New energy vehicle charging path planning system based on multidimensional data mining technology

Abstract

The present invention discloses a new energy vehicle charging route planning system based on multidimensional data mining technology, the system comprising a user data collection module, a multidimensional data mining module, and a route visualization module. The present invention acquires pre-stored route planning elements by collecting user driving elements and route recommendation tendency information. Upon receiving the user driving elements and route recommendation tendency information transmitted by the user data collection module, the present invention performs data mining and similarity matching on the user driving elements and route planning elements based on the route recommendation tendency information, obtains a set of candidate charging piles, obtains navigation information of the route planning for each charging pile in the set of candidate charging piles, and visualizes the navigation information. Thus, while considering conventional factors in charging pile route planning, traffic conditions and user living habits are added, and route recommendation tendency selection is provided, thereby achieving new energy vehicle charging route planning with higher accuracy, specificity, and real-time performance.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a new energy vehicle charging path planning system based on multidimensional data mining technology. Background Art

[0002] At present, electric vehicle users usually face the problem of difficulty in charging. On the one hand, electric vehicle charging piles have not yet been widely installed. On the other hand, electric vehicles take a long time to charge, and cannot be as fast as traditional fuel vehicles that can be refueled at a gas station and driven away after refueling. Therefore, it is difficult to reasonably allocate and wait for cars that are charging and cars to be charged.

[0003] Secondly, charging stations are typically concentrated in commercial areas near city centers. However, these areas often have high occupancy rates and are often congested. Users traveling to these stations may face queues and traffic jams. Besides commercial areas, some charging stations are also located in suburban areas, but these locations present significant distances for drivers. Existing technologies primarily consider charging station fees, distance, and station occupancy, but lack consideration for real-time conditions such as traffic congestion.

[0004] In addition, the user's living habits should also be considered in the planning of charging pile routes, such as: the user's commuting time, address, office location, etc. These factors have personalized attributes and can provide users with more exclusive planning.

[0005] Therefore, how to improve the accuracy, specificity and real-time performance of new energy vehicle charging route planning is a technical problem that needs to be solved urgently. Summary of the Invention

[0006] The main purpose of the present invention is to provide a new energy vehicle charging path planning system based on multidimensional data mining technology, aiming to solve the current technical problems of poor accuracy, specificity and real-time performance of new energy vehicle charging path planning.

[0007] To achieve the above objectives, the present invention provides a new energy vehicle charging path planning system based on multidimensional data mining technology, the system comprising:

[0008] User data collection module, which collects user travel factors and route recommendation tendency information;

[0009] The multidimensional data mining module obtains pre-stored route planning elements. Upon receiving the user driving elements and route recommendation tendency information transmitted by the user data acquisition module, the module performs data mining and similarity matching on the user driving elements and route planning elements based on the route recommendation tendency information to obtain a set of candidate charging piles.

[0010] The path visualization module obtains navigation information of the path planning of each charging pile in the candidate charging pile set and visualizes the navigation information.

[0011] Optionally, based on the route recommendation tendency information, data mining and similarity matching are performed on the user's driving elements and route planning elements to obtain a set of candidate charging piles, specifically including:

[0012] Based on the route recommendation tendency information, feature selection is performed on the user's driving factors and route planning factors to construct a multi-dimensional feature matrix;

[0013] The initial route recommendation model is trained using the multi-dimensional feature matrix to obtain a target route recommendation model integrating the user's personal features;

[0014] Similarity matching is performed based on the target path recommendation model to determine a set of candidate charging piles.

[0015] Optionally, the user driving elements include the car's starting position, the car's current cruising range and the user's living habit data; the path planning elements include charging pile charging information, charging pile queuing information and charging pile distance information.

[0016] Optionally, after collecting the user's living habit data, the method further includes: making a predictive forecast of the user's behavior after charging based on the user's living habit data.

