Chassis integrated electric vehicle health monitoring method and system

Abstract

The present application relates to the technical field of health monitoring, and particularly relates to a kind of electric vehicle health monitoring method and system based on chassis integration, method includes: the working parameter information of electric vehicle stationary state is collected, and static diagnosis is carried out, obtains stationary state health monitoring result;Collect the working parameter information of electric vehicle running state, obtain running state parameter;To the feature processing of running state parameter, obtain running state characteristic parameter;Establish health state monitoring model, running state characteristic parameter is regarded as the input of health state monitoring model, obtains running state health monitoring result;Stationary state health monitoring result and running state health monitoring result are synchronized to intelligent application platform, and visual display is carried out.The present application can realize the real-time monitoring and analysis of electric vehicle health, through timely early warning and intelligent management, can maximum degree reduce the failure risk of electric vehicle, guarantee the safety and reliability of electric vehicle.

Description

Technical Field

[0001] This invention relates to the field of health monitoring technology, and in particular to a method and system for health monitoring of electric vehicles based on chassis integration. Background Technology

[0002] As the number of electric vehicles increases and their usage time extends, some health-related issues related to electric vehicles are gradually emerging. These health issues involve various aspects of the vehicle, including battery condition, motor operation, and vehicle body structural stability. These factors directly affect the performance and lifespan of electric vehicles.

[0003] As electric vehicles (EVs) age, issues such as battery aging and improper charging can lead to a decline in battery performance, affecting the EV's range and safety. The motor's operation and wheel rotation status also require regular monitoring to ensure normal operation and driving safety. The stability of the vehicle body structure and the reliability of the braking system are also important considerations for EV health. Therefore, establishing a EV-based health monitoring system has become a crucial area of ​​development to monitor potential health problems during prolonged use. Summary of the Invention

[0004] This invention provides a method and system for health monitoring of electric vehicles based on chassis integration, which can effectively solve the problems in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A chassis-integrated method for monitoring the health of electric vehicles, the method comprising: Collect operating parameter information of electric vehicles in a stationary state, perform static diagnosis, and obtain health monitoring results in a stationary state; Collect the operating parameter information of the electric vehicle's driving status to obtain driving status parameters; The driving state parameters are subjected to feature processing to obtain driving state feature parameters; A health status monitoring model is established, and the driving status characteristic parameters are used as inputs to the health status monitoring model to obtain driving status health monitoring results; The static health monitoring results and the driving health monitoring results are synchronized to the intelligent application platform for visualization.

[0006] Furthermore, the operating parameter information of the electric vehicle in a stationary state is collected, and static diagnosis is performed to obtain the static state health monitoring results, including: Set the sampling frequency to collect the operating parameter information of the electric vehicle in a stationary state; The collected working parameter information is preprocessed to obtain the static state parameters; A threshold range is set, and the static state parameter is judged to be abnormal based on the threshold range to obtain the static state health monitoring result.

[0007] Further, the driving state parameters are subjected to feature processing to obtain driving state feature parameters, including: Define the driving state parameters and divide them into numerical data and discrete data; The numerical data is standardized, and the discrete data is encoded using one-hot encoding. The standardized numerical data and the encoded discrete data are concatenated column by column to construct a feature vector, thus obtaining the driving state feature parameters.

[0008] Furthermore, a health status monitoring model is established, and the driving status characteristic parameters are used as inputs to the health status monitoring model to obtain driving status health monitoring results, including: Collect historical driving status parameters and historical health status information of the electric vehicle, and label the collected historical driving status parameters and historical health status information to establish a historical dataset; The historical dataset is divided into a training set, a validation set, and a test set; A deep neural network is constructed as a health status monitoring model. The health status monitoring model is trained using the training set and the validation set, and evaluated using the test set. The driving state feature parameters are input into the trained health status monitoring model to obtain the driving state health monitoring results.

[0009] Furthermore, it also includes adding a batch normalization layer to the deep neural network.

[0010] Further, the health status monitoring model is trained using the training set and validation set, including: The training set is divided into several batches, and the samples in each batch are input into the deep neural network for forward propagation to calculate the output of the model. The model's output is compared with the true labels, and the loss value is calculated. The gradient of the loss function with respect to the model parameters is calculated using the backpropagation algorithm. Based on the gradient, the model parameters are updated using gradient descent to minimize the loss value; The training set is traversed several times. After each traversal of the training set, the model is evaluated using the validation set. Based on the evaluation results of the validation set, the hyperparameters of the model are adjusted.

