Method and device for predicting endurance mileage of electric vehicle

Abstract

The embodiment of the invention provides an electric vehicle endurance mileage prediction method and device, and relates to the technical field of data processing, and the method specifically comprises the steps: obtaining a target vehicle operation data set; generating a target corresponding relation between the vehicle residual electric quantity value and the vehicle endurance mileage based on the vehicle residual electric quantity value and the accumulated driving mileage in the target vehicle operation data set; generating an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relation; the input sample data set comprises a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set comprises a sample vehicle endurance mileage subset; and training an initial time sequence model through the input sample data set and the output sample data set to generate a target vehicle mileage prediction model. According to the invention, the precision of predicting the endurance mileage of the vehicle can be improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of data processing, and in particular to a method and device for predicting the cruising range of an electric vehicle. Background Art

[0002] As global environmental awareness increases, electric vehicles have been widely promoted as clean energy transportation. However, the limitations of electric vehicles in terms of range and the imperfect charging facilities have led to users' "range anxiety", that is, they are worried that the electric vehicle's power is not enough to support them to reach their destination or find a charging station. This anxiety directly affects the acceptance and popularity of electric vehicles.

[0003] At present, the prediction method of electric vehicle range usually relies on linear regression models or simple deep learning models, and predicts the range through static data such as the current battery remaining power and battery voltage. Since the current prediction method usually only considers the status of the vehicle battery and ignores the impact of different driving behaviors on the vehicle battery energy consumption, the prediction results are not accurate enough.

[0004] Therefore, how to improve the accuracy in predicting vehicle range has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present application provides a method and device for predicting the cruising range of an electric vehicle, which can improve the accuracy in predicting the cruising range of the vehicle.

[0006] In a first aspect, an embodiment of the present application provides a method for predicting the range of an electric vehicle, comprising:

[0007] Acquire a target vehicle operation data set; the target vehicle operation data set includes a subset of vehicle operation data collected from multiple vehicles at preset intervals; the vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data during vehicle driving and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal travel value, and brake pedal travel value;

[0008] Based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set, a target correspondence relationship between the vehicle remaining power value and the vehicle cruising range is generated; the target correspondence relationship includes the vehicle cruising range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states;

[0009] According to the target vehicle operation data set and the target corresponding relationship, an input sample data set and an output sample data set are generated; the input sample data set includes: a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset;

[0010] The initial time series model is trained by using the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

[0011] As an optional implementation of the embodiment of the present application, the step of obtaining the target vehicle operation data set includes:

[0012] For multiple vehicles, collecting vehicle operation data at the same time every preset time to obtain an initial vehicle operation data set;

[0013] Data cleaning is performed on a plurality of the vehicle operation data in the initial vehicle operation data set to generate the target vehicle operation data set.

[0014] As an optional implementation of the embodiment of the present application, the target corresponding relationship between the vehicle remaining power value and the vehicle cruising range is generated based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set; including:

[0015] Filtering the vehicle operation data corresponding to the continuous driving segments corresponding to each vehicle from the target vehicle operation data set;

[0016] For the vehicle operation data corresponding to the plurality of continuous driving segments, obtaining the vehicle remaining power value at the start time and the end time of the continuous driving segments, and the accumulated mileage at the start time and the end time of the continuous driving segments;

[0017] Based on the remaining power values ​​of the vehicle at the start time and the end time of the plurality of continuous driving segments and the accumulated mileage at the start time and the end time of the plurality of continuous driving segments, generating the actual power usage and the actual mileage corresponding to the plurality of continuous driving segments;

[0018] Based on the multiple actual power usage and the corresponding actual mileage, a target corresponding relationship between each vehicle remaining power value and the vehicle cruising range is inferred.

[0019] As an optional implementation of the embodiment of the present application, after generating the input sample data set and the output sample data set according to the target vehicle operation data set and the target correspondence relationship, the method further includes:

[0020] Calculating the vehicle battery available energy value corresponding to each of the vehicle operation data subsets in the target vehicle operation data set to generate a sample vehicle battery available energy value subset;

[0021] The subset of sample vehicle battery available energy values ​​is added to the input sample data set.

[0022] As an optional implementation of the embodiment of the present application, the collecting of the vehicle battery available energy value corresponding to each vehicle operation data in the target vehicle operation data set includes:

[0023] When the vehicle is charging, the available charge per unit of electricity in the battery during the charging process is calculated using the current integration formula;

[0024] The vehicle battery available energy value is calculated based on the available charge amount per unit of electricity.

[0025] As an optional implementation of the embodiment of the present application, the initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle mileage prediction model, including:

[0026] The initial time series model is trained by using the input sample data set and the output sample data set to obtain an initial vehicle mileage prediction model;

[0027] Based on a hyperparameter optimization algorithm and a cross-validation algorithm, the initial vehicle mileage prediction model is optimized and adjusted to obtain a target vehicle mileage prediction model.

[0028] As an optional implementation of the embodiment of the present application, after the initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle mileage prediction model, the method further includes:

[0029] Periodically collecting current vehicle operation data of the target vehicle, and inputting the current vehicle operation data into the target vehicle mileage test model;

[0030] The target cruising range output by the target vehicle mileage test model is obtained, and the target cruising range is updated and displayed periodically.