[0017] Optionally, the route recommendation tendency information includes a time cost priority tendency and a charging cost priority tendency; wherein:

[0018] Time cost priority, specifically: prioritize the time T from the user's departure point to the charging station c Add the time T from the charging station to the destination of the user's predicted behavior p The path with the least sum;

[0019] Charging cost priority, specifically: giving priority to user charging cost C c Add the travel cost C r The path with the smallest sum.

[0020] Optionally, based on the route recommendation tendency information, feature selection is performed on the user's driving elements and route planning elements to construct a multi-dimensional feature matrix, specifically including:

[0021] When the path recommendation tendency information is time cost priority, the recursive feature elimination method is used to select features, and a multi-dimensional feature matrix is ​​constructed based on the selected features;

[0022] When the route recommendation tendency information is charging cost priority, the principal component analysis method is used to select features, and a multidimensional feature matrix is ​​constructed based on the selected features.

[0023] Optionally, the initial path recommendation model adopts the XGBoost recommendation model.

[0024] Optionally, when the path recommendation tendency information is a time cost priority tendency, the expression of the objective function of the initial path recommendation model is specifically:

[0025] obj T =Loss+α(f T )

[0026]

[0027] Among them, α is the regularization term that suppresses the complexity of the model, f T To fit the function, which has no polynomial solution, use Taylor's second-order expansion to approximate it:

[0028]

[0029] N is the number of samples, L is the Huber Loss function, y i is the actual value, is the predicted value; the expression of Huber Loss function is:

[0030]

[0031] Among them, δ is a hyperparameter of HuberLoss.

[0032] Optionally, when the route recommendation tendency information is a charging cost priority tendency, the expression of the objective function of the initial route recommendation model is specifically:

[0033] obj C =Loss+Loss″+β(f C )

[0034]

[0035] Among them, β is the regular term that suppresses the complexity of the model, f C is the fitting function, which has a finite number of polynomial solutions, N is the number of samples, L is the Huber Loss function, y i is the actual value, is the predicted value; the expression of Huber Loss function is:

[0036]

[0037] Among them, δ is a hyperparameter of Huber Loss.

[0038] Optionally, performing similarity matching based on the target path recommendation model to determine a set of candidate charging piles specifically includes:

[0039] generating a scoring matrix based on the target path recommendation model;

[0040] The scoring matrix is ​​used to perform similarity matching on each charging pile to obtain a set of candidate charging piles; wherein the set of candidate charging piles includes a preset number of charging piles ranked by score.

[0041] Optionally, the expression of the scoring matrix is ​​specifically:

[0042] K=(λ1,λ2,λ3,θ)

[0043] Among them, λ1 is the action category of the current user predicted by the target path recommendation model, λ2 and λ3 are the time priority score corresponding to the time cost priority tendency and the charging cost priority score corresponding to the charging cost limited tendency of the target path recommendation model, respectively, and θ is the current cruising range.

[0044] Optionally, obtaining navigation information of the path planning for each charging pile in the candidate charging pile set and visualizing the navigation information specifically includes:

[0045] Sending the vehicle's starting location and the candidate charging pile set to a third-party navigation software, so that the third-party navigation software obtains historical traffic data information and generates navigation information for a path planning for each charging pile in the candidate charging pile set based on the historical traffic data information, the vehicle's starting location, and the candidate charging pile set;

[0046] Receive navigation information of the path planning of each charging pile in the candidate charging pile set transmitted by the third navigation software, and visualize the navigation information.

[0047] Optionally, obtaining navigation information of a path plan for each charging pile in the candidate charging pile set and visualizing the navigation information further includes:

[0048] Determine whether the mileage corresponding to the navigation information of the path planning for each charging pile is greater than the current cruising range of the vehicle;

[0049] If so, the charging pile is removed from the candidate charging pile set.

