Method and device for predicting pure electric cruising range, electronic equipment and vehicle

Abstract

The application relates to a pure-electric endurance mileage prediction method and device, electronic equipment and a vehicle, and relates to the technical field of data processing. The method comprises the following steps: obtaining driving data, vehicle data and driving environment information of a vehicle to be predicted; extracting feature data from the driving data, the vehicle data and the driving environment information and inputting the feature data into a preset machine learning model for calculation, so as to obtain a pure-electric residual mileage value corresponding to the same vehicle type as the vehicle to be predicted according to the feature data; wherein the feature data is used to represent the vehicle battery state and the vehicle power performance; and the pure-electric residual mileage value output by the preset machine learning model is determined as the pure-electric endurance mileage prediction value of the vehicle to be predicted. By applying the technical solution of the application, the prediction accuracy of the pure-electric endurance mileage of a new energy vehicle can be improved, the trust degree of a user for the prediction result is improved, mileage anxiety is reduced, and the driving experience of the user can be improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a method, apparatus, electronic device, and vehicle for predicting pure electric driving range. Background Technology

[0002] New energy vehicles refer to automobiles that use unconventional vehicle fuels as their power source (or use conventional vehicle fuels and adopt new on-board power devices), and integrate advanced technologies in vehicle power control and drive, resulting in vehicles with advanced technical principles, new technologies, and new structures.

[0003] Currently, the pure electric driving range of new energy vehicles is predicted by looking up tables based on the battery's state of charge (SOC) and temperature. However, this method has low prediction accuracy, which leads to low user trust in the prediction results, easily triggering range anxiety and thus affecting the user's driving experience. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, electronic device and vehicle for predicting pure electric driving range. The main purpose is to improve the technical problem that the prediction accuracy of the existing pure electric driving range prediction methods is low, which leads to low user confidence in the prediction results, easily causes range anxiety, and thus affects the user's driving experience.

[0005] Firstly, this application provides a method for predicting pure electric driving range, including:

[0006] Obtain driving data, vehicle data, and driving environment information of the vehicle to be predicted;

[0007] Feature data is extracted from the driving data, vehicle data, and driving environment information and input into a preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data; wherein, the feature data is used to characterize the vehicle battery status and vehicle power performance, and the preset machine learning model is trained based on historical driving data samples of the sample vehicle dataset and historical environment data samples corresponding to the historical driving data samples.

[0008] The remaining pure electric range value output by the preset machine learning model is determined as the predicted pure electric range value of the vehicle to be predicted.

[0009] Secondly, this application provides a device for predicting pure electric driving range, comprising:

[0010] The acquisition module is configured to acquire driving data, vehicle data, and driving environment information of the vehicle to be predicted;

[0011] The input module is configured to extract feature data from the driving data, vehicle data, and driving environment information and input it into a preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data; wherein, the feature data is used to characterize the vehicle battery status and vehicle power performance, and the preset machine learning model is trained based on historical driving data samples of the sample vehicle dataset and historical environment data samples corresponding to the historical driving data samples;

[0012] The determination module is configured to determine the remaining pure electric range value output by the preset machine learning model as the predicted pure electric range value of the vehicle to be predicted.

[0013] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the pure electric driving range prediction method described in the first aspect.

[0014] Fourthly, this application provides an electronic device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the computer program to implement the pure electric driving range prediction method described in the first aspect.

[0015] Fifthly, this application provides a vehicle including: electronic equipment as described in the fourth aspect.

[0016] By employing the above technical solution, this application provides a method, device, electronic device, and vehicle for predicting pure electric driving range. Compared with existing technologies, this application can accurately predict the pure electric driving range of a vehicle based on its driving data, vehicle data, and driving environment information using a machine learning model. Specifically, feature data characterizing the vehicle's battery state and power performance can be extracted from the driving data, vehicle data, and driving environment information of the vehicle to be predicted and input into a preset machine learning model for calculation. Based on this feature data, the remaining pure electric driving range corresponding to the same vehicle model as the vehicle to be predicted can be obtained. This preset machine learning model is trained based on historical driving data samples from a sample vehicle dataset and historical environment data samples corresponding to those historical driving data samples. Therefore, the predicted pure electric driving range of the vehicle to be predicted can be determined based on the remaining pure electric driving range value output by the preset machine learning model. By applying the technical solution of this application, the prediction accuracy of the pure electric driving range of new energy vehicles can be improved, increasing user confidence in the prediction results, reducing range anxiety, and enhancing the user's driving experience.

