Control method and device for tail gate driving state

Abstract

The invention relates to the technical field of vehicle body domain control, and provides a tail gate driving state control method and device. The method comprises the following steps: acquiring state data of the tail gate driving system; preprocessing the state data; inputting the preprocessed state data into a driving state analysis model to obtain a state probability which is output by the driving state analysis model and corresponds to each driving state, and taking the driving state corresponding to the maximum state probability as a real-time judgment driving state; and controlling the tail gate driving system based on the real-time determined driving state. According to the method, whether the current tail gate driving system is normal or needs to be adjusted can be accurately and reliably reflected by judging the driving state in real time, so that the tail gate driving system is controlled, meanwhile, the safety and accuracy of the tail gate driving system are ensured, and the user experience is improved.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle body domain control, and in particular to a method and device for controlling a tailgate driving state. Background Art

[0002] With the continuous advancement of safety technology, tailgate safety systems have become an integral part of vehicle design. The main purpose of a tailgate safety system is to prevent pinching during tailgate operation. Traditional tailgate safety systems usually compare the real-time status of physical sensors or drive motors with the anti-pinch preset threshold to detect whether there are objects or human parts in the tailgate closing path. Once an obstacle is detected, the system triggers the anti-pinch protection mechanism to stop the tailgate movement or reverse it to avoid injury.

[0003] However, traditional tailgate safety systems often need to make a trade-off between the sensitivity and false alarm rate of anti-pinch detection when designing. A higher sensitivity should be set to improve safety, but a higher sensitivity will also increase the possibility of false triggering of anti-pinch detection, that is, the accuracy of anti-pinch detection will be reduced, which will affect the user experience and may cause users to lose trust in the system; lower sensitivity can reduce the false alarm rate of anti-pinch detection and improve accuracy, but it cannot ensure that anti-pinch detection can be effectively triggered when an obstacle is detected, which will not only cause users to lose trust in the system, but it is also more likely that the tailgate will continue to move due to the inability to effectively trigger anti-pinch detection, causing direct damage to the obstacle, and reducing safety.

[0004] Therefore, how to provide a solution to the above technical problems is a problem that those skilled in the art need to solve. Summary of the invention

[0005] In view of this, an embodiment of the present application provides a method and device for controlling a tailgate driving state to solve the problem in the prior art that the safety and accuracy of anti-pinch of a traditional tailgate safety system cannot be improved simultaneously.

[0006] According to a first aspect of an embodiment of the present application, a method for controlling a tailgate driving state is provided, comprising:

[0007] Acquiring status data of a tailgate drive system; the tailgate drive system includes a drive motor and a tailgate controlled by the drive motor, and the status data includes a movement speed of the tailgate, an acceleration of the movement speed, a motor current of the drive motor, a current change value of the motor current, and an anti-pinch monitoring result signal, wherein the anti-pinch monitoring result signal is used to indicate a real-time anti-pinch monitoring result determined based on at least one of the movement speed, the acceleration, the motor current, and the current change value;

[0008] Preprocess the status data;

[0009] The pre-processed state data is input into the driving state analysis model to obtain the state probability corresponding to each driving state output by the driving state analysis model, and the driving state corresponding to the maximum state probability is used as the real-time driving state; the driving state includes a normal driving state, a correctly triggered anti-pinch state, and an incorrectly triggered anti-pinch state;

[0010] Based on the real-time determination of the driving status, the tailgate driving system is controlled.

[0011] According to a second aspect of an embodiment of the present application, a device for controlling a tailgate driving state is provided, comprising:

[0012] a data acquisition module, for acquiring status data of a tailgate drive system; the tailgate drive system includes a drive motor and a tailgate controlled by the drive motor, the status data includes a movement speed of the tailgate, an acceleration of the movement speed, a motor current of the drive motor, a current change value of the motor current, and an anti-pinch monitoring result signal, the anti-pinch monitoring result signal is used to indicate a real-time anti-pinch monitoring result determined based on at least one of the movement speed, the acceleration, the motor current, and the current change value;

[0013] Data preprocessing portal, used to preprocess the status data;

[0014] The state analysis module is used to input the preprocessed state data into the driving state analysis model, obtain the state probability corresponding to each driving state output by the driving state analysis model, and use the driving state corresponding to the maximum state probability as the real-time driving state; the driving state includes the normal driving state, the correctly triggered anti-pinch state, and the incorrectly triggered anti-pinch state;

[0015] The control module is used to control the tailgate drive system based on real-time determination of the drive state.