[0011] Furthermore, based on the evaluation results of the validation set, the hyperparameters of the model are adjusted, including: Determine the hyperparameters that need to be adjusted, and set a value range for each hyperparameter; Create a mesh by combining the values ​​of each hyperparameter into a mesh, generating several hyperparameter combinations; Traverse the grid, for each hyperparameter combination in the grid, train the model using the training set, evaluate the performance of the model using the validation set, and record the performance metrics for each hyperparameter combination; Based on the recorded performance metrics, select the optimal combination of hyperparameters.

[0012] A chassis-integrated electric vehicle health monitoring system, the system comprising: The static health monitoring module collects the operating parameter information of the electric vehicle in a static state, performs static diagnosis, and obtains the static health monitoring results. The driving status parameter acquisition module collects the working parameter information of the electric vehicle's driving status to obtain driving status parameters; The feature processing module performs feature processing on the driving state parameters to obtain driving state feature parameters; The driving status health monitoring module establishes a health status monitoring model, uses the driving status characteristic parameters as input to the health status monitoring model, and obtains the driving status health monitoring results. The visualization module synchronizes the static health monitoring results and the driving health monitoring results to the intelligent application platform for visualization.

[0013] Furthermore, the static state health monitoring module includes: The parameter acquisition unit is configured to acquire the operating parameter information of the electric vehicle in a stationary state by setting the acquisition frequency. The parameter preprocessing unit preprocesses the collected working parameter information to obtain the static state parameters; The threshold judgment unit sets a threshold range and judges whether the static state parameter is abnormal based on the threshold range, thereby obtaining the static state health monitoring result.

[0014] Furthermore, the feature processing module includes: The parameter partitioning unit defines the driving state parameters and divides the driving state parameters into numerical data and discrete data. The parameter transformation processing unit standardizes the numerical data and performs one-hot encoding on the discrete data. The feature vector construction unit concatenates the standardized numerical data and the encoded discrete data column by column to construct a feature vector, thereby obtaining the driving state feature parameters.

[0015] The technical solution of this invention can achieve the following technical effects: This invention utilizes an intelligent health monitoring system to achieve real-time monitoring and analysis of the health of electric vehicles, promptly identifying and resolving potential problems. Through timely warnings and intelligent management, it can minimize the risk of electric vehicle malfunctions, ensuring the safety and reliability of electric vehicles and extending their service life. Simultaneously, synchronizing the monitoring results to an intelligent application platform enables data visualization and remote management, providing users with a more convenient and intelligent electric vehicle experience.

[0016] 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

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a chassis-integrated electric vehicle health monitoring method. Figure 2 A flowchart illustrating the process of static diagnosis; Figure 3 A flowchart illustrating the process for obtaining driving condition health monitoring results; Figure 4 This is a schematic diagram of a chassis-integrated electric vehicle health monitoring system. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] Example 1 like Figure 1 As shown, a health monitoring method for electric vehicles based on chassis integration is described, and the method includes: S1: Collect the working parameter information of the electric vehicle in a stationary state, perform static diagnosis, and obtain the health monitoring results in a stationary state; Specifically, since the system is free from external interference when stationary, it can accurately reflect the health status of the electric vehicle itself. Therefore, when the electric vehicle is stationary, especially when parked for extended periods, sensors and data acquisition devices can collect relevant parameter information, including battery voltage and temperature. The collected parameters can be transmitted to the data processing unit via CAN bus or wireless communication for preprocessing and static diagnostics. Static diagnostics can effectively detect the health status of the electric vehicle when it is not in motion, such as battery health and motor status, helping to identify potential problems in advance and avoid malfunctions during driving.

[0022] S2: Collect working parameter information of the electric vehicle's driving status to obtain driving status parameters; S3: Perform feature processing on the driving state parameters to obtain driving state feature parameters; This step involves collecting operating parameters of the electric vehicle while it is in motion using sensors and data acquisition devices. These parameters include battery current, voltage, temperature, and motor speed, reflecting the vehicle's performance in actual operation. The collected parameters can be transmitted to the data processing unit via CAN bus or wireless communication for preprocessing and feature processing. Preprocessing of the raw data includes data cleaning, noise reduction, and outlier handling to ensure data quality and accuracy. For the preprocessed data, key features can be extracted using feature processing algorithms, providing a foundation for subsequent analysis.