[0031] In a second aspect, an embodiment of the present application provides a device for predicting the range of an electric vehicle, comprising:

[0032] an acquisition unit, configured to acquire a target vehicle operation data set; the target vehicle operation data set includes a vehicle operation data subset collected from a plurality of vehicles at a preset interval; the vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data during vehicle driving and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal travel value, and brake pedal travel value;

[0033] A generating unit, configured to generate a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set; the target corresponding relationship includes the vehicle cruising range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states;

[0034] A creation unit, configured to create an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship; the input sample data set includes: a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset;

[0035] A training unit is used to train the initial time series model through the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

[0036] As an optional implementation of the embodiment of the present application, the acquisition unit is specifically used to collect vehicle operation data at the same time for multiple vehicles at intervals of the preset time to obtain an initial vehicle operation data set; perform data cleaning on multiple vehicle operation data in the initial vehicle operation data set to generate the target vehicle operation data set.

[0037] As an optional implementation of an embodiment of the present application, the generating unit is specifically used to filter the vehicle operation data corresponding to the continuous driving segment corresponding to each vehicle from the target vehicle operation data set; for the vehicle operation data corresponding to the multiple continuous driving segments, obtain the vehicle remaining power value from the start time to the end time of the continuous driving segment, and the cumulative mileage from the start time to the end time of the continuous driving segment; based on the vehicle remaining power value from the start time to the end time of the multiple continuous driving segments, and the cumulative mileage from the start time to the end time of the continuous driving segments, generate the actual power usage and actual mileage corresponding to the multiple continuous driving segments; based on the multiple actual power usage and the corresponding actual mileage, infer the target corresponding relationship between the remaining power value of each vehicle and the vehicle cruising range.

[0038] As an optional implementation of the embodiment of the present application, the creation unit is also used to calculate the vehicle battery available energy value corresponding to each of the vehicle operation data subsets in the target vehicle operation data set, generate a sample vehicle battery available energy value subset; and add the sample vehicle battery available energy value subset to the input sample data set.

[0039] As an optional implementation of the embodiment of the present application, the creation unit is also used to calculate the available charge per unit of electricity of the battery during the charging process by a current integration formula when the vehicle is charging; and calculate the available energy value of the vehicle battery based on the available charge per unit of electricity.

[0040] As an optional implementation of the embodiment of the present application, the training unit is specifically used to train the initial time series model through the input sample data set and the output sample data set to obtain an initial vehicle mileage prediction model; based on a hyperparameter optimization algorithm and a cross-validation algorithm, the initial vehicle mileage prediction model is optimized and adjusted to obtain a target vehicle mileage prediction model.

[0041] As an optional implementation of the embodiment of the present application, the electric vehicle range prediction device also includes an application unit, which is specifically used to periodically collect current vehicle operation data of the target vehicle, input the current vehicle operation data into the target vehicle mileage test model; obtain the target range output by the target vehicle mileage test model, and periodically update and display the target range.

[0042] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory and a processor, wherein the memory is used to store a computer program; and the processor is used to enable the electronic device to implement the electric vehicle range prediction method described in any of the above embodiments when executing the computer program.

[0043] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a computing device, the computing device implements the electric vehicle range prediction method described in any one of the above embodiments.

[0044] The electric vehicle range prediction method provided by the embodiment of the present application is specifically as follows: obtaining a target vehicle operation data set; the target vehicle operation data set includes a subset of vehicle operation data collected from multiple vehicles at preset intervals; the vehicle operation data subset includes: vehicle status data and driving behavior data; the vehicle status data includes status data during vehicle driving and when the vehicle is stationary; the driving behavior data includes: driving status, braking status, accelerator pedal travel value, and brake pedal travel value; based on the vehicle remaining power value and accumulated mileage in the target vehicle operation data set, a target corresponding relationship between the vehicle remaining power value and the vehicle range is generated; the target The target correspondence relationship includes that each of the remaining power values ​​of the vehicle can support the vehicle's cruising range under different vehicle operating data states; an input sample data set and an output sample data set are generated according to the target vehicle operating data set and the target correspondence relationship; the input sample data set includes: multiple sample vehicle operating data subsets, sample vehicle remaining power value subsets; the output sample data set includes a sample vehicle cruising range subset; the initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range according to the current vehicle operating data.

[0045] The embodiment of the present application takes into account that different driving behaviors lead to different power consumption during vehicle driving. When the target vehicle operation data set is regenerated, a vehicle operation data subset is first constructed based on the vehicle status data and the driving behavior data, and then the target vehicle operation data set is obtained based on a large number of vehicle operation data subsets. Then, based on the target vehicle operation data set and the target correspondence, the generated input sample data set and output sample data set will also include relevant data on driving behavior. When the initial time series model is trained by the input sample data set and the output sample data set, the model can capture the complex relationship between the vehicle status data, driving behavior data and the change in vehicle battery power, so as to take into account the power consumption factors of electric vehicles as comprehensively as possible, thereby improving the performance of the target vehicle mileage prediction model and improving the prediction accuracy of the vehicle's cruising range. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 One of the step flow charts of the method for predicting the range of an electric vehicle provided in an embodiment of the present application;

[0049] Figure 2 The second step flow chart of the method for predicting the range of an electric vehicle provided in the embodiment of the present application;

[0050] Figure 3 A diagram of the architecture of a time series model of an electric vehicle range prediction method provided in an embodiment of the present application;

[0051] Figure 4 A schematic diagram of the structure of a device for predicting the range of an electric vehicle provided in an embodiment of the present application;

[0052] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0054] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0055] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way. In addition, in the description of the embodiments of the present application, unless otherwise specified, the meaning of "multiple" refers to two or more.

