy consumption value and the charging pile queue number are converted into the time compensation coefficient, the actual time window is obtained by adjusting the time difference value according to the time compensation coefficient, the upper limit of the driving time is calculated based on the actual time window, and the charging time boundary is generated; According to the charging position boundary and the charging time boundary, the geographical position and the service period are extracted from the charging pile information, the density value of the geographical position in the space distribution and the occupancy rate of the service period in the time dimension are calculated, and the candidate charging pile list meeting the charging position boundary and the charging time boundary is screened based on the density value and the occupancy rate; The target electric quantity and the remaining electric quantity in the vehicle state information are used to calculate the total charging quantity, the charging power in the candidate charging pile list and the total charging quantity are combined to generate a task allocation matrix, the charging quantity interval is divided according to the task allocation matrix, and the charging task list is output.

6. The method of claim 1, wherein, The method comprises the following steps: obtaining a value weight sequence of the candidate charging piles within the charging time boundary from the value field, mapping the value weight sequence to the time dimension to form a value distribution curve, performing interval division on the value distribution curve to generate a time period value matrix, and calculating a value overlap amount between the charging piles according to the time period value matrix; comparing the value overlap amount between the charging piles with the preset weight threshold to calculate a value matching degree, and selecting the charging piles meeting the preset weight threshold requirement as target charging piles according to the value matching degree to generate a target charging pile value mapping diagram; calculating a value overlap area between the charging piles through the target charging pile value mapping diagram, normalizing the value overlap area to obtain a connection degree value, constructing a target charging pile connection matrix according to the connection degree value, and extracting a maximum connected component based on the connection matrix to form a charging collaborative group.

7. The method of claim 1, wherein, The method comprises the following steps: mapping the charging power of the target charging pile according to the service time period to obtain a charging power distribution sequence, and mapping the value weight according to the service time period to obtain a weight distribution sequence; calculating a time period supply capacity for the charging power distribution sequence, calculating a time period service value for the weight distribution sequence, combining the time period supply capacity and the time period service value to form a target charging pile service feature, and constructing a task feature according to the demand quantity of the charging subtask and the time period distribution; calculating a matching degree based on the service feature and the task feature, and constructing a task allocation matrix; dividing the task allocation matrix into subblocks according to the charging collaborative group, calculating the similarity of the service feature and the task feature in the subblock to obtain a matching coefficient, sorting the charging subtasks according to the matching coefficient to generate a task sequence, and distributing the charging subtasks in the task sequence to the charging collaborative group according to the matching coefficient to form a distribution scheme; calculating a charging collaborative group task load value based on the charging collaborative group distribution scheme, constructing a charging service network according to the charging collaborative group distribution scheme and the charging collaborative group task load value, extracting a path combination between the charging subtask and the target charging pile in the charging service network, calculating a distance cost and a time cost of the path combination to obtain a path evaluation value, and selecting a path combination with the optimal path evaluation value to generate a charging route scheme.

8. A route planning system for new energy vehicle charging piles, for implementing the method of any one of the preceding claims 1-7, characterized in that, The method comprises the following steps: a first unit is configured to obtain vehicle state information and charging pile information; a second unit is configured to extract, from historical charging data, a time distribution feature of a vehicle arriving at a charging pile and a charging duration distribution feature for each charging pile in the charging pile information, and construct a load state atlas; a third unit is configured to calculate, according to the load state atlas and the vehicle state information, a path energy consumption value and a waiting time prediction value of the vehicle arriving at each charging pile for charging, assign a value weight to each charging pile at different time points, and construct a value field. The fourth unit is configured to calculate, based on the charging pile information and the path energy consumption value, a charging position boundary and a charging time boundary reversely with the vehicle state information as a reference, filter candidate charging piles according to the charging position boundary, calculate total charging quantity required by the vehicle, and decompose the total charging quantity into a plurality of charging sub-tasks according to charging powers of the candidate charging piles; The fifth unit is configured to select, from the candidate charging piles, a target charging pile whose value weight satisfies a preset weight threshold within the charging time boundary according to the value field, and construct a charging cooperative group. The sixth unit is configured to allocate the charging sub-tasks to the charging cooperative group according to the charging power and the value weight of the target charging pile, and generate a charging route scheme according to an allocation result of the target charging pile and the charging sub-tasks of the charging cooperative group.

9. An electronic device, comprising: comprise: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon computer program instructions, wherein, The computer program instructions, when executed by the processor, implement the method of any one of claims 1 to 7.

Images