[0016] According to a third aspect of an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0017] Compared with the prior art, the beneficial effects of the embodiments of the present application include at least the following: the embodiments of the present application obtain multiple status data based on the drive motor and the tailgate, pre-process the status data and input them into the drive status analysis model to obtain the state probability of each drive state corresponding to the current state data, and use the drive state with the largest state probability as the real-time determination drive state for controlling the tailgate drive system. The drive states with state probabilities include a normal drive state, a correctly triggered anti-pinch state and an incorrectly triggered anti-pinch state. The real-time determination of the drive state can accurately and reliably reflect whether the current tailgate drive system is normal or needs adjustment, thereby controlling the tailgate drive system, while ensuring the safety and accuracy of the tailgate drive system and improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying any creative work.

[0019] Figure 1 It is a schematic diagram of an application scenario of an embodiment of the present application;

[0020] Figure 2 is a flow chart of a method for controlling a tailgate driving state provided in an embodiment of the present application;

[0021] Figure 3 This is a schematic diagram of processing status data provided by an embodiment of the present application;

[0022] Figure 4 is a structural schematic diagram of a control device for a tailgate driving state provided in an embodiment of the present application;

[0023] Figure 5 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0025] A method and device for controlling a tailgate driving state according to an embodiment of the present application will be described in detail below with reference to the accompanying drawings.

[0026] Figure 1 1 is a schematic diagram of an application scenario of an embodiment of the present application. The application scenario may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a server 104 and a network 105.

[0027] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be hardware or software. When the first terminal device 101, the second terminal device 102, and the third terminal device 103 are hardware, they can be various electronic devices with display screens and support communication with the server 104, including but not limited to vehicle systems, smart phones, tablet computers, laptop portable computers, and desktop computers; when the first terminal device 101, the second terminal device 102, and the third terminal device 103 are software, they can be installed in the above electronic devices. The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be implemented as multiple software or software modules, or as a single software or software module, which is not limited in the embodiment of the present application. Further, various applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as data processing applications, instant messaging tools, social platform software, search applications, shopping applications, etc.

[0028] The server 104 may be a server that provides various services, for example, a background server that receives a request sent by a terminal device that establishes a communication connection with the server, and the background server may receive and analyze the request sent by the terminal device, and generate a processing result. The server 104 may be a single server, or a server cluster composed of several servers, or a cloud computing service center, which is not limited in the present embodiment of the application.

[0029] It should be noted that the server 104 can be hardware or software. When the server 104 is hardware, it can be various electronic devices that provide various services for the first terminal device 101, the second terminal device 102, and the third terminal device 103. When the server 104 is software, it can be multiple software or software modules that provide various services for the first terminal device 101, the second terminal device 102, and the third terminal device 103, or it can be a single software or software module that provides various services for the first terminal device 101, the second terminal device 102, and the third terminal device 103, and the embodiment of the present application does not limit this.

[0030] The network 105 can be a wired network connected by coaxial cable, twisted pair and optical fiber, or it can be a wireless network that can interconnect various communication devices without wiring, such as Bluetooth, Near Field Communication (NFC), infrared, etc., which is not limited in the embodiments of the present application.

[0031] It should be noted that the specific types, quantities and combinations of the first terminal device 101, the second terminal device 102, the third terminal device 103, the server 104 and the network 105 can be adjusted according to the actual needs of the application scenario, and the embodiments of the present application do not limit this.

[0032] Figure 2 It is a flow chart of a method for controlling a tailgate driving state provided in an embodiment of the present application. Figure 2 The control method can be Figure 1 The first terminal device, the second terminal device, the third terminal device or the server executes. Figure 2 As shown, the control method includes:

[0033] S201: Acquire status data of a tailgate drive system; the tailgate drive system includes a drive motor and a tailgate controlled by the drive motor, and the status data includes a movement speed of the tailgate, an acceleration of the movement speed, a motor current of the drive motor, a current change value of the motor current, and an anti-pinch monitoring result signal, wherein the anti-pinch monitoring result signal is used to indicate a real-time anti-pinch monitoring result determined based on at least one of the movement speed, the acceleration, the motor current, and the current change value;

[0034] S202: pre-processing the status data;

[0035] S203: inputting the preprocessed state data into the driving state analysis model, obtaining the state probabilities corresponding to each driving state output by the driving state analysis model, and taking the driving state corresponding to the maximum state probability as the real-time driving state; the driving state includes a normal driving state, a correctly triggered anti-pinch state, and an incorrectly triggered anti-pinch state;

[0036] S204: Controlling the tailgate driving system based on the real-time determination of the driving state.

