Snowmobile battery life estimation method, electronic device and readable storage medium

Abstract

The present application discloses a method for estimating the battery life of a snowmobile, an electronic device, and a readable storage medium. The method comprises: obtaining relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load, and vehicle acceleration; using a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value; using a second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value; wherein the first power consumption estimation module and the second power consumption estimation module have different estimation methods; combining the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; and using the third power consumption estimation value to calculate the battery life of the snowmobile. Through the above method, the accuracy of the battery life estimation of the snowmobile is improved.

Description

Technical Field

[0001] The present application relates to the field of battery life technology, and in particular to a snowmobile battery life estimation method, an electronic device, and a readable storage medium. Background Art

[0002] The transportation industry is looking at electric vehicles as the future mode of transportation due to the depletion of fossil fuels and the increasing need to overcome global warming caused by vehicle pollution. For this purpose, it is necessary to overcome the barriers facing the mass adoption of electric vehicles. A large number of studies have been conducted to analyze the major factors affecting the market penetration of electric vehicles. The studies found that short driving range, lack of public charging infrastructure, long charging time, and high initial cost are the major factors. Consumers have shown greater interest in purchasing plug-in hybrid vehicles compared to pure electric vehicles. The main reason for this is the short driving range of electric vehicles, which makes drivers worried whether they can reach their destination with the current battery state of charge. Therefore, accurate battery range estimation is a pressing issue that needs to be addressed. Summary of the invention

[0003] The present application provides a snowmobile battery life estimation method, an electronic device, and a readable storage medium, which can improve the accuracy of snowmobile battery life estimation.

[0004] In a first aspect, the present application provides a method for estimating the battery life of a snowmobile, the method comprising: obtaining relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration; using a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value; using a second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value; wherein the first power consumption estimation module and the second power consumption estimation module have different estimation methods; combining the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; and using the third power consumption estimation value to calculate the battery life of the snowmobile.

[0005] Among them, the first power consumption estimation module is used to estimate the power consumption of relevant data to obtain a first power consumption estimation value, including: respectively using the corresponding feature extraction modules in the first power consumption estimation module to extract features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features; connecting the vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features to obtain a first connection feature; connecting the first connection feature, ambient temperature and battery charge state to obtain a second connection feature; and obtaining the first power consumption estimation value according to the second connection feature.

[0006] Among them, each feature extraction module includes a residual branch and three convolution branches; the kernel sizes of the three convolution branches are different; the residual branch and the three convolution branches in the corresponding feature extraction module are used to extract features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features.

[0007] Among them, each convolution branch includes a first convolution kernel activation layer, a first flat pooling layer, a second convolution kernel activation layer, a second flat pooling layer, a third convolution kernel activation layer, a dropout layer and a third flat pooling layer connected in sequence.

[0008] The method of combining the first power consumption estimation value and the second power consumption estimation value to obtain the third power consumption estimation value includes: inputting the first power consumption estimation value and the second power consumption estimation value into the bagged decision tree for adjustment to obtain the third power consumption estimation value.

[0009] Among them, the second power consumption estimation module is used to estimate the power consumption of the relevant data to obtain the second power consumption estimation value, including: using the diagonal forgetting factor recursive least squares method in the second power consumption estimation module to estimate the parameters of the relevant data to obtain the target parameters; and providing the target parameters to the extended Kalman filter in the second power consumption estimation module; using the extended Kalman filter, the power consumption of the relevant data is estimated to obtain the second power consumption estimation value.

[0010] Among them, the forgetting factor in the diagonal forgetting factor recursive least squares method is updated in the following way: generate a diagonal forgetting factor test data set; use the adaptive recursive least squares method to perform parameter estimation to obtain several parameters; determine the target optimal parameter set from the several parameters according to the set time step; in response to the error of the target optimal parameter set being less than the error of the current optimal parameter set, use the target optimal parameter set as the current optimal parameter set.