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a method for predicting pure electric driving range provided in an embodiment of this application is shown.

[0021] Figure 2 This illustration shows the display effect of the vehicle remaining mileage prediction value provided in the embodiment of this application;

[0022] Figure 3 A flowchart illustrating another method for predicting pure electric driving range provided in an embodiment of this application is shown.

[0023] Figure 4 A schematic diagram illustrating an example of using the machine learning model provided in an embodiment of this application is shown;

[0024] Figure 5 A schematic diagram illustrating an example of data feature extraction provided in an embodiment of this application is shown;

[0025] Figure 6 A schematic diagram illustrating a tag calculation example provided in an embodiment of this application is shown;

[0026] Figure 7 A schematic diagram of the structure of a pure electric driving range prediction device provided in an embodiment of this application is shown. Detailed Implementation

[0027] To better understand the above-mentioned objectives, features, and advantages of this application, the solution of this application will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0028] To address the technical issue of low prediction accuracy in current pure electric driving range prediction methods, which leads to low user trust in the prediction results, easily triggers range anxiety, and consequently affects the user's driving experience, this embodiment provides a method for predicting pure electric driving range, such as... Figure 1 As shown, the method includes:

[0029] Step 101: Obtain the driving data, vehicle data, and driving environment information of the vehicle to be predicted.

[0030] Historical driving data can be data related to user driving behavior, specifically including: energy consumption per kilometer of the vehicle's current journey, average current, average vehicle speed, acceleration, etc. Specifically, energy consumption per kilometer can be calculated using voltage, current, and time signals recorded by the in-vehicle controller, along with the distance traveled. Average current can be calculated using current and time signals recorded by the in-vehicle controller. Average vehicle speed can be calculated using the vehicle speed from the speedometer and the time signal recorded by the controller. Acceleration can be calculated using the vehicle's acceleration sensor.

[0031] Vehicle data can be vehicle status data, specifically including: battery state of health (SOH), maximum battery pack temperature, minimum battery pack temperature, state of charge (SOC), and current battery voltage. SOH is calculated based on the physical properties of the battery's internal voltage and current, as well as its charge level. The maximum and minimum battery pack temperatures are obtained using temperature sensors. SOC is calculated based on the physical properties of the battery's internal voltage and current. The current battery voltage is calculated using a voltage sensor.

[0032] The driving environment information can be the environmental information outside the vehicle, specifically including information such as the vehicle's outdoor temperature. The vehicle's outdoor temperature can be calculated using a temperature sensor.

[0033] Step 102: Extract feature data from the driving data, vehicle data and driving environment information of the vehicle to be predicted and input it into the preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data.

[0034] In this method, feature data can be used to characterize the vehicle's battery status and power performance. The pre-set machine learning model can be trained in advance using sample data from a large number of vehicles of the same model as the vehicle to be predicted. Specifically, the pre-set machine learning model can be trained based on historical driving data samples and corresponding historical environmental data samples from the sample vehicle dataset. Each vehicle model can have its own corresponding pre-set machine learning model. In practical applications, the pre-set machine learning model corresponding to the vehicle model to be predicted can be found for calculation. For example, by calculating using this model, the remaining pure electric range of the sample vehicle of the same model that is most similar to the current situation of the vehicle to be predicted (feature data from driving data, vehicle data, and driving environment information used to characterize the vehicle's battery status and power performance) can be found and used as the model's output. Compared with the traditional prediction method based on battery SOC and temperature through table lookup, the method proposed in this embodiment considers more comprehensive factors. It can use the pre-set machine learning model to calculate based on the feature data from driving data, vehicle data, and driving environment information used to characterize the vehicle's battery status and power performance, and then combine this with historical driving data samples from the vehicle dataset and corresponding historical environmental data samples for analysis to accurately predict the pure electric range.

[0035] Step 103: Determine the remaining pure electric range value output by the preset machine learning model as the predicted pure electric range value of the vehicle to be predicted.

[0036] For example, after obtaining the corresponding remaining pure electric range value using a preset machine learning model, it can be displayed as the predicted pure electric range value of the vehicle on the dashboard in the driver's seat, such as... Figure 2 As shown, this information accurately helps drivers understand the vehicle's remaining pure electric range, enabling them to plan their route. Based on the predicted remaining pure electric range, drivers can choose to continue to their destination or visit the nearest charging station.