[0050] Optionally, obtaining navigation information of a path plan for each charging pile in the candidate charging pile set and visualizing the navigation information further includes:

[0051] Obtaining current traffic data, and the path visualization module transmits the current traffic data to the multidimensional data mining module, so that the multidimensional data mining module updates the candidate charging pile set;

[0052] The path visualization module sends the updated candidate charging pile set to the third-party navigation software to update the navigation information of the path planning of each charging pile in the candidate charging pile set.

[0053] In addition, in order to achieve the above-mentioned purpose, the present invention also provides a new energy vehicle charging path planning method based on multidimensional data mining technology, the method comprising:

[0054] Collect user travel factors and route recommendation tendency information;

[0055] Obtaining pre-stored route planning elements, upon receiving user travel elements and route recommendation tendency information transmitted by the user data acquisition module, performing data mining and similarity matching on the user travel elements and route planning elements based on the route recommendation tendency information to obtain a set of candidate charging piles;

[0056] Obtain navigation information for the path planning of each charging pile in the candidate charging pile set, and visualize the navigation information.

[0057] In addition, in order to achieve the above-mentioned purpose, the present invention also provides a new energy vehicle charging path planning device based on multidimensional data mining technology, and the device includes: a memory, a processor, and a new energy vehicle charging path planning program based on multidimensional data mining technology stored on the memory and runnable on the processor. When the new energy vehicle charging path planning program based on multidimensional data mining technology is executed by the processor, the steps of the above-mentioned new energy vehicle charging path planning method based on multidimensional data mining technology are implemented.

[0058] In addition, in order to achieve the above-mentioned purpose, the present invention also provides a storage medium, on which a new energy vehicle charging path planning program based on multidimensional data mining technology is stored. When the new energy vehicle charging path planning program based on multidimensional data mining technology is executed by a processor, the steps of the above-mentioned new energy vehicle charging path planning method based on multidimensional data mining technology are implemented.

[0059] An embodiment of the present invention proposes a new energy vehicle charging route planning system based on multidimensional data mining technology, which includes a user data collection module, a multidimensional data mining module, and a route visualization module. The present invention obtains pre-stored route planning elements by collecting user driving elements and route recommendation tendency information. When receiving the user driving elements and route recommendation tendency information transmitted by the user data collection module, the present invention performs data mining and similarity matching on the user driving elements and route planning elements based on the route recommendation tendency information, obtains a set of candidate charging piles, obtains navigation information of the route planning of each charging pile in the candidate charging pile set, and visualizes the navigation information. Thus, under the premise of considering the conventional factors of charging pile route planning, traffic conditions and user living habits are added, and route recommendation tendency selection is provided, thereby achieving new energy vehicle charging route planning with higher accuracy, specificity, and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a schematic diagram of an embodiment of a new energy vehicle charging path planning system based on multidimensional data mining technology in an embodiment of the present invention;

[0061] Figure 2 Schematic diagram of the execution flow of the multidimensional data mining module of the present invention;

[0062] Figure 3 This is a schematic diagram of an embodiment of a new energy vehicle charging path planning method based on multidimensional data mining technology in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0064] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] The embodiment of the present invention provides a new energy vehicle charging path planning system based on multidimensional data mining technology, referring to Figure 1 , Figure 1 This is a schematic diagram of an embodiment of a new energy vehicle charging path planning system based on multidimensional data mining technology of the present invention.

[0066] In this embodiment, the new energy vehicle charging route planning system based on multidimensional data mining technology includes: a user data collection module, a multidimensional data mining module, and a route visualization module. The user data collection module is located on the user side and collects multidimensional data for charging route planning. The multidimensional data mining module is located on the cloud server side and performs mining and similarity matching on the received multidimensional data. The route visualization module is located on the user side and visualizes the route planning and navigation information for each charging pile in the candidate charging pile set, providing the user with the optimal route plan.