[0011] Providing the target parameter to the extended Kalman filter in the second power consumption estimation module includes: acquiring an excitation tag corresponding to the target parameter; and providing the target parameter to the extended Kalman filter in the second power consumption estimation module in response to the excitation tag being a preset tag.

[0012] In a second aspect, the present application provides an electronic device, comprising a processor and a memory coupled to the processor, wherein a computer program is stored in the memory, and the processor is used to execute the computer program to implement the method provided in the first aspect.

[0013] In a third aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the method provided in the first aspect.

[0014] The beneficial effects of the present application are as follows: different from the prior art, the snowmobile battery life estimation method, electronic device and readable storage medium provided by the present application utilize a first power consumption estimation module and a second power consumption estimation module of different algorithms to combine vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and at least one of vehicle acceleration to estimate power consumption, and obtain corresponding first power consumption estimation value and second power consumption estimation value; then combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; use the third power consumption estimation value to calculate the battery life of the snowmobile, and be able to utilize a variety of vehicle-related data and a combination of two algorithms to estimate the battery life of the snowmobile, thereby improving the accuracy of the snowmobile battery life estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0016] Figure 1 It is a flow chart of an embodiment of a method for estimating the battery life of a snowmobile provided by the present application;

[0017] Figure 2 It is a flowchart of another embodiment of the snowmobile battery life estimation method provided by the present application;

[0018] Figure 3 It is a flowchart of another embodiment of the snowmobile battery life estimation method provided by the present application;

[0019] Figure 4 It is a structural schematic diagram of an embodiment of an electronic device provided by the present application;

[0020] Figure 5 It is a structural schematic diagram of an embodiment of a computer-readable storage medium provided by the present application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It will be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some but not all structures related to the present application are shown in the drawings. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0022] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0023] The transportation industry is looking at electric vehicles as the future mode of transportation due to the depletion of fossil fuels and the increasing need to overcome global warming caused by vehicle pollution. For this purpose, it is necessary to overcome the barriers facing the mass adoption of electric vehicles. A large number of studies have been conducted to analyze the major factors affecting the market penetration of electric vehicles. The studies found that short driving range, lack of public charging infrastructure, long charging time, and high initial cost are the major factors. Consumers have shown greater interest in purchasing plug-in hybrid vehicles compared to pure electric vehicles. The main reason for this is the short driving range of electric vehicles, which makes drivers worried whether they can reach their destination with the current battery state of charge. Therefore, accurate battery range estimation is a pressing issue that needs to be addressed.

[0024] Based on this, the present application proposes to use a first power consumption estimation module and a second power consumption estimation module of different algorithms to combine at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration to estimate power consumption, and obtain a corresponding first power consumption estimation value and a second power consumption estimation value; then combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; use the third power consumption estimation value to calculate the battery life of the snowmobile, and use a variety of vehicle-related data and a combination of two algorithms to estimate the battery life of the snowmobile, thereby improving the accuracy of the battery life estimation of the snowmobile. For details, please refer to any of the following technical solutions.

[0025] See also Figure 1 , Figure 1 1 is a flow chart of an embodiment of a method for estimating the battery life of a snowmobile provided by the present application. The method comprises:

[0026] Step 11: Obtain relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary load and vehicle acceleration.

[0027] The battery state of charge indicates the current available charge in the battery.

[0028] Vehicle speed can be obtained from the snowmobile itself, vehicle acceleration can be calculated based on vehicle speed, a Geographic Information System (GIS) can be used to obtain elevation data, and a freely available weather API can be used to obtain wind speed and ambient temperature, etc.

[0029] In some embodiments, the relevant data includes vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, and vehicle acceleration.

[0030] In some embodiments, the relevant data includes vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary loads, and vehicle acceleration.

[0031] Step 12: Use the first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value.

[0032] In some embodiments, the first power consumption estimation module may be constructed based on a neural network model, that is, the trained neural network model may be used to perform power consumption estimation on relevant data to obtain a first power consumption estimation value.

[0033] In some embodiments, when the relevant data includes vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary load and vehicle acceleration, each data can be input into the corresponding neural network module separately, and power consumption estimation can be performed separately, and then multiple power consumption estimates can be integrated to obtain a first power consumption estimation value.