[0023] S4: Establish a health status monitoring model, use driving status characteristic parameters as input to the health status monitoring model, and obtain driving status health monitoring results; Specifically, machine learning algorithms, including decision trees, random forests, and neural networks, can be used to build predictive models for the health status of electric vehicles. These models can effectively predict and assess the health status of electric vehicles, improving the accuracy and real-time performance of monitoring. During model training, supervised learning can be performed using known health status labels, or unsupervised learning can be attempted using unlabeled data. Then, in the data processing unit, the trained model is used to predict real-time data. By inputting feature parameters into the model for analysis, the operating status of the electric vehicle can be assessed in real time, helping to detect and prevent faults in advance.

[0024] S5: Synchronize the health monitoring results of stationary state with the health monitoring results of driving state to the intelligent application platform for visualization display.

[0025] To achieve real-time monitoring and data management of the health status of electric vehicles, the data processing unit can transmit health monitoring data to the intelligent application platform via wireless communication technologies such as Bluetooth, Wi-Fi, and cellular networks. This not only provides an intuitive reflection of the electric vehicle's health status but also helps reduce data transmission latency and enhance real-time performance. The results of electric vehicle health status monitoring can be visualized, including real-time health monitoring data, historical trend charts, maintenance reminders, and alarm notifications.

[0026] In the above embodiments, various sensors, data processing units, and communication modules are integrated into the chassis of the electric vehicle, enabling the system to collect and process various data in real time during the electric vehicle's stationary and running states, generate health status monitoring results for both stationary and driving states, achieve comprehensive monitoring of the electric vehicle's health status, and transmit the results to smart devices for visualization.

[0027] This invention utilizes an intelligent health monitoring system to achieve real-time monitoring and analysis of the health of electric vehicles, promptly identifying and resolving potential problems. Through timely warnings and intelligent management, it can minimize the risk of electric vehicle malfunctions, ensuring the safety and reliability of electric vehicles and extending their service life. Simultaneously, synchronizing the monitoring results to an intelligent application platform enables data visualization and remote management, providing users with a more convenient and intelligent electric vehicle experience.

[0028] Based on the above embodiments, such as Figure 2 As shown, the operating parameter information of the electric vehicle in a stationary state is collected, and static diagnosis is performed to obtain the health monitoring results in a stationary state, including: S11: Set the acquisition frequency to collect the working parameter information of the electric vehicle in a stationary state; S12: Preprocess the collected working parameter information to obtain the static state parameters; S13: Set a threshold range, determine whether the static state parameters are abnormal based on the threshold range, and obtain the static state health monitoring results.

[0029] Specifically, to ensure the stability and continuity of data acquisition, a parameter acquisition frequency can be set. With the electric vehicle stationary, operating parameter information is collected from various sensors at the set frequency. These stationary parameters are the foundation for subsequent diagnosis and evaluation; therefore, the accuracy and reliability of the data must be ensured. The collected data can be cleaned and smoothed to remove outliers and noise. This preprocessing step improves data quality, removes errors and noise, and makes subsequent analysis more accurate. Then, a threshold judgment algorithm is used to compare the collected data with a set threshold range to determine whether the collected data exceeds the normal range, effectively enabling a health assessment of the electric vehicle in a stationary state.

[0030] Furthermore, feature processing is performed on the driving state parameters to obtain driving state feature parameters, including: S31: Define driving state parameters, and divide driving state parameters into numerical data and discrete data; S32: Standardize numerical data and perform one-hot encoding on discrete data; By dividing driving state parameters into numerical and discrete data, appropriate preprocessing methods can be adopted for different types of data. After standardization, numerical data eliminates the influence of dimensions and makes the data more uniform. For example, the data is converted into a distribution with a mean of 0 and a standard deviation of 1. After one-hot encoding, discrete data can be converted into binary vectors, which avoids unordered comparisons between discrete values ​​and meets the input requirements of most machine learning algorithms.

[0031] S33: Concatenate the standardized numerical data and the encoded discrete data column by column to construct a feature vector and obtain the driving state feature parameters.

[0032] By concatenating the processed numerical and discrete data into a complete feature vector, which is then used as input to a deep neural network model, this method simplifies the data processing process while retaining sufficient feature information, thereby improving the training effect of the model.