[0037] Compared with existing technologies, this embodiment can accurately predict the pure electric range of a vehicle based on its driving data, vehicle data, and driving environment information using a machine learning model. Specifically, feature data characterizing the vehicle's battery status and power performance are extracted from the driving data, vehicle data, and driving environment information and input into a preset machine learning model for calculation. This model is trained using historical driving data samples and corresponding historical environment data samples from a sample vehicle dataset. The predicted pure electric range of the vehicle is then determined based on the remaining pure electric range output by the preset machine learning model. By applying the technical solution of this embodiment, the prediction accuracy of the pure electric range of new energy vehicles can be improved, increasing user confidence in the prediction results, reducing range anxiety, and enhancing the user's driving experience.

[0038] Furthermore, as a refinement and extension of the above embodiments, in order to fully illustrate the specific implementation process of the method in this embodiment, this embodiment provides the following: Figure 3 The specific method shown includes:

[0039] Step 201: Obtain historical driving data samples, vehicle data samples, and historical environmental data samples for each trip of the target model's sample vehicle, and obtain the historical remaining driving range value corresponding to each trip of the sample vehicle.

[0040] When it is necessary to build a preset machine learning model corresponding to the target vehicle model, the historical driving data, vehicle data and historical driving environment information of the sample vehicle for each trip of the target vehicle can be obtained, as well as sample data such as the historical remaining driving range value corresponding to each trip of the sample vehicle.

[0041] Step 202: Construct a dataset based on historical driving data samples, vehicle data samples, and historical environmental data samples for each trip of the sample vehicle, as well as the historical remaining driving range value corresponding to each trip.

[0042] In this embodiment, the process of building a machine learning model can be divided into several steps, including data feature selection, label calculation, dataset partitioning, model modeling and training, and model prediction. Among these, for example... Figure 4 As shown, features are the inputs to the model, and labels are the outputs of the model.

[0043] The process of selecting data features can correspond to the process of determining the sample feature data of the dataset in step 202. Optionally, step 202 may specifically include: extracting energy consumption per kilometer of travel, historical average current, historical average vehicle speed, and acceleration from historical driving data samples; and extracting battery health, maximum battery pack temperature, minimum battery pack temperature, battery state of charge, and battery voltage from vehicle data samples; and extracting the vehicle's outdoor temperature from historical environmental data samples; then, the extracted energy consumption per kilometer of travel, battery state of charge, battery voltage, historical average current, historical average vehicle speed, acceleration, battery health, maximum battery pack temperature, minimum battery pack temperature, and vehicle outdoor temperature for each trip are used as data. The sample feature data is collected, and the remaining driving range value corresponding to each trip can be used as the sample label information corresponding to the sample feature data. Among them, the energy consumption per kilometer of trip, the historical average current, and the historical average vehicle speed can be non-dynamic data, while acceleration, battery health, maximum battery pack temperature, minimum battery pack temperature, battery state of charge, battery voltage, and vehicle outdoor temperature are dynamic instantaneous data. These data can be periodically reported on the vehicle side. Subsequently, the model can be trained through these massive data samples to obtain a machine learning model that can accurately predict the remaining pure electric range, as shown in step 203.

[0044] For example, based on the historical driving data, environmental information, and vehicle data of the studied vehicle model, feature selection and processing are performed, and the selected features include... Figure 5 As shown in the figure. Among them, except for history_energy_per_km, BMS_RESSCur_mean, and ESP_VehicleSpeed_mean, the other features are signals collected from the CAN bus. history_energy_per_km is calculated by dividing the electricity consumed in the historical trip by the historical mileage. BMS_RESSCur_mean and ESP_VehicleSpeed_mean are the average values ​​of historical current and historical vehicle speed, respectively.

[0045] The label calculation process corresponds to the process of determining the sample label information of the dataset in step 202. Optionally, the historical remaining driving range value corresponding to each trip of the sample vehicle can be obtained, specifically including: calculating the remaining driving range value of the sample vehicle at time tm using Formula 1.