[0067] Specifically, in this embodiment, the user data collection module is used to collect user driving elements and path recommendation tendency information; the multidimensional data mining module is used to obtain pre-stored path planning elements, and when receiving the user driving elements and path recommendation tendency information transmitted by the user data collection module, the multidimensional data mining module performs data mining and similarity matching on the user driving elements and path planning elements based on the path recommendation tendency information to obtain a set of candidate charging piles; the path visualization module is used to obtain navigation information of the path planning of each charging pile in the candidate charging pile set, and visualize the navigation information.

[0068] It should be noted that user travel factors include the car's starting location, the car's current range, and the user's lifestyle data. Among them, the user's lifestyle data includes work hours, off-get off work hours, office location, and historical movement information (such as going to work, going home, shopping, dining, etc.).

[0069] In a preferred embodiment, after collecting the user's living habit data, the method further includes: making a predictive forecast of the user's behavior after charging based on the user's living habit data.

[0070] Therefore, this embodiment, by incorporating elements of user living habits into route planning, can generate recommendations based on personal specificity for users, and can provide predictive action predictions based on users' daily habits, that is, predict where the user will go next after charging, and use this prediction result to assist in planning the charging route.

[0071] In order to explain the present application more clearly, the following specifically describes the implementation of charging path planning based on predictive prediction.

[0072] For example, if we know based on user lifestyle data that the user drives to a gym at a certain location around 7 PM every Thursday, we can predict that the user's predicted action at 7 PM next Thursday will be to go to that location. Therefore, in this scenario, when planning charging routes based on this predictive prediction, we prioritize charging stations near that location.

[0073] Therefore, by predicting the user's predictive actions to assist in the planning of charging routes, more personalized charging route recommendations are generated.

[0074] It should be noted that the route planning elements include charging pile charging information, charging pile queuing information and charging pile distance information.

[0075] Based on this, the multidimensional data mining module can perform data mining and similarity matching based on the received user travel factors and route planning factors to obtain a set of candidate charging piles. Specifically, the user travel factors provide user-side feature data for charging route planning, while the route planning factors provide charging pile feature data for charging route planning. This allows the multidimensional data mining module to generate accurate and personalized route plans based on these user-side and charging pile feature data.

[0076] Furthermore, the path visualization module uses historical traffic data as an evaluation and real-time traffic data as it navigates the path planning for each charging station in the candidate charging station set. This approach takes a holistic approach, incorporating two new factors—traffic conditions and user lifestyle—into the five commonly used elements of charging station fee information, charging station queue information, charging station distance information, vehicle starting location, and vehicle current range—to plan the user's charging path. This allows for real-time dynamic planning of the optimal path, minimizing the impact of unexpected situations such as traffic congestion on the user.

[0077] In another embodiment, when data mining and similarity matching are performed on user driving elements and path planning elements based on path recommendation tendency information to obtain a set of candidate charging piles, feature selection is performed on user driving elements and path planning elements based on the path recommendation tendency information to construct a multidimensional feature matrix; the initial path recommendation model is trained using the multidimensional feature matrix to obtain a target path recommendation model that integrates user personal features; similarity matching is performed based on the target path recommendation model to determine the set of candidate charging piles.

[0078] Easy to understand, such as Figure 2 As shown, after uploading the data collected by the user data collection module to the multidimensional data mining module, the multidimensional data mining module will first perform data cleaning to filter out dirty data. After that, feature selection, feature processing and feature construction are performed.

[0079] It should be noted that the route recommendation tendency information includes time cost priority and charging cost priority. Among them: time cost priority, specifically: give priority to the time T from the user's departure point to the charging station c Add the time T from the charging station to the destination of the user's predicted behavior p The path with the least total cost; the priority of charging cost, specifically: giving priority to user charging cost C c Add the travel cost C r The path with the smallest sum.

[0080] Specifically, the recommendation logic of the two recommendation methods of path recommendation tendency information is as follows:

[0081] (1) Time cost priority: The charging station fee and the distance length are not considered. Only the path with the shortest time is considered. However, it should be noted that the shortest time here does not refer to the time T from the user's starting point to the charging station. c At least, it refers to the time T from the departure point to the charging station c Add the charging pile to the predictive action prediction destination time T p The total is the least. Let the sum of the two times be T = T c +T p .