[0033] Based on the above embodiments, such as Figure 3 As shown, a health status monitoring model is established, and driving status characteristic parameters are used as inputs to obtain driving status health monitoring results, including: S41: Collect historical driving status parameters and historical health status information of electric vehicles, and label the collected historical driving status parameters and historical health status information to establish a historical dataset; S42: Divide the historical dataset into a training set, a validation set, and a test set; Specifically, historical driving state parameters and health status information of electric vehicles are collected and labeled, forming the foundation for the monitoring model. Labeling the collected data involves associating each data point with a corresponding health status label. This labeling process ensures the supervised learning model can correctly learn the relationship between driving state and health status. To improve model performance, data preprocessing can be performed, including handling missing values, outliers, standardization, or normalization, ensuring data quality and consistency. The historical dataset is partitioned for model selection and tuning during training and for evaluating the model's generalization performance during final evaluation. The historical dataset can be randomly divided into training, validation, and test sets according to a certain ratio.

[0034] S43: Construct a deep neural network as a health status monitoring model, train the health status monitoring model using a training set and a validation set, and evaluate the health status monitoring model using a test set; S44: Input the driving status feature parameters into the trained health status monitoring model to obtain the driving status health monitoring results.

[0035] In this embodiment, multi-layer networks can better learn complex features. To construct a deep neural network structure, the model architecture can be defined first, determining the number of layers and the number of neurons in each layer, including input layers, hidden layers, and output layers. The input layer is responsible for receiving the features of the raw data and passing them to the next layer for processing. The number of nodes in the input layer is equal to the number of features. The hidden layer is used to learn the features and patterns in the data and consists of multiple neurons. Each neuron is connected to all neurons in the previous layer and is calculated using weights and biases. The output layer is responsible for integrating and transforming the features learned by the hidden layer and outputting the final prediction result.

[0036] After defining the model architecture, select the activation function, loss function, and optimizer. The activation function can be ReLU, sigmoid, tanh, etc., applied to each neuron to introduce non-linearity. The loss function can be chosen appropriately based on the task's nature; for example, cross-entropy loss is suitable for classification tasks, while mean squared error is suitable for regression tasks. The optimizer can be Adam, SGD, etc., used to minimize the loss function. Before training the model, compile it to prepare for training. Use a compiler to compile the model architecture, loss function, and optimizer information into executable code. Then, use training, validation, and test sets to train and evaluate the model. When evaluating the model's performance using the test dataset, calculate metrics such as accuracy, precision, and recall on the test set to assess the model's generalization ability. If the model achieves satisfactory performance, use the trained model to predict new data.

[0037] Furthermore, it also includes adding batch normalization layers to deep neural networks.

[0038] Specifically, to avoid drastic changes in the distribution of input features, alleviate the vanishing or exploding gradient problem, and accelerate the model's convergence process, batch normalization layers can be added to the hidden layers of deep neural networks. Batch normalization is a regularization technique that normalizes each feature before the activation function of each hidden layer in a deep neural network, keeping the distribution of input features within a stable range. This helps accelerate model convergence and improves the model's training speed, stability, and generalization ability.

[0039] Furthermore, the health status monitoring model is trained using a training set and a validation set, including: A1: Divide the training set into several batches, and input the samples in each batch into a deep neural network for forward propagation, and calculate the output of the model; Specifically, mini-batch training can balance computational efficiency and memory usage. Therefore, the training dataset can be divided into multiple mini-batches, with each batch being input into the neural network one by one, and the output results calculated.

[0040] A2: Compare the model's output with the true labels and calculate the loss value; Calculating the loss value is a standard method for evaluating model performance. The loss function quantifies the gap between the model's prediction and the true label.

[0041] A3: Calculate the gradient of the loss function with respect to the model parameters using the backpropagation algorithm; Backpropagation is the core algorithm for training deep learning models. It can calculate the gradient of the loss function with respect to each parameter using the chain rule. The gradient represents the rate of change of the loss function with respect to each parameter, and we can deduce from the gradient how to adjust the parameters to minimize the loss value.

[0042] A4: Update the model parameters using gradient descent based on the gradient to minimize the loss value; This step updates the model parameters using gradient descent based on the calculated gradient. The update step size is determined by the learning rate. Gradient descent and its variants are standard methods for optimizing neural network parameters. By adjusting the parameters, the loss value is minimized. Examples include Adam and RMSprop.