[0056] It should be noted that, in this article, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0057] The present application provides a method for predicting the range of an electric vehicle. Figure 1 As shown, the electric vehicle range prediction method includes the following steps S101-S104:

[0058] S101. Obtain a target vehicle operation data set.

[0059] Among them, the target vehicle operation data set includes a subset of vehicle operation data collected from multiple vehicles at preset intervals; the vehicle operation data subset includes: vehicle status data and driving behavior data; the vehicle status data includes status data during vehicle driving and when the vehicle is stationary; the driving behavior data includes: driving status, braking status, accelerator pedal travel value and brake pedal travel value.

[0060] In some embodiments, the large amount of vehicle operation data collected from multiple vehicles at preset intervals can be downloaded from some official data websites. For example, operation information covering 100 pure electric vehicles can be downloaded from an official data website. These operation data information are obtained every 10 seconds, that is, the preset time is 10 seconds. In order to ensure the timeliness of the sample data obtained according to the target vehicle operation data set, vehicle operation data of more recent years should be collected as much as possible.

[0061] Among them, the target vehicle operation data set includes vehicle status data and driving behavior data. This is because the embodiment of the present application takes into account different driving behaviors and the differences in power consumption caused by vehicle driving. Therefore, on the one hand, starting from the vehicle's own status data, and on the other hand, starting from the driving behavior data, the target vehicle operation data set is composed based on the driving behavior data and the vehicle status data, and the power consumption factors of the electric vehicle are taken into account as comprehensively as possible to improve the comprehensiveness of the sample data generated subsequently, thereby improving the accuracy of the target vehicle mileage prediction model obtained.

[0062] Specifically, the vehicle status data may include: timestamp, vehicle battery status, vehicle speed, total voltage, total current, accumulated mileage, vehicle status, charging status, battery power, battery minimum temperature value, battery maximum temperature value. The driving status in the driving behavior data can be understood as the status of the current vehicle's accelerator pedal, the braking status can be understood as the status of the current vehicle's brake pedal, the accelerator pedal travel value can be understood as a quantitative value of the degree to which the accelerator pedal is depressed; and the brake pedal travel value can be understood as a quantitative value of the degree to which the brake pedal is depressed.

[0063] In an embodiment of the present application, the vehicle status data includes status data when the vehicle is traveling and when the vehicle is stationary. This is because when the vehicle is stationary, for example, when the vehicle is powered on but not traveling, a weak current will be generated during this period, causing the battery power to drop, thereby reducing the vehicle's cruising range. Therefore, the operating data when the vehicle is stationary and the operating data when the vehicle is traveling are combined into the vehicle status data, and then the target vehicle operating data set is jointly generated for generating sample data for subsequent training.

[0064] It should be noted that the electric vehicle range prediction method provided in the embodiment of the present application is applicable to pure electric vehicles. When generating the target vehicle operation data set, different driving behaviors are taken into account, and the power consumption caused by the vehicle during driving is different. Therefore, on the one hand, starting from the vehicle's own status data, and on the other hand, starting from the driving behavior data, the target vehicle operation data set is composed based on the driving behavior data and the vehicle status data, and the factors of the electric vehicle's power consumption are taken into account as comprehensively as possible to improve the comprehensiveness of the sample data generated subsequently, thereby improving the accuracy of the target vehicle mileage prediction model obtained.

[0065] S102: Based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set, a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range is generated.

[0066] Among them, the target correspondence relationship includes the vehicle range that each vehicle remaining power value can support the vehicle to travel under different vehicle operating data states.

[0067] In some embodiments, the electric vehicle monitors the remaining power value and the accumulated mileage of the vehicle during use, and displays them on the central control screen to inform the driver of the remaining power value and the accumulated mileage of the current vehicle; the embodiment of the present application starts from the remaining power value and the accumulated mileage of the vehicle, and selects multiple continuous driving segments corresponding to each vehicle from the target vehicle operation data set, and then determines the actual mileage corresponding to the continuous driving segment by the accumulated mileage corresponding to the start time and the end time of each continuous driving segment, and determines the actual power usage corresponding to the continuous driving segment by the vehicle remaining power value corresponding to the start time and the end time of each continuous driving segment; then, referring to the actual power usage and the corresponding actual mileage, when the remaining power value of the vehicle is exactly the actual power usage, the vehicle's cruising range can be inferred according to the actual mileage; through the above, the vehicle cruising range corresponding to the remaining power value of each vehicle under different driving behaviors and vehicle states can be inferred; then, the target corresponding relationship between the remaining power value of the vehicle and the vehicle cruising range is generated, and the following sample data is created through the corresponding relationship to train the initial time series model.