[0037] It can be understood that the method of this embodiment monitors the tailgate drive system, that is, by acquiring the state data of the tailgate drive system, analyzing the real-time determination drive state of the current tailgate drive system, and further controlling the tailgate drive system based on the real-time determination drive state. In the process, the method of this embodiment selects reliable state data, pre-processes the state data, and then inputs the state data into the drive state analysis model, which can derive the state probabilities of the normal drive state, the correctly triggered anti-pinch state, and the wrongly triggered anti-pinch state, wherein the drive state with the largest state probability is used as the real-time determination drive state, reflecting the most likely actual state of the current tailgate drive system, that is, the current tailgate drive system is in the normal drive state, in the anti-pinch state and correctly triggered, in the anti-pinch state and wrongly triggered, and finally uses the real-time determination drive state to achieve accurate control of the tailgate drive system, avoiding both the wrong triggering problem caused by high sensitivity and the unsafe problem caused by low sensitivity, while improving the safety of the tailgate and the user experience.

[0038] Specifically, this embodiment collects the necessary state data of the tailgate drive system in the anti-pinch mechanism, and these state data change over time. The method of this embodiment will continuously obtain these state data and perform subsequent operations. The state data in this embodiment include but are not limited to the movement speed and acceleration of the tailgate, the motor current and current change value of the drive motor, and the anti-pinch monitoring result signal. To ensure the reliability and accuracy of the data, high-precision sensors such as MEMS accelerometers and Hall effect current sensors can be used. Among them, the anti-pinch monitoring result signal is used to indicate the real-time anti-pinch monitoring result determined based on at least one of the movement speed, acceleration, motor current and current change value. The real-time anti-pinch monitoring result is whether there is an obstacle between the current tailgate and the vehicle body. There are two real-time anti-pinch monitoring results corresponding to whether there is an obstacle. The anti-pinch monitoring result signal is represented by two signals of different states. For example, if the real-time anti-pinch monitoring result at a certain moment is that there is an obstacle between the current tailgate and the vehicle body, then the anti-pinch monitoring result signal at that moment is represented by a high level. If there is no obstacle between the current tailgate and the vehicle body, then the anti-pinch monitoring result signal is represented by a low level. In the traditional anti-pinch mechanism, the anti-pinch monitoring result signal is used as the triggering basis for the anti-pinch action. The trigger condition can be set to trigger the anti-pinch action when the high level state of the anti-pinch monitoring result signal remains for a certain window length. The length of the window length in this trigger condition affects the sensitivity of the anti-pinch trigger, making it impossible to take into account both the safety and accuracy of the anti-pinch mechanism.

[0039] Therefore, the method of this embodiment inputs all state data into the driving state analysis model after preprocessing, and further analyzes the state probability of the driving state corresponding to the current state data on the basis of the original triggering conditions. Each state probability indicates the possibility of the driving state occurring under the state data, wherein the normal driving state means that the tailgate driving system is working normally, the normal triggering anti-pinch state means that the tailgate driving system is triggered to take anti-pinch action and the triggering is correct, and the tailgate does need to trigger the anti-pinch action, and the wrong triggering anti-pinch state means that the tailgate driving system is triggered to take anti-pinch action but the triggering is wrong, and the current tailgate does not need to trigger the anti-pinch action. The most likely driving state is used as the basis for the subsequent control of the tailgate driving system, which further improves the accuracy of the control of the tailgate driving system.

[0040] It is understandable that before the state data is input into the driving state analysis model, the state data needs to be preprocessed, which may include data filtering, outlier removal, data normalization, etc., to improve the quality of the state data input into the driving state analysis model. Therefore, the process of preprocessing the state data includes:

[0041] Perform noise filtering on the state data;

[0042] Remove outliers from the state data after noise filtering;

[0043] Normalize the state data after removing outliers.

[0044] Specifically, noise filtering requires targeted selection of filtering methods based on the characteristics of various state data.