[0034] In other embodiments, the sampling module is used to sample the relevant data to obtain relevant sampled data with the same output frequency as the first power consumption estimation value.

[0035] Since the first power consumption estimation value adopts the corresponding calculation, the output first power consumption estimation value also has the corresponding output frequency. Therefore, in order to ensure the subsequent frequency consistency, the relevant data needs to be resampled to obtain the relevant sampling data with the same output frequency as the first power consumption estimation value for subsequent calculation.

[0036] In some embodiments, the vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration may be downsampled respectively to obtain the vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration with the same output frequency as the initial power consumption estimation value. Then, the ambient temperature and battery state of charge may be upsampled respectively to obtain the ambient temperature and battery state of charge with the same output frequency as the first power consumption estimation value.

[0037] For example, the output frequency corresponding to the first power consumption estimate is 1 Hz, that is, one data point per second, so it is necessary to resample the relevant data to match the 1 Hz frequency, and then use the resampled relevant data to adjust the first power consumption estimate. Because the ambient temperature and battery state of charge are considered constant within each 10-second interval. Therefore, vehicle speed, wind speed, altitude data, auxiliary loads, and vehicle acceleration other than ambient temperature and battery state of charge will be downsampled from 10 Hz to 1 Hz, while ambient temperature and battery state of charge are upsampled to 1 Hz by repeating the same value 10 times.

[0038] In some embodiments, the first power consumption estimation value may be adjusted using a corresponding decision tree model.

[0039] Step 13: Use the second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value.

[0040] The first power consumption estimation module and the second power consumption estimation module have different estimation modes.

[0041] In some embodiments, the second power consumption estimation module may use Kalman filtering to perform power consumption estimation on relevant data to obtain a second power consumption estimation value.

[0042] Step 14: Combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value.

[0043] In some embodiments, corresponding weights may be set for the first power consumption estimation value and the second power consumption estimation value. After obtaining the first power consumption estimation value and the second power consumption estimation value, weighted averages are taken to obtain the third power consumption estimation value.

[0044] In some embodiments, if the first power consumption estimation value is adjusted in the above manner, corresponding weights can be set for the adjusted first power consumption estimation value and the second power consumption estimation value. After obtaining the adjusted first power consumption estimation value and the second power consumption estimation value, weighted average is taken to obtain the third power consumption estimation value.

[0045] Step 15: Calculate the battery life of the snowmobile using the third power consumption estimation value.

[0046] In some embodiments, the battery life of a snowmobile may be calculated using the following formula:

[0047]

[0048] Among them, SOC t represents the battery state of charge at time t (corresponding to the battery life of the snowmobile), SOC0 represents the battery state of charge at time 0, and Est pow represents the estimated value of the intermediate power consumption in the time interval 0-t, Ecap Indicates the rated energy capacity of the battery.

[0049] That is, in practical applications, SOC0 represents the battery state of charge at the current moment, and SOC t Indicates the battery state of charge at time t.

[0050] Among them, SOC t The battery state of charge at the tth moment can represent the battery life of the snowmobile.

[0051] In this embodiment, a first power consumption estimation module and a second power consumption estimation module using different algorithms are combined with at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration to obtain corresponding first power consumption estimation value and second power consumption estimation value; then the first power consumption estimation value and the second power consumption estimation value are combined to obtain a third power consumption estimation value; the third power consumption estimation value is used to calculate the battery life of the snowmobile, and the battery life of the snowmobile can be estimated by utilizing a variety of vehicle-related data and combining the two algorithms, thereby improving the accuracy of the snowmobile battery life estimation.

[0052] Furthermore, because snowmobiles are usually used in scenarios where the temperature drops, the solution proposed in this application can more accurately estimate the battery life of the snowmobile in combination with the ambient temperature, which helps to improve the user experience of the snowmobile and the accuracy of battery power testing.