[0043] A5: Iterate through the training set several times. After each iteration, evaluate the model using the validation set. Adjust the model's hyperparameters based on the validation set evaluation results.

[0044] To ensure the model fully learns the patterns in the data, the training set can be traversed multiple times. Each complete traversal of the entire dataset is called an epoch. As the number of epochs increases, the model parameters are gradually tuned to a better state, thereby improving the model's performance. At the end of an epoch, all batches of data have been processed, and the model parameters have been updated accordingly. To monitor for overfitting, the model performance can be evaluated using a validation set at the end of an epoch. After training for multiple epochs, the evaluation results on the validation set are accumulated. If the performance on the validation set does not improve significantly, or if overfitting occurs, grid search, random search, or other automatic hyperparameter tuning methods can be used to adjust the model's hyperparameters, including the learning rate and batch size. After finding the optimal combination of hyperparameters, the final model training is performed, ensuring that the model parameters and hyperparameters are optimally tuned, thus improving the model's generalization ability.

[0045] Furthermore, based on the evaluation results of the validation set, the hyperparameters of the model are adjusted, including: B1: Determine the hyperparameters that need to be adjusted and set the value range for each hyperparameter; B2: Create a grid, combining the values ​​of each hyperparameter into a grid to generate several hyperparameter combinations; B3: Traverse the grid, train the model using the training set for each hyperparameter combination in the grid, evaluate the model's performance using the validation set, and record the performance metrics for each hyperparameter combination. B4: Select the optimal combination of hyperparameters based on the recorded performance metrics.

[0046] When building and optimizing a model, some parameters cannot be automatically adjusted from the training data. These parameters are called hyperparameters, including the learning rate, batch size, and number of training epochs. To find the optimal combination of hyperparameters, a grid search method can be used. This method generates all possible hyperparameter combinations by constructing a grid, ensuring that each combination is tried to find the best configuration. Specifically, it iterates through all hyperparameter combinations, training and evaluating the model on each combination, recording performance metrics to facilitate the selection of the optimal combination later. After training and evaluating all hyperparameter combinations, the best-performing hyperparameter combination is selected based on performance metrics, including accuracy on the validation set and loss value.

[0047] Example 2: like Figure 4 As shown, the electric vehicle health monitoring system based on chassis integration includes: The static health monitoring module collects the operating parameter information of the electric vehicle in a static state, performs static diagnosis, and obtains the static health monitoring results. The driving status parameter acquisition module collects the working parameter information of the electric vehicle's driving status and obtains the driving status parameters. The feature processing module performs feature processing on the driving state parameters to obtain driving state feature parameters; The driving status health monitoring module establishes a health status monitoring model, takes driving status characteristic parameters as input to the health status monitoring model, and obtains driving status health monitoring results. The visualization module synchronizes the health monitoring results in both stationary and driving states to the intelligent application platform for visualization.

[0048] The adjustment system described above in this invention can effectively realize the health monitoring method for electric vehicles based on chassis integration, and the technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0049] Furthermore, the resting state health monitoring module includes: The parameter acquisition unit is configured to collect operating parameter information of the electric vehicle when it is stationary, by setting the acquisition frequency. The parameter preprocessing unit preprocesses the collected working parameter information to obtain the static state parameters; The threshold judgment unit sets a threshold range and judges whether the static state parameters are abnormal based on the threshold range, thus obtaining the static state health monitoring results.

[0050] Furthermore, the feature processing module includes: The parameter partitioning unit defines the driving state parameters and divides them into numerical data and discrete data. The parameter transformation and processing unit standardizes numerical data and performs one-hot encoding on discrete data. The feature vector construction unit concatenates the standardized numerical data and the encoded discrete data column by column to construct a feature vector, thereby obtaining the driving state feature parameters.

[0051] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0052] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for health monitoring of electric vehicles based on chassis integration, characterized in that, The method includes: Collect operating parameter information of electric vehicles in a stationary state, perform static diagnosis, and obtain health monitoring results in a stationary state; Collect the operating parameter information of the electric vehicle's driving status to obtain driving status parameters; The driving state parameters are subjected to feature processing to obtain driving state feature parameters; A health status monitoring model is established, and the driving status characteristic parameters are used as inputs to the health status monitoring model to obtain driving status health monitoring results; The static state health monitoring results and the driving state health monitoring results are synchronized to the intelligent application platform for visualization. A health status monitoring model is established, and the driving status characteristic parameters are used as inputs to the health status monitoring model to obtain driving status health monitoring results, including: Collect historical driving status parameters and historical health status information of the electric vehicle, and label the collected historical driving status parameters and historical health status information to establish a historical dataset; The historical dataset is divided into a training set, a validation set, and a test set; A deep neural network is constructed as a health status monitoring model. The health status monitoring model is trained using the training set and the validation set, and evaluated using the test set. The driving state feature parameters are input into the trained health status monitoring model to obtain the driving state health monitoring results; It also includes adding a batch normalization layer to the deep neural network.