[0045] The data of the tailgate's speed is generally continuous and contains short-term fluctuations and noise, which may be caused by sensor accuracy, environmental changes, or mechanical vibrations. Moving average filtering is a simple and effective smoothing technique that smoothes data by calculating the local average of data points. It is suitable for removing short-term fluctuations and noise in the speed of movement.

[0046] Correspondingly, the acceleration data of the motion speed changes more dramatically, especially when the tailgate starts or stops, there will be large fluctuations. Kalman filtering is a recursive filtering algorithm that is suitable for processing noise in dynamic systems. It can update the state estimate at each time step. It is very suitable for real-time processing of acceleration data and can continuously perform state estimation and noise suppression in the data stream.

[0047] Correspondingly, the motor current data contains high-frequency noise, which is caused by the electromagnetic interference of the motor or power supply fluctuations. The current changes are usually relatively stable, but may fluctuate greatly during startup or load changes. Low-pass filtering allows low-frequency signals to pass through and blocks high-frequency noise. It is suitable for removing high-frequency noise from motor current data and retaining useful low-frequency signals.

[0048] Correspondingly, the current change value contains spike noise, which may be caused by instantaneous current fluctuations or sensor failures. Median filtering smoothes the data by taking the median of the data in a local window to remove spike noise. It is particularly effective in removing spike noise while retaining the edge information of the data.

[0049] Correspondingly, the data of the anti-pinch monitoring result signal contains noise caused by sensor delay or system response time fluctuation. Exponential smoothing smoothes the data by assigning exponential decay weights to historical data, which can effectively deal with the noise of the anti-pinch monitoring result signal.

[0050] Therefore, the process of noise filtering the state data may include:

[0051] Perform moving average filtering on the motion speed;

[0052] Perform Kalman filtering on acceleration;

[0053] Low-pass filtering of the motor current;

[0054] Perform median filtering on the current change value;

[0055] Perform exponential smoothing filtering on the anti-pinch monitoring result signal.

[0056] It can be understood that, in this embodiment, when noise filtering is performed on the status data, a specific filtering algorithm can be selected according to the actual performance of the status data and user needs. The above is only an example and is not limited here.

[0057] Furthermore, the process of removing outliers from the state data after noise filtering includes:

[0058] Based on the Z-score analysis method, outliers are removed from each state data after noise filtering.

[0059] It can be understood that the Z-score analysis method is used to identify and remove outliers in a data set. The specific processing steps include:

[0060] Calculate the mean and standard deviation;

[0061] For each data point in the dataset, calculate its Z score.

[0062] Data points whose absolute values ​​of Z scores exceed the preset threshold are removed as outliers.

[0063] The calculation formula of Z score is: Z = (X-μ) / σ. X is the value of the current data point, μ is the mean value of the data set, and σ is the standard deviation of the data set. The preset threshold can be set to 3 or other values, depending on the requirements for the concentration of the data. After removing outliers from the original state data, the remaining state data will be more concentrated and consistent, which helps to improve the accuracy of subsequent analysis.

[0064] It can be understood that in this embodiment, the data normalization of the state data adopts the maximum and minimum normalization method to linearly change and scale the data values ​​of all the state data to the numerical range of 0 to 1, and the original relationship of the numerical values ​​remains unchanged. The specific processing process for each state data includes: determining the maximum value Max and the minimum value Min, and for each data point, determining the value of the data after normalization according to the normalization formula. The normalization formula is: X'=(X-Min) / (Max-Min), X is the value before data normalization, and X' is the value after data normalization.

[0065] It is understandable that the noise filtering, outlier removal and data normalization of the state data can be specifically described as follows: Figure 3 After that, the state data is input into the driving state analysis model to obtain the state probabilities corresponding to the three driving states. The driving state analysis model can adopt the neural network structure of MLP (Multilayer Perceptron). The specific structure and training process of the model can be found in the content of the embodiments below.