[0053] See also Figure 2 , Figure 2 1 is a flow chart of another embodiment of a method for estimating the battery life of a snowmobile provided by the present application. The method comprises:

[0054] Step 21: Obtain relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary load and vehicle acceleration.

[0055] Step 22: Use the corresponding feature extraction modules in the first power consumption estimation module to extract features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features.

[0056] Among them, each feature extraction module includes a residual branch and three convolution branches; the kernel sizes of the three convolution branches are different; the residual branch and the three convolution branches in the corresponding feature extraction module are used to extract features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features.

[0057] That is, the vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration each have a corresponding feature extraction module.

[0058] Among them, each convolution branch includes a first convolution kernel activation layer, a first flat pooling layer, a second convolution kernel activation layer, a second flat pooling layer, a third convolution kernel activation layer, a dropout layer and a third flat pooling layer connected in sequence.

[0059] Among them, the residual branch includes the first average pooling layer, the second average pooling layer, and the third average pooling layer connected in sequence. The residual branch makes the gradient back propagation better during the training process, so the network is easy to optimize, and finally improves the accuracy of the network. The residual branch acts like an identity function, but three average pooling layers are used instead of providing the same output as the input. The average pooling layer downsamples the data by taking the average value of the input data in a specific area. The pooling layer makes the model locally invariant, that is, the model can extract the same features from the input no matter how the features are rotated, scaled, or shifted. Therefore, the pooling layer not only reduces the size of the network, but also extracts important features from the input. This also helps to avoid overfitting of the network. Each convolution branch has three convolutional layers followed by an average pooling layer, except for the last convolutional layer, which is followed by a dropout layer and then a pooling layer. Convolutional layers are different from traditional fully connected layers because they apply convolutional filters to the input. Convolutional filters extract local features by focusing on specific areas of the input. A convolutional layer can have hundreds of such filters and extract hundreds of features from the input. This helps the convolutional layers learn complex patterns.

[0060] Step 23: Connect the vehicle speed feature, the wind speed feature, the altitude data feature, the auxiliary load feature and the vehicle acceleration feature to obtain a first connection feature.

[0061] Step 24: Connect the first connection characteristic, the ambient temperature and the battery state of charge to obtain a second connection characteristic.

[0062] The ambient temperature and battery state of charge in the above seven data do not change within the 10-second time interval. Therefore, the ambient temperature and battery state of charge are considered constant for each partition, and no feature extraction is performed on the ambient temperature and battery state of charge. Each of the other five data is passed to a separate feature extraction module, and each feature extraction module has four separate branches to extract features. Each feature extraction module has a residual branch and three similar convolution branches.

[0063] Step 25: Obtain a first power consumption estimation value according to the second connection characteristic.

[0064] In a single branch, the convolutional layers differ from each other in the number of kernels used, for example, in each convolutional branch, the first convolutional kernel activation layer has 8 kernels, while the second and third convolutional kernel activation layers have 4 kernels. The convolutional branches differ from each other in kernel size, i.e., one convolutional branch has a kernel of size 3, the second convolutional branch has a kernel of size 5, and the third has a kernel of size 7. Branches with different kernel sizes help the network to extract and learn features based on different temporal resolutions. Each convolutional layer has a padding of n, which depends on the kernel size, such as n = (kernel size-1) / 2. ReLU is used as the activation function for each convolutional layer and can be defined as ReLU(x) = max(0,x). There are two main advantages of activation functions in CNN. First, the activation function converts the output to a scaled range, which helps in training the network quickly. Second, the combination of activation functions helps the network learn more complex nonlinear patterns from the input. Each convolutional branch has a dropout layer as a regularizer. During training, the dropout layer randomly ignores some of the outputs of the previous layer. This makes the training process noisy and forces each layer to jointly adapt to the mistakes made by previous layers, ultimately making the network more robust.