[0046]

[0047] In each trip, the sample vehicle used pure electric mode from time t1 to time tn, where t1 <tm<tn,Label tmis the remaining cruising range value of the sample vehicle at time tm, S is the actual travel distance of the sample vehicle within the time interval from time tm to time tn, and SOE tn is the remaining battery energy of the sample vehicle at time tn, and e is the average energy consumption per kilometer of the sample vehicle within the time interval from time t1 to time tn.

[0048] For example, the label (Label) is the remaining electric range. Since there is almost no case in the current historical data where the user completely depletes the battery, accurate remaining electric range cannot be obtained from the training data. Therefore, the above method is adopted in this embodiment to replace the calculation of the real Label. As Figure 6 shown, assume that the user drives from time t1 to time tn in pure electric mode, and t1 < tm < tn. Then the remaining electric cruising range value of the sample vehicle at time tm can be shown as in Formula 1.

[0049] At the end of the journey at time tn, the battery level is greater than 0 in most scenarios. For the remaining electric cruising range corresponding to the battery level at this time, it is calculated by dividing the available remaining battery power at this time by the average energy consumption per kilometer of the whole journey. Specifically, it can be shown as in Formula 2. For example, the remaining electric cruising range value of the sample vehicle at time tn is calculated through Formula 2.

[0050]

[0051] Among them, Label tn is the remaining electric cruising range value of the sample vehicle at time tn, SOE tn is the remaining battery energy of the sample vehicle at time tn, and e is the average energy consumption per kilometer of the sample vehicle within the time interval from time t1 to time tn.

[0052] Step 203: Train a preset machine learning model through the constructed data set.

[0053] For the process of dividing the data set, it corresponds to the process of model training in Step 203. Optionally, Step 203 may specifically include: using the cross-validation method (Cross Validation, CV) to divide the data set into a training set and a validation set, and subsequent cross-validation. For example, using the cross-validation method for training set / validation set division and subsequent cross-validation means dividing the data set into k mutually exclusive subsets of similar sizes. Each time, k - 1 subsets are used for training and 1 subset is used for validation, and training is performed k times. Finally, the mean of the k training results is returned.

[0054] The model modeling and training process can correspond to the model training process in step 203. Optionally, step 203 may also include: first, using the Xgboost ensemble learning algorithm to tune the model parameters based on the dataset to select the optimal parameter combination; and then completing the model training after the optimal parameter combination is determined.

[0055] There are many types of machine learning models, and the chosen algorithm is XGBoost, an ensemble learning algorithm. It's an optimized distributed gradient boosting tree algorithm and currently the fastest and best open-source boosting tree toolkit, more than 10 times faster than common toolkits. XGBoost is an ensemble algorithm that builds multiple weak evaluators on the data, aggregating the modeling results of all weak evaluators to achieve better regression or classification performance than a single model. By building weak evaluators one by one, multiple iterations are gradually accumulated to form a strong evaluator formed by ensembled tree models. For a regression tree (predicting pure electric range is a regression problem), the value at each leaf node is the mean of all samples at that leaf node.

[0056] Each leaf node has a prediction score, also known as a leaf weight. This leaf weight is the regression value of all samples at that leaf node in that tree, denoted by fk(xi) or w, where fk represents the k-th decision tree and xi represents the feature vector corresponding to sample i. When there are multiple trees, the regression result of the ensemble model is the sum of the prediction scores of all trees.

[0057] The core algorithm of Xgboost is as follows: a. Continuously add trees and perform feature splits to grow a tree. Each time a tree is added, it learns a new function fk(x) to fit the residual of the previous prediction; b. When training is complete and k trees are obtained, to predict the score of a sample, it is based on the features of the sample, which will fall into a corresponding leaf node in each tree, and each leaf node corresponds to a score; c. Finally, the scores corresponding to each tree are added together to obtain the predicted value of the sample.

[0058] The process involves continuously enumerating different tree structures, then using a scoring function to find the optimal tree structure, which is then added to the model. This process is repeated continuously. This search process uses a greedy algorithm. A feature is selected for splitting, the minimum loss function is calculated, another feature is selected for splitting, and another minimum loss function is obtained. By enumerating and finding the best-performing structure, the tree is split, resulting in a sapling. After preparing the feature and label data and deciding to use the XGBoost model, the next step is to build and train the XGBoost model.