[0066] Further, after obtaining the real-time determination of the driving state, the real-time determination of the driving state is used as a control reference to control the tailgate drive system. It is understandable that the operation of the tailgate drive system is subject to many aspects of control and constraints, such as the received user instructions, the monitoring of the triggering conditions of the anti-pinch action, and the real-time determination of the driving state obtained based on the status data of the tailgate drive system. The monitoring of the triggering conditions of the anti-pinch action refers to the tailgate drive system obtaining a trigger determination result that meets the triggering conditions based on the anti-pinch monitoring result signal, and preparing to perform the anti-pinch action. At this time, a triggering anti-pinch event occurs; according to whether the trigger determination result is consistent with the facts, such events will be further divided into a correctly triggered anti-pinch event and an erroneously triggered anti-pinch event, wherein the correctly triggered anti-pinch event corresponds to the tailgate drive system should perform the anti-pinch action, and the erroneously triggered anti-pinch event corresponds to the tailgate drive system should not perform the anti-pinch action, that is, when the erroneously triggered anti-pinch event occurs, that is, when the tailgate drive system is ready to perform the anti-pinch action, if it is considered that the anti-pinch action is triggered by error, the anti-pinch action should not be performed. Therefore, this embodiment actually adds the real-time determination of the driving state to the control factors of the tailgate drive system, and uses it as an important control constraint. The tailgate drive system will be supervised by the real-time determination of the driving state during operation. Once the tailgate drive system triggers an anti-pinch event, but the event is an erroneous triggering of the anti-pinch event, the real-time determination of the driving state at this moment is an erroneous triggering of the anti-pinch state, and the tailgate drive system should not be allowed to perform the anti-pinch action. This prohibition effect is usually achieved by prohibiting the enabling of the anti-pinch function. Specifically, the process of controlling the tailgate drive system based on the real-time determination of the driving state includes:

[0067] When it is determined in real time that the driving state is a normal driving state or the anti-pinch state is correctly triggered, the configuration of the tailgate driving system is set to enable the anti-pinch function;

[0068] When it is determined in real time that the driving state is an erroneously triggered anti-pinch state, the configuration of the tailgate driving system is set to prohibit enabling the anti-pinch function at the current moment.

[0069] It is understandable that when the tailgate drive system is configured to enable the anti-pinch function, the anti-pinch action is allowed to be performed, and when the tailgate drive system is configured to prohibit the anti-pinch function, the anti-pinch action is not allowed to be performed. The method of this embodiment is continuously executed at a specific frequency, and whether the anti-pinch function is enabled will change with the update of the real-time determination drive state. If the real-time determination drive state is updated from the error-triggered anti-pinch state to the normal start state or the normal trigger anti-pinch state, the configuration will be changed to enable the anti-pinch function, thereby allowing the anti-pinch action to be performed when the trigger condition is met.

[0070] The method of the embodiment of the present application obtains multiple status data based on the drive motor and the tailgate, pre-processes the status data and then inputs them into the drive status analysis model to obtain the state probability of each drive state corresponding to the current state data, and uses the drive state with the largest state probability as the real-time determination drive state for controlling the tailgate drive system. The drive states with state probabilities include a normal drive state, a correctly triggered anti-pinch state and an incorrectly triggered anti-pinch state. The real-time determination of the drive state can accurately and reliably reflect whether the current tailgate drive system is normal or needs adjustment, so as to control the tailgate drive system, while ensuring the safety and accuracy of the tailgate drive system and improving user experience.

[0071] When executing the method of this embodiment, the driving state analysis model can adopt the neural network structure of MLP (Multilayer Perceptron), the number of neurons in the input layer corresponds to the type of state data, the number of neurons in the output layer corresponds to the number of driving states, the number of hidden layers and the number of neurons in the hidden layers can be set according to actual conditions or user needs, and the number of neurons in each hidden layer is usually not less than the number of neurons in the input layer and the number of neurons in the output layer. Therefore, the neural network structure of the driving state analysis model includes an input layer, at least one hidden layer and an output layer connected in sequence;

[0072] The input layer includes five input layer neurons, each of which is used to receive a corresponding state data;

[0073] The output layer includes three output layer neurons, and each output layer neuron is used to output the state probability of a corresponding driving state.

[0074] Further, each hidden layer includes a plurality of hidden layer neurons;

[0075] Each hidden layer neuron and each output layer neuron includes a linear transformation unit and a nonlinear transformation unit connected in sequence.

[0076] It is understandable that due to the different positions of neurons, the activation functions of the corresponding nonlinear transformation units are also different. Specifically:

[0077] The nonlinear transformation unit of the hidden layer neurons is used to perform nonlinear transformation based on the ReLU activation function;

[0078] The nonlinear transformation unit of the output layer neurons is used to perform nonlinear transformation based on the softmax activation function.