[0065] The output features of the convolutional branch and the residual branch are concatenated at the end to obtain the output features of size 13×12 extracted by the feature extraction module. The features extracted from each feature extraction module are further concatenated and then flattened in consecutive layers. These flattened feature vectors of size 780 are concatenated with two parameters, ambient temperature and battery state of charge, to form a final feature vector of size 782, which is then passed through two consecutive fully connected layers to obtain the final output of the estimated power consumption of size 10 in a 10-second interval, that is, the output of the first power consumption estimate.

[0066] Step 26: Use the diagonal forgetting factor recursive least square method in the second power consumption estimation module to perform parameter estimation on the relevant data to obtain target parameters; and provide the target parameters to the extended Kalman filter in the second power consumption estimation module.

[0067] In some embodiments, the target parameter may be the noise covariance in an extended Kalman filter, such as Q and R in an extended Kalman filter.

[0068] In some embodiments, the forgetting factor in the diagonal forgetting factor recursive least squares method is updated in the following manner: generate a diagonal forgetting factor test data set; use an adaptive recursive least squares method to perform parameter estimation to obtain several parameters; determine a target optimal parameter set from the several parameters according to a set time step; in response to the error of the target optimal parameter set being less than the error of the current optimal parameter set, use the target optimal parameter set as the current optimal parameter set.

[0069] In some embodiments, in the diagonal forgetting factor recursive least squares method, the parameters and error covariance matrix are first initialized, and then the update gain and the estimated error are calculated. The error covariance matrix, the update gain and the estimated error are used to perform parameter estimation on the relevant data to obtain the target parameters.

[0070] Furthermore, the sensitivity of the estimated target parameter is determined by using a condition number detection algorithm, and the target parameter with high sensitivity is provided to the extended Kalman filter in the second power consumption estimation module. The condition number is introduced to numerically represent the accuracy of the estimated target parameter, and automatic adjustment is performed based on the condition number. Therefore, high estimation accuracy can be obtained without fine-tuning the target parameter, and adaptive response can be achieved through automatic adjustment even if the driving mode of the snowmobile changes.

[0071] Step 27: Use an extended Kalman filter to estimate the power consumption of the relevant data to obtain a second power consumption estimation value.

[0072] In some embodiments, step 27 may be the following process: obtaining an excitation tag corresponding to the target parameter; in response to the excitation tag being a preset tag, providing the target parameter to an extended Kalman filter in a second power consumption estimation module. The excitation tag may characterize whether the target parameter is reliable. When the excitation tag is a preset tag, it indicates that the target parameter is reliable, and it may be provided to the extended Kalman filter in the second power consumption estimation module so that the extended Kalman filter uses the updated parameters to perform power consumption estimation on the relevant data and obtain a second power consumption estimation value.

[0073] Step 28: Combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value.

[0074] Step 29: Calculate the battery life of the snowmobile using the third power consumption estimation value.

[0075] In this embodiment, a first power consumption estimation module and a second power consumption estimation module using different algorithms are combined with at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration to obtain corresponding first power consumption estimation value and second power consumption estimation value; then the first power consumption estimation value and the second power consumption estimation value are combined to obtain a third power consumption estimation value; the third power consumption estimation value is used to calculate the battery life of the snowmobile, and the battery life of the snowmobile can be estimated by utilizing a variety of vehicle-related data and combining the two algorithms, thereby improving the accuracy of the snowmobile battery life estimation.

[0076] Furthermore, by using the network structure of the above-mentioned network model for prediction, a more accurate first power consumption estimation value can be obtained, and by combining the diagonal forgetting factor recursive least squares method and the extended Kalman filter, a more accurate second power consumption estimation value can also be obtained, that is, the final battery life of the snowmobile is also more accurate.

[0077] See also Figure 3 , Figure 3 1 is a flow chart of another embodiment of a method for estimating the battery life of a snowmobile provided by the present application. The method comprises:

[0078] Step 31: Obtain relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary load and vehicle acceleration.

[0079] Step 32: Use the first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value.

[0080] Step 33: Use the second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value.

[0081] Steps 31 to 33 have the same or similar technical solutions as any embodiment of the present application and are not described in detail here.