[0059] Model building is implemented using XgBoost's own libraries. The specific process is as follows: a. Import the necessary data packages (including those); b. Construct the training and test sets into a format usable by XgBoost; c. Train the model. The model training process is the process of tuning the model parameters based on the training dataset to select the optimal parameter combination. The grid search algorithm is a method to optimize model performance by traversing a given combination of parameters.

[0060] The model parameters may include: objective, eta, min_child_weight, gamma, max_delta_step, subsample, and lambda.

[0061] `objective` [default: `reg:linear`] defines the loss function to be minimized (generally, "binary:logistic" and "rank:pairwise" are used for binary classification problems). The most commonly used values ​​are (and can be customized): `binary:logistic` performs logistic regression for binary classification, returning the predicted probability (not the class). `multi:softmax` uses a softmax multi-classifier, returning the predicted class (not the probability). In this case, an additional parameter, `num_class` (the number of classes), is required. `multi:softprob` has the same parameters as `multi:softmax`, but returns the probability of each data point belonging to each class.

[0062] eta [default 0.3]: Learning rate; used to update the weights of leaf nodes. Multiplying by this coefficient prevents excessively large step sizes. A larger parameter value may prevent convergence. Setting the learning rate eta to a smaller value allows for more careful learning in subsequent iterations.

[0063] `min_child_weight` [default 1]: The sum of the minimum weights of the leaf nodes; this parameter is used to avoid overfitting. When this value is large, it can prevent the model from learning local special samples. This parameter has a significant impact on the results. The smaller this parameter is, the easier it is to overfit; however, if this value is too high, it will lead to underfitting. This parameter needs to be adjusted using cross-validation (CV).

[0064] `gamma` [default 0]: When splitting a node, the node will only be split if the loss function value decreases after the split. `gamma` specifies the minimum decrease in the loss function required for node splitting. The larger this parameter is, the more conservative the algorithm. A value between 0.1 and 0.2 is acceptable. This parameter also needs to be adjusted later.

[0065] max_delta_step [default 0]: This parameter takes effect in the update step, limiting the maximum step size for the weight change of each tree.

[0066] subsample [default 1]: The proportion of random samples used to generate each tree. Lower values ​​make the algorithm more conservative and prevent overfitting, while values ​​that are too small can lead to underfitting.

[0067] lambda [default 1]: This parameter controls the regularization part of XgBoost. It can be used to further reduce overfitting.

[0068] The general order for XgBoost tuning can be as follows: determine a relatively large learning rate of 0.1, tune num_boost_round, tune max_depth and min_weight parameters, tune gamma parameter, tune regularization parameter, and then reduce the learning rate. After the model parameters are determined, the model training is complete. The next step is to save the model, which is the preset machine learning model.

[0069] The pre-trained model can then be called to predict the pure electric driving range (Label) on new data and evaluate its performance. The accuracy of the pure electric driving range prediction is evaluated by the mean absolute error (MAE).

[0070] Step 204: Upon receiving the instruction to predict the pure electric range of the vehicle to be predicted, obtain the driving data, vehicle data, and driving environment information of the vehicle to be predicted.

[0071] Step 205: Extract feature data from the driving data, vehicle data and driving environment information of the vehicle to be predicted and input it into the preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data.

[0072] Optionally, step 205 may specifically include: extracting the energy consumption per kilometer of the current trip of the vehicle to be predicted, the battery state of charge, battery voltage, average current, average vehicle speed, acceleration, battery health, maximum battery pack temperature, minimum battery pack temperature, and outdoor temperature, and inputting them into a preset machine learning model for calculation, so as to obtain the pure electric remaining mileage value of sample vehicles of the same model as the vehicle to be predicted under similar historical conditions based on these feature data.

[0073] Step 206: Determine the remaining pure electric range value output by the preset machine learning model as the predicted pure electric range value of the vehicle to be predicted.

[0074] This embodiment provides a machine learning-based method for predicting pure electric driving range. It combines historical driving behavior data and environmental data with a machine learning model to predict driving range with high accuracy. This improves the accuracy of range prediction by incorporating real-time vehicle information, environmental information, and user driving behavior data into the calculated range prediction results, increasing user trust in the predictions, reducing range anxiety, and enhancing the user's driving experience.

[0075] Furthermore, as Figure 1 and Figure 3 The specific implementation of the method shown in this embodiment provides a device for predicting pure electric driving range, such as... Figure 7 As shown, the device includes: an acquisition module 31, an input module 32, and a determination module 33.