[0068] Specifically, in the embodiment of the present application, the step of refining the target corresponding relationship between the vehicle remaining power value and the vehicle cruising range based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set to generate the target corresponding relationship includes the following steps A to D:

[0069] Step A: Filtering the vehicle operation data corresponding to the continuous driving segments corresponding to each vehicle from the target vehicle operation data set.

[0070] In some embodiments, the target vehicle operation data set includes a timestamp for each vehicle operation data. Based on the timestamp and other vehicle operation data under the timestamp, combined with the vehicle's operation rules, multiple continuous driving segments corresponding to each vehicle can be determined.

[0071] Step B: for the vehicle operation data corresponding to the multiple continuous driving segments, obtain the vehicle remaining power value at the start time and the end time of the continuous driving segment, and the accumulated mileage at the start time and the end time of the continuous driving segment.

[0072] In some embodiments, after determining multiple continuous driving segments corresponding to each vehicle, the start time and the end time of the current continuous driving segment can be determined based on all vehicle operation data contained in each continuous driving segment, and then the vehicle remaining power value at the start time and the end time of the continuous driving segment and the accumulated mileage at the start time and the end time of the continuous driving segment can be obtained through the vehicle operation data subsets corresponding to the start time and the end time respectively.

[0073] Step C: based on the vehicle remaining power values ​​from the start time to the end time of the plurality of continuous driving segments and the accumulated mileage from the start time to the end time of the plurality of continuous driving segments, generate the actual power usage and actual mileage corresponding to the plurality of continuous driving segments.

[0074] Exemplarily, for a continuous driving segment, after obtaining the vehicle remaining power value 1 and the vehicle remaining power value 2 corresponding to any moment and the end moment respectively, as well as the cumulative mileage 1 and the cumulative mileage 2 corresponding to any moment and the end moment respectively, the vehicle remaining power value 1 is subtracted from the vehicle remaining power value 2 to obtain the actual power usage corresponding to the continuous driving mileage, and the cumulative mileage 1 is subtracted from the cumulative mileage 2 to obtain the actual mileage corresponding to the continuous driving mileage.

[0075] Step D: infer the target corresponding relationship between the remaining power value of each vehicle and the vehicle cruising range based on the multiple actual power usage and the corresponding actual mileage.

[0076] In some embodiments, according to the above steps S1021 to S1023, the actual power usage and the corresponding actual mileage corresponding to multiple continuous driving segments under different driving behaviors and vehicle conditions can be obtained, and then referring to the actual power usage and the corresponding actual mileage, when the remaining power value of the vehicle is exactly the actual power usage, the vehicle's cruising range can be inferred based on the actual mileage; data reasoning is performed according to the above process to infer the vehicle's cruising range corresponding to each remaining power value under different driving behaviors and vehicle conditions, and then the target correspondence is generated.

[0077] S103: Generate an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship.

[0078] Among them, the input sample data set includes: multiple sample vehicle operation data subsets, and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset.

[0079] Specifically, before model training, it is necessary to create an input sample data set and an output sample data set for training the model, so that the model can use the input sample data set for training and continuously adjust its parameters to minimize the difference between the predicted output and the actual output sample data set. By repeatedly adjusting the parameters, the model can find the structure and parameter values ​​that best suit the data, thereby improving the prediction accuracy.

[0080] In an embodiment of the present application, a model is required to predict the vehicle's range based on the current vehicle's operating data and the remaining power value. Therefore, the input sample data set is generated based on the target vehicle operating data set and the vehicle's remaining power value in the target correspondence relationship, and the output sample data set includes the vehicle's range in the target correspondence relationship.

[0081] S104: Training an initial time series model using the input sample data set and the output sample data set to generate a target vehicle mileage prediction model.

[0082] The target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

[0083] In some embodiments, the initial time series model can be generated based on the KAN-LSTM network architecture, which is a model architecture that combines Kolmogorov-Arnold Networks (KAN) and Long-Short-Term Memory (LSTM). This combination can take advantage of LSTM's powerful processing capabilities for sequence data and KAN's advantages in knowledge perception and attention mechanisms, thereby playing a role in some tasks that require combining knowledge and sequence information for prediction or classification. By using the KAN network to replace the fully connected layer in LSTM, the prediction accuracy and adaptability of the model are significantly improved.

[0084] Furthermore, due to the impact of different driving behaviors on battery power consumption, for example, frequent braking, sudden acceleration and high-speed driving will accelerate the vehicle's power consumption rate. Therefore, when predicting the cruising range, it is necessary to make full use of the driving behavior data in the corresponding time period to capture the impact of driving behavior on power changes; therefore, the time series model used in the embodiment of the present application can effectively learn the complex relationship between driving behavior and battery power changes, thereby significantly improving the accuracy of cruising range prediction.

[0085] After the target vehicle mileage prediction model is generated, it is deployed to the vehicle system. Then, the target vehicle mileage prediction model is used to update the vehicle mileage prediction result at preset intervals based on the real-time collected vehicle status data and driving behavior data, and is displayed in real time on the vehicle display screen to ensure that the driver is aware of the vehicle's mileage information at any time.