[0082] (2) Charging cost priority: The reference indicator for this recommendation is the charging cost C c and the distance cost C r Minimum, recorded as total cost C = C c +C p It should be noted that the distance cost C r It is not only determined by the length of the journey, but also by the traffic conditions on the current road section, because new energy vehicles and gasoline vehicles also have the problem of idling energy consumption. The longer the journey, the worse the traffic conditions, and the r The larger the value.

[0083] In this embodiment, the user can manually select the time cost priority tendency or the charging cost priority tendency, thereby providing a feature selection and training method for the new energy vehicle charging path planning system based on multidimensional data mining technology, and providing users with more suitable path recommendations.

[0084] Specifically, when the path recommendation tendency information is time cost priority, the recursive feature elimination method is used to select features, and a multidimensional feature matrix is ​​constructed based on the selected features; when the path recommendation tendency information is charging cost priority, the principal component analysis method is used to select features, and a multidimensional feature matrix is ​​constructed based on the selected features.

[0085] (1) For time cost priority: This recommendation method uses recursive feature elimination to select features, and the constructed multidimensional feature matrix H is m×n dimensional.

[0086] (2) For charging cost priority: This recommendation method uses principal component analysis to select features, and the constructed multidimensional feature matrix H is dimension.

[0087] After obtaining the multi-dimensional feature matrix, it is necessary to select the corresponding recommendation model to adapt to the needs of different path recommendation tendencies. Among them, the initial path recommendation model uses the XGBoost recommendation model. Due to the user's choice of recommendation method, the objective function of different recommendation methods is different. The following describes the objective functions of the recommendation models of different recommendation methods:

[0088] (1) When the path recommendation tendency information is a time cost priority tendency, the expression of the objective function of the initial path recommendation model is specifically:

[0089] obj T =Loss+α(f T )

[0090]

[0091] Among them, α is the regularization term that suppresses the complexity of the model, f T To fit the function, which has no polynomial solution, use Taylor's second-order expansion to approximate it:

[0092]

[0093] N is the number of samples, L is the Huber Loss function, y i is the actual value, is the predicted value; the expression of Huber Loss function is:

[0094]

[0095] Among them, δ is a hyperparameter of HuberLoss, and δ is the connection position of the two losses of MSE and MAE.

[0096] (2) When the route recommendation tendency information is the charging cost priority tendency, the expression of the objective function of the initial route recommendation model is specifically:

[0097] obj C =Loss+Loss″+β(f C )

[0098]

[0099] Among them, β is the regular term that suppresses the complexity of the model, f C is the fitting function, which has a finite number of polynomial solutions, N is the number of samples, L is the Huber Loss function, y i is the actual value, is the predicted value; the expression of Huber Loss function is:

[0100]

[0101] Among them, δ is a hyperparameter of Huber Loss, and δ is the connection position of the two losses of MSE and MAE.

[0102] Therefore, after obtaining the multidimensional feature matrix and the initial route recommendation model, the multidimensional feature matrix can be used to train the initial route recommendation model to obtain the target route recommendation model that integrates the user's personal features.

[0103] This embodiment considers the path recommendation tendency information, adopts different multidimensional feature matrix construction and different model selection methods, and then after training the initial path recommendation model using the multidimensional feature matrix, the obtained target path recommendation model integrates the user's personal characteristics and path recommendation tendency.

[0104] After that, this embodiment performs similarity matching based on the target path recommendation model to determine a set of candidate charging piles.

[0105] Specifically, a scoring matrix is ​​generated based on the target path recommendation model; similarity matching is performed on each charging pile using the scoring matrix to obtain a set of candidate charging piles; wherein the set of candidate charging piles includes a preset number of charging piles ranked by score.