[0076] The acquisition module 31 is configured to acquire driving data, vehicle data and driving environment information of the vehicle to be predicted;

[0077] Input module 32 is configured to extract feature data from the driving data, vehicle data, and driving environment information and input it into a preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data; wherein, the feature data is used to characterize the vehicle battery status and vehicle power performance, and the preset machine learning model is trained based on historical driving data samples of the sample vehicle dataset and historical environment data samples corresponding to the historical driving data samples;

[0078] The determination module 33 is configured to determine the remaining pure electric range value output by the preset machine learning model as the predicted pure electric range value of the vehicle to be predicted.

[0079] In specific application scenarios, this device also includes: a model building module;

[0080] The model building module is configured to acquire historical driving data samples, vehicle data samples, and historical environmental data samples of the target model for each trip, and to acquire the historical remaining driving range value corresponding to each trip of the sample vehicle; to build a dataset based on the historical driving data samples, vehicle data samples, and historical environmental data samples of each trip of the sample vehicle, and the historical remaining driving range value corresponding to each trip; and to train the preset machine learning model using the dataset.

[0081] In specific application scenarios, the model building module is specifically configured to extract energy consumption per kilometer of trip, historical average current, historical average vehicle speed, and acceleration from the historical driving data samples; and to extract battery health, maximum battery pack temperature, minimum battery pack temperature, battery state of charge, and battery voltage from the vehicle data samples; and to extract the vehicle's outdoor temperature from the historical environmental data samples. The extracted energy consumption per kilometer of trip, battery state of charge, battery voltage, historical average current, historical average vehicle speed, acceleration, battery health, maximum battery pack temperature, minimum battery pack temperature, and vehicle's outdoor temperature for each trip are used as sample feature data of the dataset, and the historical remaining driving range value corresponding to each trip is used as sample label information corresponding to the sample feature data.

[0082] In specific application scenarios, the input module 32 is specifically configured to extract the energy consumption per kilometer of the current journey of the vehicle to be predicted, the battery state of charge, the battery voltage, the average current, the average vehicle speed, the acceleration, the battery health, the maximum temperature of the battery pack, the minimum temperature of the battery pack, and the outdoor temperature, and input them into a preset machine learning model for calculation.

[0083] In specific application scenarios, the model building module is further configured to use formulas. The remaining pure electric driving range of the sample vehicle at time tm was calculated, wherein the sample vehicle used pure electric mode from time t1 to time tn during each trip. <tm<tn,Label tm S represents the remaining pure electric driving range of the sample vehicle at time tm, and S represents the actual distance traveled by the sample vehicle during the time interval from time tm to time tn. tn Let e ​​be the remaining battery energy of the sample vehicle at time tn, and let e be the average energy consumption per kilometer of the sample vehicle during the time interval from time t1 to time tn.

[0084] In specific application scenarios, the model building module is further configured to use cross-validation to divide the dataset into training and validation sets, and then perform subsequent cross-validation.

[0085] In specific application scenarios, the model building module is further configured to use the Xgboost ensemble learning algorithm to tune the model parameters based on the dataset in order to select the optimal parameter combination; and to complete model training after the optimal parameter combination is determined.

[0086] It should be noted that other corresponding descriptions of the functional units involved in the pure electric driving range prediction device provided in this embodiment can be found in [reference]. Figure 1 and Figure 3 The corresponding descriptions in [the document] will not be repeated here.

[0087] Based on the above, Figure 1 and Figure 3 Accordingly, this embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. Figure 1 and Figure 3 The method shown.

[0088] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0089] Based on the above, Figure 1 and Figure 3 The method shown, and Figure 7 To achieve the above objectives, this application also provides an electronic device, which can be configured on the end side of a vehicle (such as a new energy vehicle), as shown in the virtual device embodiment. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-described... Figure 1 and Figure 3 The method shown.

[0090] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0091] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0092] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0093] Based on the aforementioned electronic device, this application also provides a vehicle, which may specifically include the aforementioned electronic device. This vehicle may specifically be a new energy vehicle, etc.

[0094] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented using software plus necessary general-purpose hardware platforms, or it can be implemented through hardware. By applying the solution of this embodiment, the prediction accuracy of the pure electric range of new energy vehicles can be improved, increasing users' confidence in the prediction results, reducing range anxiety, and enhancing the user's driving experience.