[0086] The electric vehicle range prediction method provided in the embodiment of the present application is specifically as follows: obtaining a target vehicle operation data set; the target vehicle operation data set includes a subset of vehicle operation data collected from multiple vehicles at preset intervals; the vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data during vehicle driving and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal stroke value and brake pedal stroke value; based on the vehicle remaining power value and accumulated mileage in the target vehicle operation data set, a target correspondence relationship between the vehicle remaining power value and the vehicle range is generated; the target correspondence relationship includes the vehicle range that the remaining power value of each vehicle can support the vehicle to travel under different vehicle operation data states; according to the target vehicle operation data set and the target correspondence relationship, an input sample data set and an output sample data set are generated; the input sample data set includes: multiple sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle range subset; the initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle range prediction model; the target vehicle mileage prediction model is used to predict the vehicle range according to the current vehicle operation data. The embodiment of the present application takes into account that different driving behaviors lead to different power consumption during vehicle driving. When the target vehicle operation data set is regenerated, a vehicle operation data subset is first constructed based on the vehicle status data and the driving behavior data, and then the target vehicle operation data set is obtained based on a large number of vehicle operation data subsets. Then, based on the target vehicle operation data set and the target correspondence, the generated input sample data set and output sample data set will also include relevant data on driving behavior. When the initial time series model is trained by the input sample data set and the output sample data set, the model can capture the complex relationship between the vehicle status data, driving behavior data and the change in vehicle battery power, so as to take into account the power consumption factors of electric vehicles as comprehensively as possible, thereby improving the performance of the target vehicle mileage prediction model and improving the prediction accuracy of the vehicle's cruising range.

[0087] As an extension and refinement of the above embodiment, refer to Figure 2 As shown, another method for predicting the cruising range of an electric vehicle provided in an embodiment of the present application includes the following steps S201 to S206:

[0088] S201. For a plurality of vehicles, collect vehicle operation data at the same time every preset time to obtain an initial vehicle operation data set.

[0089] In some embodiments, when generating the target vehicle operation data, vehicle operation data from multiple vehicles can be collected in advance. During the collection, the vehicle operation data can be collected once every preset time, that is, the vehicle status data and driving behavior data at the same time are collected to generate the initial vehicle operation data set.

[0090] S202: Perform data cleaning on the plurality of vehicle operation data in the initial vehicle operation data set to generate the target vehicle operation data set.

[0091] Specifically, data cleaning is a key step in the data preprocessing process. Its main purpose is to detect and correct (or delete) errors, missing values, duplicate data, noisy data, and inconsistent data in the data set to improve the quality of the data and make the data more suitable for subsequent model training, thereby obtaining the target vehicle operation data set after data cleaning.

[0092] S203: Based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set, a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range is generated.

[0093] The target correspondence relationship includes the vehicle range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states;

[0094] S204: Generate an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship.

[0095] Among them, the input sample data set includes: multiple sample vehicle operation data subsets, and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset.

[0096] Step 1: Calculate the vehicle battery available energy value corresponding to each of the vehicle operation data subsets in the target vehicle operation data set to generate a sample vehicle battery available energy value subset.

[0097] In the embodiment of the present application, it is considered that the battery will gradually age with the use time; at the same time, the battery aging will cause the actual charge provided by the battery to gradually decrease under the same power conditions. Therefore, it is particularly important to accurately estimate the available energy of the battery when predicting the driving range. Since the internal structure of the battery is complex and aging is difficult to accurately predict through the model, the available energy of the vehicle battery is estimated through charging data, so that the available energy of the vehicle battery is used as an input sample feature during the training process; specifically, the refinement method of calculating the available energy value of the vehicle battery corresponding to each of the vehicle operation data subsets in the target vehicle operation data set can refer to the following steps 11 to 13:

[0098] Step 11: Collect charging data sets of vehicles involved in the target vehicle operation data set.

[0099] In some embodiments, the charging data set also includes a vehicle status data subset and a driving behavior data subset. The vehicle status data subset includes data reception time, charging status, vehicle speed, cumulative mileage, total voltage, total current, battery power, battery minimum temperature, and battery maximum temperature; the driving behavior data subset includes driving force, braking force, accelerator pedal travel value, and brake pedal travel value, which are numbers that can reflect the current driving behavior; it should be noted that the data in the driving behavior data subset can also be ignored during the charging process, because the data in the driving behavior data subset will not be used when calculating the available charge per unit of electricity during the charging process.

[0100] Step 12: Based on the charging data set, calculate the available charge per unit of electricity of the battery during the charging process by using a current integration formula.

[0101] Specifically, the specific calculation formula for calculating the available charge per unit of electricity of the battery during the charging process by the current integration formula is as follows:

[0102]

[0103] ΔSOC=SOC t2-SOC t1

[0104] Q=ΔQ / ΔSOC

[0105] Among them, ΔQ is the change in charge from t1 to t2 during charging; I(t) is the charging current; ΔSOC is the change in battery power from t1 to t2; SOC t1 is the battery power at t1; SOC t2 is the battery power at t2; Q is the available charge per unit of power.

[0106] Step 13: Based on the available charge per unit of electricity, obtain the available energy value of the vehicle battery.