2. The method for health monitoring of electric vehicles based on chassis integration according to claim 1, characterized in that, Collect operating parameter information of electric vehicles in a stationary state, perform static diagnosis, and obtain static health monitoring results, including: Set the sampling frequency to collect the operating parameter information of the electric vehicle in a stationary state; The collected working parameter information is preprocessed to obtain the static state parameters; A threshold range is set, and the static state parameter is judged to be abnormal based on the threshold range to obtain the static state health monitoring result.

3. The method for health monitoring of electric vehicles based on chassis integration according to claim 1, characterized in that, The driving state parameters are subjected to feature processing to obtain driving state feature parameters, including: Define the driving state parameters and divide them into numerical data and discrete data; The numerical data is standardized, and the discrete data is encoded using one-hot encoding. The standardized numerical data and the encoded discrete data are concatenated column by column to construct a feature vector, thus obtaining the driving state feature parameters.

4. The method for monitoring the health of electric vehicles based on chassis integration according to claim 1, characterized in that, Training the health status monitoring model using the training and validation sets includes: The training set is divided into several batches, and the samples in each batch are input into the deep neural network for forward propagation to calculate the output of the model. The model's output is compared with the true labels, and the loss value is calculated. The gradient of the loss function with respect to the model parameters is calculated using the backpropagation algorithm. Based on the gradient, the model parameters are updated using gradient descent to minimize the loss value; The training set is traversed several times. After each traversal of the training set, the model is evaluated using the validation set. Based on the evaluation results of the validation set, the hyperparameters of the model are adjusted.

5. The method for health monitoring of electric vehicles based on chassis integration according to claim 4, characterized in that, Based on the evaluation results of the validation set, the hyperparameters of the model are adjusted, including: Determine the hyperparameters that need to be adjusted, and set a value range for each hyperparameter; Create a mesh by combining the values ​​of each hyperparameter into a mesh, generating several hyperparameter combinations; Traverse the grid, for each hyperparameter combination in the grid, train the model using the training set, evaluate the performance of the model using the validation set, and record the performance metrics for each hyperparameter combination; Based on the recorded performance metrics, select the optimal combination of hyperparameters.

6. A chassis-integrated electric vehicle health monitoring system, characterized in that, The system includes: The static health monitoring module collects the operating parameter information of the electric vehicle in a static state, performs static diagnosis, and obtains the static health monitoring results. The driving status parameter acquisition module collects the working parameter information of the electric vehicle's driving status to obtain driving status parameters; The feature processing module performs feature processing on the driving state parameters to obtain driving state feature parameters; The driving status health monitoring module establishes a health status monitoring model, uses the driving status characteristic parameters as input to the health status monitoring model, and obtains the driving status health monitoring results. The visualization module synchronizes the static health monitoring results and the driving health monitoring results to the intelligent application platform for visualization.

7. The electric vehicle health monitoring system based on chassis integration according to claim 6, characterized in that, The static health monitoring module includes: The parameter acquisition unit is configured to acquire the operating parameter information of the electric vehicle in a stationary state by setting the acquisition frequency. The parameter preprocessing unit preprocesses the collected working parameter information to obtain the static state parameters; The threshold judgment unit sets a threshold range and judges whether the static state parameter is abnormal based on the threshold range, thereby obtaining the static state health monitoring result.

8. The electric vehicle health monitoring system based on chassis integration according to claim 6, characterized in that, The feature processing module includes: The parameter partitioning unit defines the driving state parameters and divides the driving state parameters into numerical data and discrete data. The parameter transformation processing unit standardizes the numerical data and performs one-hot encoding on the discrete data. The feature vector construction unit concatenates the standardized numerical data and the encoded discrete data column by column to construct a feature vector, thereby obtaining the driving state feature parameters.

Images