[0095] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0096] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for predicting pure electric driving range, characterized in that, include: Obtain driving data, vehicle data, and driving environment information of the vehicle to be predicted; Feature data is extracted from the driving data, vehicle data, and driving environment information and input into a preset machine learning model for calculation. This model calculates the remaining pure electric mileage corresponding to the same vehicle model as the vehicle to be predicted. The feature data characterizes the vehicle's battery status and power performance. The preset machine learning model is trained using historical driving data samples from a sample vehicle dataset and corresponding historical environment data samples. The labels used in training the preset machine learning model include historical remaining driving range values ​​calculated based on the actual mileage traveled by the sample vehicle during each trip, the remaining battery energy at the end of the trip, and the average energy consumption. The remaining pure electric range value output by the preset machine learning model is determined as the predicted pure electric range value of the vehicle to be predicted.

2. The method according to claim 1, characterized in that, The construction process of the preset machine learning model includes: Obtain historical driving data samples, vehicle data samples, and historical environmental data samples for each trip of the target model's sample vehicle, and obtain the historical remaining driving range value corresponding to each trip of the sample vehicle; A dataset is constructed based on historical driving data samples, vehicle data samples, and historical environmental data samples of each trip of the sample vehicles, as well as the historical remaining driving range value corresponding to each trip; The preset machine learning model is obtained by training the dataset.

3. The method according to claim 2, characterized in that, The dataset is constructed based on historical driving data samples, vehicle data samples, and historical environmental data samples of each trip of the sample vehicles, as well as the historical remaining driving range value corresponding to each trip, including: Extract energy consumption per kilometer of travel, historical average current, historical average vehicle speed, and acceleration from the historical driving data samples; and... From the vehicle data sample, extract battery health, maximum battery pack temperature, minimum battery pack temperature, battery state of charge, and battery voltage; and, Extract the vehicle's outdoor temperature from the historical environmental data sample; The extracted data for each trip, including energy consumption per kilometer, battery state of charge, battery voltage, historical average current, historical average vehicle speed, acceleration, battery health, maximum battery pack temperature, minimum battery pack temperature, and vehicle outdoor temperature, are used as sample feature data for the dataset. The historical remaining driving range for each trip is used as sample label information corresponding to the sample feature data.

4. The method according to claim 2, characterized in that, The step of obtaining the historical remaining driving range value for each trip of the sample vehicle includes: Through formula The remaining pure electric driving range of the sample vehicle at time tm is calculated, wherein the sample vehicle uses pure electric mode from time t1 to time tn during each trip. <tm<tn, This represents the remaining pure electric driving range of the sample vehicle at time tm. The actual distance traveled by the sample vehicle during the time interval from time tm to time tn. Let be the remaining battery energy of the sample vehicle at time tn. The average energy consumption per kilometer of the sample vehicle during the time interval from time t1 to time tn.

5. The method according to claim 2, characterized in that, The step of training the preset machine learning model using the dataset includes: The dataset is divided into training and validation sets using cross-validation, followed by subsequent cross-validation.

6. The method according to claim 2, characterized in that, The process of training the preset machine learning model using the dataset includes: Using the Xgboost ensemble learning algorithm, the model parameters are tuned based on the dataset to select the optimal parameter combination; Model training is completed after the optimal parameter combination is determined.

7. A device for predicting pure electric driving range, characterized in that, include: The acquisition module is configured to acquire driving data, vehicle data, and driving environment information of the vehicle to be predicted; The input module is configured to extract feature data from the driving data, vehicle data, and driving environment information and input it into a preset machine learning model for calculation, so as to obtain the pure electric remaining range value corresponding to the same model of the vehicle to be predicted based on the feature data; wherein, the feature data is used to characterize the vehicle battery status and vehicle power performance, and the preset machine learning model is trained based on historical driving data samples of the sample vehicle dataset and historical environment data samples corresponding to the historical driving data samples; wherein, the labels used for training the preset machine learning model include historical remaining range values ​​calculated based on the actual mileage driven by the sample vehicle in each trip, the remaining battery energy at the end of the trip, and the average energy consumption. The determination module is configured to determine the remaining pure electric range value output by the preset machine learning model as the predicted pure electric range value of the vehicle to be predicted.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

9. An electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

10. A vehicle, characterized in that, include: The electronic device as described in claim 9.

Images