[0107] The current vehicle battery available energy value can be obtained by multiplying the available charge per unit by the number of all units in the vehicle battery.

[0108] It should be noted that the aging characteristics of vehicle batteries change little in the short term. Therefore, when applying the model after obtaining the trained vehicle mileage prediction model, the vehicle charging data set can be obtained once at a preset interval, such as one month, and then the Q value can be calculated. The Q value in the model input data can be updated regularly, and the input of the model's vehicle battery available energy value can be adjusted according to the actual situation of the vehicle battery, thereby improving the prediction accuracy.

[0109] Step 2: Add the subset of available energy values ​​of the sample vehicle batteries to the input sample data set.

[0110] Specifically, after obtaining the subset of available energy values ​​of the sample vehicle battery, the subset of available energy values ​​of the sample vehicle battery can be used as input sample data to train the model when training the initial time series model, so that the model can comprehensively learn the available energy value of the vehicle battery, the complex relationship between driving behavior and battery power changes, thereby significantly improving the accuracy of the range prediction.

[0111] S205 . Training an initial time series model using the input sample data set and the output sample data set to obtain an initial vehicle mileage prediction model.

[0112] Specifically, refer to Figure 3 As shown, it is the architecture diagram of the initial time series model, where the architecture of the initial time series model includes a long short-term memory network 31 and a KAN layer 32. The KAN network is used to replace the fully connected layer in the long short-term memory network LSMT because the fully connected layer in the LSTM is mainly used to linearly transform the features output by the LSTM unit to map to the target dimension, while the KAN network may have a more complex and effective feature extraction mechanism. For example, the KAN network simulates complex nonlinear relationships through edges parameterized by spline functions, and can mine more complex nonlinear relationships in the data. This is in contrast to the relatively simple linear combination method of the fully connected layer, so that when processing sequence data, more representative features can be extracted, which is helpful for subsequent prediction tasks; thereby, the prediction accuracy and adaptability of the target vehicle mileage prediction model can be improved.

[0113] S206: Based on a hyperparameter optimization algorithm and a cross-validation algorithm, the initial vehicle mileage prediction model is optimized and adjusted to obtain a target vehicle mileage prediction model.

[0114] In some embodiments, after the initial vehicle mileage prediction model is obtained in step S205, it is necessary to further optimize and adjust the initial vehicle mileage prediction model to obtain a more accurate target vehicle mileage prediction model; specifically, in an embodiment of the present application, a specific method for optimizing and adjusting the initial vehicle mileage prediction model is to first perform a performance evaluation on the initial vehicle mileage prediction model, and then better optimize and adjust the initial vehicle mileage prediction model to obtain the target vehicle mileage prediction model.

[0115] Specifically, the performance of the initial vehicle mileage prediction model can be evaluated using a 10-fold cross validation method; the evaluation indicators include mean absolute error (MAE), root mean square error (RMSE) and determination coefficient (R2), and the calculation formula of the evaluation indicators is as follows:

[0116]

[0117] Among them, Y i is the true value, are the predicted values; these indicators will be used to measure the predictive ability and accuracy of the model.

[0118] It should be noted that after obtaining the target vehicle mileage prediction model, the target vehicle mileage prediction model will be deployed to the vehicle side, and then the vehicle side will predict the vehicle mileage corresponding to the current remaining power based on the real-time collected data, and dynamically update it on the vehicle's central control screen. At the same time, it is also necessary to regularly save and update data based on the user's driving habits, and gradually form a time series database. This database is used to optimize the reference data within the platform, so that the model can more accurately predict future mileage and provide personalized driving suggestions.

[0119] Specifically, after the initial time series model is trained by the input sample data set and the output sample data set to generate the target vehicle mileage prediction model, in the actual model application process, the specific steps include the following steps A and B:

[0120] Step A: periodically collect current vehicle operation data of the target vehicle, and input the current vehicle operation data into the target vehicle mileage test model.

[0121] Specifically, the target vehicle mileage test model can be used to periodically predict the cruising range of the current vehicle. For example, the vehicle operation data of the current target vehicle, i.e., the vehicle status data and driving behavior data, can be collected every 10 seconds, and the latest vehicle battery available energy value can be obtained. The above data is input into the target vehicle mileage test model, and the cruising range corresponding to the current remaining power is then obtained through model reasoning.

[0122] Step B: obtaining the target cruising range output by the target vehicle mileage test model, and periodically updating and displaying the target cruising range.

[0123] By periodically inputting data into the target vehicle mileage test model, the target vehicle mileage corresponding to the current remaining power can be updated on the vehicle's central control screen each time the target mileage is obtained, so that the driver can obtain the target mileage in time, thereby helping the driver to effectively plan the itinerary and set the charging strategy.

[0124] For example, in order to better illustrate the technical solution of this application, the following is a detailed introduction in conjunction with a specific embodiment. The electric vehicle operation data set used in the present invention comes from the Tianjin Industry and Information Technology Bureau, covering the operation information of 100 pure electric vehicles. The data recording time is from March 30, 2024 to April 7, 2024, and one piece of data is updated every 10 seconds to ensure the timeliness and integrity of the data.