[0106] In a preferred embodiment, the expression of the scoring matrix is ​​specifically:

[0107] K=(λ1,λ2,λ3,θ)

[0108] Among them, λ1 is the action category of the current user predicted by the target path recommendation model, λ2 and λ3 are the time priority score corresponding to the time cost priority tendency and the charging cost priority score corresponding to the charging cost limited tendency of the target path recommendation model, respectively, and θ is the current cruising range.

[0109] In order to explain this application more clearly, the following specifically describes how to implement charging path planning based on the scoring matrix.

[0110] For example, based on user lifestyle data, we know that the user typically prioritizes time costs when heading to work at 9:00 AM on weekdays, and prioritizes charging costs when heading home at 5:00 PM on weekdays. We can predict that the user will prioritize time costs when going to work at 9:00 AM on weekdays, and prioritize charging costs when leaving work at 5:00 PM on weekdays. Based on this, when generating the scoring matrix, we assign a higher time-priority score to the time of going to work and a higher charging-cost-priority score to the time of leaving work.

[0111] As a result, when planning a charging route for a user's commute, priority will be given to charging stations that prioritize time, while when planning a charging route for the user's return, priority will be given to charging stations that prioritize charging costs. Furthermore, by considering route recommendation preference information, a charging route plan that integrates the user's personal characteristics and route recommendation preferences is obtained.

[0112] In a preferred embodiment, the candidate charging pile set includes the top three charging piles ranked by score.

[0113] In another embodiment, the path visualization module obtains the navigation information of the path planning of each charging pile in the candidate charging pile set and visualizes the navigation information by calling a third-party navigation software.

[0114] Specifically, the car's starting position and the set of candidate charging piles are sent to a third-party navigation software so that the third-party navigation software obtains historical traffic data information, and based on the historical traffic data information, the car's starting position and the set of candidate charging piles, generates navigation information for path planning of each charging pile in the candidate charging pile set; receives the navigation information for path planning of each charging pile in the candidate charging pile set transmitted by the third navigation software, and visualizes the navigation information.

[0115] It should be noted that when obtaining the navigation information of the path planning of each charging pile in the candidate charging pile set and visualizing the navigation information, it is also necessary to consider the distance to the current charging pile and whether the current vehicle can reach it. If the current cruising range of the car does not support the car to reach the current charging pile, the path planning of the current charging pile will not be considered.

[0116] Specifically, it is determined whether the mileage corresponding to the navigation information of the path planning of each charging pile is greater than the current cruising range of the vehicle; if so, the charging pile is eliminated from the candidate charging pile set.

[0117] It should be noted that when obtaining the navigation information of the path planning of each charging pile in the candidate charging pile set and visualizing the navigation information, it is also necessary to consider the current traffic congestion. If traffic congestion occurs at the location of the current path, the path may not belong to the path of the charging pile with the highest score in the candidate charging pile set. At this time, it is necessary to consider regenerating the candidate charging pile set.

[0118] Specifically, the current traffic data is obtained, and the path visualization module transmits the current traffic data to the multidimensional data mining module so that the multidimensional data mining module updates the candidate charging pile set; the path visualization module sends the updated candidate charging pile set to the third-party navigation software to update the navigation information of the path planning of each charging pile in the candidate charging pile set.

[0119] Therefore, this embodiment generates a more real-time route planning solution by considering both historical and real-time traffic data, providing users with a better route planning experience when faced with unexpected factors. This embodiment provides a new energy vehicle charging route planning system based on multidimensional data mining technology. By considering conventional factors in charging pile route planning, adding traffic conditions and user living habits, and providing route recommendation preferences, it achieves new energy vehicle charging route planning with higher accuracy, specificity, and real-time performance.

[0120] Reference Figure 3 , Figure 3 This is a schematic diagram of an embodiment of a new energy vehicle charging path planning method based on multidimensional data mining technology of the present invention.