[0079] In an exemplary embodiment, the neural network structure of the driving state analysis model includes an input layer, two hidden layers and an output layer connected in sequence, wherein:

[0080] The input layer includes five input layer neurons, each of which is used to receive a kind of state data;

[0081] The first hidden layer can be set to 64 hidden layer neurons. A larger number of hidden layer neurons helps to extract high-dimensional features in the state data. These high-dimensional features can better represent the complexity of the input data, thereby improving the model's ability to distinguish different events, especially for error-triggered anti-pinch events. The analysis model needs to identify subtle feature differences, and high-dimensional features can provide richer information. The 64 hidden layer neurons in the first hidden layer all include a linear transformation unit and a nonlinear transformation unit connected in sequence.

[0082] Here, the linear transformation unit maps the data of the five input layer neurons to a 64-dimensional feature space. The calculation of the linear transformation is expressed as:

[0083]

[0084] Among them, z(j) is the result of the linear transformation of the jth hidden layer neuron, w(i,j) is the weight corresponding to the i-th input layer neuron and the j-th hidden layer neuron, x(i) is the data feature of the i-th input layer neuron, b(j) is the bias term of the j-th hidden layer neuron, the value of i is the serial number of the input layer neuron, and the value of j is the serial number of the hidden layer neuron.

[0085] Furthermore, the nonlinear transformation unit performs nonlinear transformation based on the ReLU (Rectified Linear Unit) activation function. The nonlinear transformation can introduce nonlinear features. The ReLU activation function can set all negative values ​​to zero and keep the positive values ​​unchanged. The formula is: a(j)=max(0,z(j)), where a(j) is the result of the nonlinear transformation of the jth hidden layer neuron.

[0086] Similarly, the second hidden layer can be set to 32 hidden layer neurons, and the structure of each hidden layer neuron in the two hidden layers is the same. The 64-dimensional features output by the first hidden layer are passed to the second hidden layer. Each hidden layer neuron in the second hidden layer performs linear and nonlinear transformations on the 64-dimensional features. The nonlinear transformation also uses the ReLU activation function. By reducing the number of neurons, the second hidden layer can compress and combine the high-order features extracted by the first hidden layer, which helps reduce the complexity of the data, prevent overfitting, and retain important feature information.

[0087] Similarly, the output layer neurons receive the output data of the second hidden layer. The linear transformation unit in each output layer neuron performs linear transformation on the output data, and then the nonlinear transformation unit performs nonlinear transformation. Assume that the result of the linear transformation is zo(k), where k represents the serial number of the output layer neuron, and the value range can be 1-3. The nonlinear transformation uses the softmax activation function. The output of the softmax activation function is a probability distribution, which indicates the probability that the output data belongs to this category. The formula is:

[0088]

[0089] σ(k) is the output result of the kth output layer neuron, and n is the neuron index of the output layer.

[0090] It can be understood that through the layer-by-layer feature extraction and combination of the first hidden layer and the second hidden layer, the model forms a hierarchical feature representation, which enables the model to more accurately identify and distinguish different event types, especially the false triggering anti-pinch event. The nonlinearity introduced by the ReLU activation function and the probability distribution output provided by the Softmax activation function further improve the classification ability and confidence of the model. Feature compression and combination reduce the complexity of the model and reduce the risk of overfitting. Combining these mechanisms, the model can significantly improve the prediction efficiency and accuracy of false triggering anti-pinch events.

[0091] It is understandable that the activation function of the hidden layer selects ReLU, which performs well in multi-layer neural networks and can alleviate the gradient vanishing problem, thereby improving the training efficiency and learning ability of the model. The calculation of the ReLU activation function is very simple, and only a threshold operation is required, which is particularly important for large-scale data sets and complex models, and can significantly reduce the training time. The ReLU activation function introduces nonlinearity, which can better capture the complex patterns and features in the input data, thereby improving the prediction accuracy of the false triggering anti-pinch event. The activation function of the output layer uses Softmax, which is suitable for multi-classification problems. The Softmax activation function makes the output of each neuron represent the probability of belonging to the corresponding category, which is particularly important for the multi-classification problem of multiple driving states in this embodiment. The Softmax activation function can amplify the probability of the correct category while compressing the probability of other categories, thereby improving the confidence of the classification, and can more accurately distinguish different categories, especially for the prediction of false triggering anti-pinch events.