[0125] First, obtain the original data of the electric vehicle operation information and extract valid information. Each record of the original data consists of 23 fields, including: vehicle VIN, data collection time, latitude and longitude, vehicle status, charging status, operating mode, vehicle speed, cumulative mileage, total voltage, total current, battery power, DC-DC status, gear, insulation resistance, minimum temperature value, maximum temperature value, driving force, braking force, accelerator pedal travel value and brake pedal travel value. According to the needs of the embodiment of the present application, after extracting valid information, the data vehicle operation data set and charging data set are collected. Some of the data are shown in Table 1 and Table 2:

[0126] Table 1. Some vehicle operation data sets

[0127]

[0128] Table 2. Some vehicle charging data sets

[0129]

[0130] In the above Tables 1 and 2, a charging status of "1" indicates that the vehicle is currently in a charging state; a charging status of "3" indicates that the vehicle is currently in a non-charging state; a driving status of "1" and "0" can be used to indicate that the driving pedal is pressed and the driving pedal is not pressed respectively; a braking status of "1" and "0" can be used to indicate that the brake pedal is pressed and the brake pedal is not pressed respectively.

[0131] Furthermore, the corresponding remaining power value and cumulative mileage of the vehicle, as well as the available energy value of the vehicle battery, are obtained. After integrating all valid data, the final partial sample data is formed. Some of the data are shown in Table 3:

[0132] Table 3. Some sample input data sets and sample output data sets

[0133]

[0134] Finally, the initial time series model is trained using the sample data set as exemplified in Table 3 above to obtain the target vehicle mileage prediction model.

[0135] Based on the same inventive concept, as an implementation of the above method, an embodiment of the present application also provides an electric vehicle range prediction device, which corresponds to the above method embodiment. For ease of reading, this embodiment will no longer repeat the details of the above method embodiment one by one, but it should be clear that an electric vehicle range prediction device in this embodiment can correspond to all the contents of the above method embodiment.

[0136] The present application provides a device for predicting the range of an electric vehicle. Figure 4 is a schematic diagram of the structure of the electric vehicle range prediction device, such as Figure 4 As shown, the electric vehicle cruising range prediction device 400 includes:

[0137] The acquisition unit 401 is used to acquire a target vehicle operation data set; the target vehicle operation data set includes a vehicle operation data subset collected from multiple vehicles at a preset interval; the vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data when the vehicle is traveling and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal travel value, and brake pedal travel value;

[0138] A generating unit 402 is configured to generate a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set; the target corresponding relationship includes the vehicle cruising range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states;

[0139] The creation unit 403 is used to create an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship; the input sample data set includes: a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset;

[0140] The training unit 404 is used to train the initial time series model through the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

[0141] As an optional implementation of the embodiment of the present application, the acquisition unit 401 is specifically used to collect vehicle operation data at the same time for multiple vehicles at intervals of the preset time to obtain an initial vehicle operation data set; perform data cleaning on multiple vehicle operation data in the initial vehicle operation data set to generate the target vehicle operation data set.

[0142] As an optional implementation of an embodiment of the present application, the generating unit 402 is specifically used to filter the vehicle operation data corresponding to the continuous driving segment corresponding to each vehicle from the target vehicle operation data set; for the vehicle operation data corresponding to the multiple continuous driving segments, obtain the vehicle remaining power value from the start time to the end time of the continuous driving segment, and the cumulative mileage from the start time to the end time of the continuous driving segment; based on the vehicle remaining power value from the start time to the end time of the multiple continuous driving segments, and the cumulative mileage from the start time to the end time of the continuous driving segment, generate the actual power usage and actual mileage corresponding to the multiple continuous driving segments.

[0143] As an optional implementation of the embodiment of the present application, the creation unit 403 is also used to calculate the vehicle battery available energy value corresponding to each of the vehicle operation data subsets in the target vehicle operation data set, generate a sample vehicle battery available energy value subset; and add the sample vehicle battery available energy value subset to the input sample data set.

[0144] As an optional implementation of the embodiment of the present application, the creation unit 403 is also used to calculate the available charge per unit of electricity of the battery during the charging process by using a current integration formula when the vehicle is charging; and calculate the available energy value of the vehicle battery based on the available charge per unit of electricity.

[0145] As an optional implementation of the embodiment of the present application, the training unit 404 is specifically used to train the initial time series model through the input sample data set and the output sample data set to obtain an initial vehicle mileage prediction model; based on a hyperparameter optimization algorithm and a cross-validation algorithm, the initial vehicle mileage prediction model is optimized and adjusted to obtain a target vehicle mileage prediction model.

[0146] As an optional implementation of the embodiment of the present application, the electric vehicle range prediction device also includes an application unit, which is specifically used to periodically collect current vehicle operation data of the target vehicle, input the current vehicle operation data into the target vehicle mileage test model; obtain the target range output by the target vehicle mileage test model, and periodically update and display the target range.

[0147] Based on the same inventive concept, an embodiment of the present disclosure also provides an electronic device. Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure, such as Figure 5 As shown, the electronic device provided in this embodiment includes: a memory 501 and a processor 502, wherein the memory 501 is used to store a computer program; and the processor 502 is used to execute the audio data processing method provided in the above embodiment when executing the computer program.