[0082] Step 34: Input the first power consumption estimation value and the second power consumption estimation value into the bagged decision tree for adjustment to obtain a third power consumption estimation value.

[0083] Among them, the bagged decision tree includes multiple sub-decision trees, the first power consumption estimation value and the second power consumption estimation value are input into the bagged decision tree, the first power consumption estimation value and the second power consumption estimation value are adjusted to obtain a third power consumption estimation value, including: inputting the first power consumption estimation value and the second power consumption estimation value into each sub-decision tree, adjusting the first power consumption estimation value and the second power consumption estimation value to obtain several predicted values; averaging the several predicted values ​​to obtain the third power consumption estimation value.

[0084] A bagged decision tree is an ensemble of multiple decision trees using bagging. The main goal of bagging is to reduce the variance of the learners (in this case, the sub-decision trees). This is achieved by averaging and combining the outputs of multiple learners, which ultimately helps in achieving high prediction accuracy. In a bagged decision tree, each learner is a separate decision tree (in this case, the sub-decision tree), and each decision tree is trained separately on a subset of the training dataset. The outputs of all the individual decision trees are averaged to obtain the final prediction (third power estimate). In this work, the number of decision trees and other parameters of the bagged decision tree are optimized by minimizing the cross-validation error, such as by using Bayesian optimization to optimize the number of trees and other parameters. The optimized model contains 10 weak tree learners, which are integrated in the bagged decision tree with the lowest prediction accuracy.

[0085] Step 35: Calculate the battery life of the snowmobile using the third power consumption estimation value.

[0086] In this embodiment, a first power consumption estimation module and a second power consumption estimation module using different algorithms are combined with at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration to obtain corresponding first power consumption estimation value and second power consumption estimation value; then the first power consumption estimation value and the second power consumption estimation value are combined to obtain a third power consumption estimation value; the third power consumption estimation value is used to calculate the battery life of the snowmobile, and the battery life of the snowmobile can be estimated by utilizing a variety of vehicle-related data and combining the two algorithms, thereby improving the accuracy of the snowmobile battery life estimation.

[0087] See also Figure 4 , Figure 4 1 is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. The electronic device 40 includes a processor 41 and a memory 42 coupled to the processor 41. The memory 42 stores a computer program. The processor 41 is used to execute the computer program to implement the following method:

[0088] Obtain relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration; use a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value; use a second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value; wherein the first power consumption estimation module and the second power consumption estimation module have different estimation methods; combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; and use the third power consumption estimation value to calculate the battery life of the snowmobile.

[0089] It can be understood that the processor 41 is also used to execute a computer program to implement the method of any of the above embodiments.

[0090] See also Figure 5 , Figure 5 1 is a schematic diagram of a computer readable storage medium according to an embodiment of the present application. The computer readable storage medium 50 stores a computer program 51. When the computer program 51 is executed by a processor, the following method is implemented:

[0091] Obtain relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and vehicle acceleration; use a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value; use a second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value; wherein the first power consumption estimation module and the second power consumption estimation module have different estimation methods; combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; and use the third power consumption estimation value to calculate the battery life of the snowmobile.

[0092] It can be understood that when the computer program 51 is executed by a processor, it can also implement the method of any of the above embodiments.

[0093] In summary, the snowmobile battery life estimation method, electronic device and readable storage medium provided in the present application use a first power consumption estimation module and a second power consumption estimation module of different algorithms to combine vehicle speed, ambient temperature, wind speed, altitude data, battery charge state, auxiliary load and at least one of vehicle acceleration to estimate power consumption, and obtain corresponding first power consumption estimation value and second power consumption estimation value; then combine the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; use the third power consumption estimation value to calculate the battery life of the snowmobile, and can use a variety of vehicle-related data and a combination of two algorithms to estimate the battery life of the snowmobile, thereby improving the accuracy of the snowmobile battery life estimation.

[0094] Furthermore, the proposed method does not require any vehicle-specific parameters such as the internal resistance of the battery, the efficiency curve of the motor, the discharge / charge curve of the battery, etc. to estimate the energy consumption. Therefore, the proposed method can be easily extended to any other electric vehicle.