[0121] like Figure 3 As shown, the new energy vehicle charging path planning method based on multidimensional data mining technology proposed in an embodiment of the present invention includes the following steps:

[0122] S100: Collecting user travel factors and route recommendation tendency information;

[0123] S200: Obtaining pre-stored route planning elements, upon receiving user travel elements and route recommendation tendency information transmitted by the user data acquisition module, performing data mining and similarity matching on the user travel elements and route planning elements based on the route recommendation tendency information to obtain a set of candidate charging piles;

[0124] S300: Obtain navigation information of the path planning of each charging pile in the candidate charging pile set, and visualize the navigation information.

[0125] Other embodiments or specific implementations of the new energy vehicle charging path planning method based on multidimensional data mining technology of the present invention can refer to the above-mentioned system embodiments and will not be repeated here.

[0126] In addition, the present invention also proposes a new energy vehicle charging path planning device based on multidimensional data mining technology, and the new energy vehicle charging path planning device based on multidimensional data mining technology includes a memory, a processor, and a new energy vehicle charging path planning program based on multidimensional data mining technology stored on the memory and run on the processor, wherein: when the new energy vehicle charging path planning program based on multidimensional data mining technology is executed by the processor, the new energy vehicle charging path planning method based on multidimensional data mining technology described in each embodiment of the present invention is implemented.

[0127] The specific implementation of the new energy vehicle charging path planning device based on multidimensional data mining technology in this application is basically the same as the above-mentioned embodiments of the new energy vehicle charging path planning system based on multidimensional data mining technology, and will not be repeated here.

[0128] In addition, the present invention further provides a readable storage medium, comprising a computer-readable storage medium, on which is stored a new energy vehicle charging path planning program based on multidimensional data mining technology. The readable storage medium may be at least one of a ROM (Read-Only Memory) / RAM (Random Access Memory), a magnetic disk, and an optical disk, and the readable storage medium includes a plurality of instructions for enabling a new energy vehicle charging path planning device based on multidimensional data mining technology having a processor to execute the new energy vehicle charging path planning method based on multidimensional data mining technology described in various embodiments of the present invention.

[0129] The specific implementation of the new energy vehicle charging path planning program based on multidimensional data mining technology in the readable storage medium of this application is basically the same as the above-mentioned embodiments of the new energy vehicle charging path planning system based on multidimensional data mining technology, and will not be repeated here.

[0130] It should be understood that, in the description of this specification, reference to terms such as "one embodiment," "another embodiment," "other embodiments," or "first to Nth embodiments" means that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples.

[0131] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0132] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0133] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A new energy vehicle charging path planning system based on multidimensional data mining technology, characterized in that the system includes: User data collection module, which collects user travel factors and route recommendation tendency information; The multidimensional data mining module obtains pre-stored route planning elements. Upon receiving the user driving elements and route recommendation tendency information transmitted by the user data acquisition module, the module performs data mining and similarity matching on the user driving elements and route planning elements based on the route recommendation tendency information to obtain a set of candidate charging piles. A path visualization module obtains navigation information of the path planning for each charging pile in the candidate charging pile set and visualizes the navigation information; Among them, based on the path recommendation tendency information, data mining processing and similarity matching are performed on the user's driving elements and path planning elements to obtain a set of candidate charging piles, which specifically includes: based on the path recommendation tendency information, feature selection is performed on the user's driving elements and path planning elements to construct a multidimensional feature matrix; the initial path recommendation model is trained using the multidimensional feature matrix to obtain a target path recommendation model that integrates the user's personal characteristics; similarity matching is performed based on the target path recommendation model to determine a set of candidate charging piles; wherein, the user's driving elements include the car's starting position, the car's current cruising range and the user's living habit data; the path planning elements include charging pile charging information, charging pile queuing information and charging pile distance information.

2. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 1, characterized in that: After collecting the user's living habit data, it also includes: making predictive predictions on the user's behavior after charging based on the user's living habit data.

3. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 2, characterized in that: The route recommendation tendency information includes a time cost priority tendency and a charging cost priority tendency; wherein: Prioritize time cost, specifically: prioritize the time it takes for users to travel from their starting point to the charging station Add the time from the charging station to the destination of the user's predicted behavior The path with the least sum; Charging cost priority, specifically: giving priority to user charging costs Plus travel costs The path with the smallest sum.

4. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 3, characterized in that: Based on the route recommendation tendency information, feature selection is performed on the user's driving factors and route planning factors to construct a multi-dimensional feature matrix, which specifically includes: When the path recommendation tendency information is time cost priority, the recursive feature elimination method is used to select features, and a multi-dimensional feature matrix is ​​constructed based on the selected features; When the route recommendation tendency information is charging cost priority, the principal component analysis method is used to select features, and a multidimensional feature matrix is ​​constructed based on the selected features.

5. The new energy vehicle charging path planning system based on multidimensional data mining technology as claimed in claim 3, characterized in that: The initial path recommendation model uses the XGBoost recommendation model; When the path recommendation tendency information is a time cost priority tendency, the expression of the objective function of the initial path recommendation model is specifically: in, is a regularization term to suppress the complexity of the model. To fit the function, which has no polynomial solution, use Taylor's second-order expansion to approximate it: ； is the sample size, is the Huber Loss function, is the actual value, is the predicted value; the expression of Huber Loss function is: in, It is a hyperparameter of Huber Loss; When the route recommendation tendency information is a charging cost priority tendency, the expression of the objective function of the initial route recommendation model is specifically: in, is a regularization term to suppress the complexity of the model. is the fitting function, which has a finite number of polynomial solutions. is the sample size, is the Huber Loss function, is the actual value, is the predicted value; the expression of Huber Loss function is: in, It is a hyperparameter of Huber Loss.

6. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 1, characterized in that: Performing similarity matching based on the target path recommendation model to determine a set of candidate charging piles specifically includes: generating a scoring matrix based on the target path recommendation model; The scoring matrix is ​​used to perform similarity matching on each charging pile to obtain a set of candidate charging piles; wherein the set of candidate charging piles includes a preset number of charging piles ranked by score; the expression of the scoring matrix is ​​specifically: in, The action category of the current user predicted by the target path recommendation model, and They are respectively the time priority score corresponding to the time cost priority tendency and the charging cost priority score corresponding to the charging cost limited tendency of the target path recommendation model. The current cruising range.

7. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 6, characterized in that: Obtaining navigation information for the path planning of each charging pile in the candidate charging pile set and visualizing the navigation information, specifically including: Sending the vehicle's starting location and the candidate charging pile set to a third-party navigation software, so that the third-party navigation software obtains historical traffic data information and generates navigation information for a path planning for each charging pile in the candidate charging pile set based on the historical traffic data information, the vehicle's starting location, and the candidate charging pile set; Receive navigation information of the path planning of each charging pile in the candidate charging pile set transmitted by the third-party navigation software, and visualize the navigation information.

8. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 7, characterized in that: Obtaining navigation information of the path planning for each charging pile in the candidate charging pile set and visualizing the navigation information further includes: Determine whether the mileage corresponding to the navigation information of the path planning for each charging pile is greater than the current cruising range of the vehicle; If so, the charging pile is removed from the candidate charging pile set.

9. The new energy vehicle charging path planning system based on multidimensional data mining technology according to claim 7, characterized in that: Obtaining navigation information of the path planning for each charging pile in the candidate charging pile set and visualizing the navigation information further includes: Obtaining current traffic data, and the path visualization module transmits the current traffic data to the multidimensional data mining module, so that the multidimensional data mining module updates the candidate charging pile set; The path visualization module sends the updated candidate charging pile set to the third-party navigation software to update the navigation information of the path planning of each charging pile in the candidate charging pile set.

Images