[0148] Based on the same inventive concept, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the computing device implements the electric vehicle range prediction method provided in the above embodiment.

[0149] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media that include computer-usable program code.

[0150] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0151] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0152] Computer readable media include permanent and non-permanent, removable and non-removable storage media. Storage media can be implemented by any method or technology to store information, and the information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements 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 application.

Claims

1. A method for predicting the cruising range of an electric vehicle, characterized in that: include: Obtain target vehicle operation data set; The target vehicle operation data set includes a subset of vehicle operation data collected from a plurality of vehicles at preset time intervals; The vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data when the vehicle is in motion and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal travel value and brake pedal travel value; Based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set, a target correspondence relationship between the vehicle remaining power value and the vehicle cruising range is generated; the target correspondence relationship includes the vehicle cruising range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states; According to the target vehicle operation data set and the target corresponding relationship, an input sample data set and an output sample data set are generated; the input sample data set includes: a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset; The initial time series model is trained by using the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

2. The method according to claim 1, characterized in that The step of obtaining a target vehicle operation data set includes: For multiple vehicles, collecting vehicle operation data at the same time every preset time to obtain an initial vehicle operation data set; Data cleaning is performed on a plurality of the vehicle operation data in the initial vehicle operation data set to generate the target vehicle operation data set.

3. The method according to claim 1, characterized in that The method of generating a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set comprises: Filtering the vehicle operation data corresponding to the continuous driving segments corresponding to each vehicle from the target vehicle operation data set; For the vehicle operation data corresponding to the plurality of continuous driving segments, obtaining the vehicle remaining power value at the start time and the end time of the continuous driving segments, and the accumulated mileage at the start time and the end time of the continuous driving segments; Based on the remaining power values ​​of the vehicle from the start time to the end time of the plurality of continuous driving segments and the accumulated mileage from the start time to the end time of the plurality of continuous driving segments, generating the actual power usage and actual mileage corresponding to the plurality of continuous driving segments; Based on the multiple actual power usage and the corresponding actual driving mileage, a target corresponding relationship between each vehicle remaining power value and the vehicle cruising range is inferred.

4. The method according to claim 1, characterized in that: After generating an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship, the method further includes: Calculating the vehicle battery available energy value corresponding to each of the vehicle operation data subsets in the target vehicle operation data set to generate a sample vehicle battery available energy value subset; The subset of sample vehicle battery available energy values ​​is added to the input sample data set.

5. The method according to claim 4, characterized in that The collecting of the vehicle battery available energy value corresponding to each vehicle operation data in the target vehicle operation data set includes: Collecting charging data sets of vehicles involved in the target vehicle operation data set; Based on the charging data set, calculating the available charge per unit of electricity of the battery during the charging process by a current integration formula; Based on the available charge amount per unit of electricity, the available energy value of the vehicle battery is obtained.

6. The method according to claim 1, characterized in that The initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle mileage prediction model, including: The initial time series model is trained by using the input sample data set and the output sample data set to obtain an initial vehicle mileage prediction model; Based on a hyperparameter optimization algorithm and a cross-validation algorithm, the initial vehicle mileage prediction model is optimized and adjusted to obtain a target vehicle mileage prediction model.

7. The method according to claim 1, characterized in that After the initial time series model is trained by the input sample data set and the output sample data set to generate a target vehicle mileage prediction model, the method further includes: Periodically collecting current vehicle operation data of the target vehicle, and inputting the current vehicle operation data into the target vehicle mileage test model; The target cruising range output by the target vehicle mileage test model is obtained, and the target cruising range is updated and displayed periodically.

8. A device for predicting the range of an electric vehicle, characterized in that: include: An acquisition unit, used for acquiring a target vehicle operation data set; The target vehicle operation data set includes a subset of vehicle operation data collected from a plurality of vehicles at preset time intervals; The vehicle operation data subset includes: vehicle state data and driving behavior data; the vehicle state data includes state data when the vehicle is in motion and when the vehicle is stationary; the driving behavior data includes: driving state, braking state, accelerator pedal travel value and brake pedal travel value; A generating unit, configured to generate a target corresponding relationship between the vehicle remaining power value and the vehicle cruising range based on the vehicle remaining power value and the accumulated mileage in the target vehicle operation data set; the target corresponding relationship includes the vehicle cruising range that each vehicle remaining power value can support the vehicle to travel under different vehicle operation data states; A creation unit, configured to create an input sample data set and an output sample data set according to the target vehicle operation data set and the target corresponding relationship; the input sample data set includes: a plurality of sample vehicle operation data subsets and a sample vehicle remaining power value subset; the output sample data set includes a sample vehicle cruising range subset; A training unit is used to train the initial time series model through the input sample data set and the output sample data set to generate a target vehicle mileage prediction model; the target vehicle mileage prediction model is used to predict the vehicle's cruising range based on current vehicle operation data.

9. An electronic device, characterized in that: include: A memory and a processor, wherein the memory is used to store a computer program; and the processor is used to enable the electronic device to implement the electric vehicle range prediction method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a computing device, the computing device implements the electric vehicle range prediction method according to any one of claims 1 to 7.

Images