[0095] Furthermore, due to the large amount of data, the technical solution of the present application can also learn the nonlinear relationship between different data quite accurately.

[0096] Furthermore, the above mentioned network models can be converted into TensorFlow Lite format and can then be easily run on different microcontrollers or can be deployed to in-vehicle embedded systems.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only illustrative, for example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0098] If the integrated units in the above other embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), disk or optical disk and other media that can store program codes.

[0099] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for estimating the battery life of a snowmobile, characterized in that: The method comprises: Acquiring relevant data of the snowmobile, the relevant data including at least one of vehicle speed, ambient temperature, wind speed, altitude data, battery state of charge, auxiliary load, and vehicle acceleration; Using a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value; Using a second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value; wherein the first power consumption estimation module and the second power consumption estimation module have different estimation modes; Combining the first power consumption estimation value and the second power consumption estimation value to obtain a third power consumption estimation value; calculating the battery life of the snowmobile using the third power consumption estimate; The using a first power consumption estimation module to perform power consumption estimation on the relevant data to obtain a first power consumption estimation value includes: respectively extracting features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration using the corresponding feature extraction modules in the first power consumption estimation module to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features; Connecting the vehicle speed feature, the wind speed feature, the altitude data feature, the auxiliary load feature and the vehicle acceleration feature to obtain a first connection feature; Connecting the first connection characteristic, the ambient temperature and the battery state of charge to obtain a second connection characteristic; Obtaining the first power consumption estimation value according to the second connection characteristic; The using the second power consumption estimation module to perform power consumption estimation on the relevant data to obtain a second power consumption estimation value includes: Using the diagonal forgetting factor recursive least squares method in the second power consumption estimation module to perform parameter estimation on the relevant data to obtain target parameters; and providing the target parameters to the extended Kalman filter in the second power consumption estimation module; Using the extended Kalman filter, performing power consumption estimation on the relevant data to obtain the second power consumption estimation value; The combining the first power consumption estimation value and the second power consumption estimation value to obtain the third power consumption estimation value includes: inputting the first power consumption estimation value and the second power consumption estimation value into a bagged decision tree for adjustment to obtain the third power consumption estimation value.

2. The method according to claim 1, characterized in that Each feature extraction module includes a residual branch and three convolution branches; the kernel sizes of the three convolution branches are different; The residual branch and three convolution branches in the corresponding feature extraction module are used to extract features of vehicle speed, wind speed, altitude data, auxiliary load and vehicle acceleration, respectively, to obtain vehicle speed features, wind speed features, altitude data features, auxiliary load features and vehicle acceleration features.

3. The method according to claim 2, characterized in that Each convolution branch includes a first convolution kernel activation layer, a first average pooling layer, a second convolution kernel activation layer, a second average pooling layer, a third convolution kernel activation layer, a dropout layer, and a third average pooling layer, which are connected sequentially.

4. The method according to claim 1, characterized in that: The forgetting factor in the diagonal forgetting factor recursive least squares method is updated in the following way: Generate a diagonal forgetting factor test data set; Adaptive recursive least squares method is used for parameter estimation to obtain several parameters; Determine a target optimal parameter set from the plurality of parameters according to a set time step; In response to the error of the target optimal parameter set being smaller than the error of the current optimal parameter set, the target optimal parameter set is used as the current optimal parameter set.

5. The method according to claim 1, characterized in that Providing the target parameter to an extended Kalman filter in the second power consumption estimation module includes: Obtaining an incentive label corresponding to the target parameter; In response to the stimulus tag being a preset tag, providing the target parameter to an extended Kalman filter in the second power consumption estimation module.

6. An electronic device, characterized in that: The electronic device comprises a processor and a memory coupled to the processor, wherein a computer program is stored in the memory, and the processor is configured to execute the computer program to implement the method according to any one of claims 1 to 5.

7. 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 processor, the computer program implements the method according to any one of claims 1 to 5.

Images