[0092] It is understandable that the driving state analysis model needs to be trained with training samples before use until it reaches certain performance indicators before it can be applied in this embodiment. The training process includes:

[0093] Select the loss function and optimizer. The loss function is cross-entropy loss, which is a loss function commonly used in multi-classification problems. It measures the difference between the predicted value and the true value. By minimizing the cross-entropy loss, the model can more accurately predict the probability of each category, helping the model to better distinguish between the three situations of normal operation of the tailgate, correct triggering of anti-pinch, and incorrect triggering of anti-pinch. The cross-entropy loss function provides a smooth and continuous gradient when calculating the gradient, which helps the optimization algorithm find the optimal solution in the parameter space. The smooth gradient makes the optimization process more stable and efficient, reduces the shock and instability during the training process, and thus improves the training efficiency and convergence speed of the model. The optimizer is used to update the network parameters to minimize the loss function. The optimizer uses the SGD (Stochastic Gradient Descent) optimization algorithm, which uses the gradient of each batch to update the model parameters. This is different from the traditional gradient descent algorithm that uses the entire data set to calculate the gradient each time. Updating parameters batch by batch makes SGD more efficient when processing large-scale data sets, reducing memory requirements and computing time. For the prediction of error-triggered anti-pinch events, SGD can iterate and update model parameters faster, improving training efficiency.

[0094] Set hyperparameters during training. The learning rate is one of the most important hyperparameters in the optimization algorithm. If the learning rate is too high, the training may not converge; if it is too low, the training process may be very slow. The learning rate is set to 0.001, which can provide an appropriate update pace while ensuring model convergence, improving training efficiency and stability. The batch size determines the amount of data to be processed before each parameter update. Smaller batches can provide more frequent updates, but the gradient estimate will be unstable; larger batches can provide more stable gradient estimates, but the update frequency is lower. The batch size is set to 64 to strike a balance between the stability of the gradient estimate and the update frequency, ensuring that the model can stably update parameters while maintaining high training efficiency. The number of iterations determines the number of times the entire dataset is used to train the model. Too few iterations may lead to underfitting, while too many iterations may lead to overfitting. The number of iterations is set to 200 to ensure that the model has enough time to learn the data characteristics and avoid underfitting, while preventing overfitting by monitoring the verification performance.

[0095] Train the neural network. Before starting training, split the data into a training set and a validation set. The training set is used to train the model, and the validation set is used to evaluate the performance of the model and adjust hyperparameters. Then set the number of Epochs that the model traverses the entire training set. In this neural network, Epochs is set to 100 to ensure that the model has enough training time to learn data features and avoid underfitting. Underfitting will cause the model to be unable to capture complex patterns in the data, thereby affecting prediction accuracy. Finally, start training the model and train the model using the training data set. After each training, use the validation set to evaluate the performance of the model, which helps monitor whether the model is learning and whether overfitting occurs.

[0096] Model validation. Evaluate the trained model on the test data set. First, ensure that the test data has been preprocessed in the same way as the training data, load the best model saved during training, use the test data set to make predictions for the model, and calculate the performance metrics of the model on the test data.

[0097] The method of the embodiment of the present application obtains multiple status data based on the drive motor and the tailgate, pre-processes the status data and then inputs them into the drive status analysis model to obtain the state probability of each drive state corresponding to the current state data, and uses the drive state with the largest state probability as the real-time determination drive state for controlling the tailgate drive system. The drive states with state probabilities include a normal drive state, a correctly triggered anti-pinch state and an incorrectly triggered anti-pinch state. The real-time determination of the drive state can accurately and reliably reflect whether the current tailgate drive system is normal or needs adjustment, so as to control the tailgate drive system, while ensuring the safety and accuracy of the tailgate drive system and improving user experience.

[0098] All the above optional technical solutions can be combined in any way to form optional embodiments of the present application, and will not be described in detail here. It should be understood that the size of the sequence number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0099] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0100] Figure 4 Schematic diagram of a control device for a tailgate driving state provided in an embodiment of the present application. Figure 4 As shown, the control device includes:

[0101] The data acquisition module 401 is used to acquire the status data of the tailgate drive system; the tailgate drive system includes a drive motor and a tailgate controlled by the drive motor, and the status data includes the movement speed of the tailgate, the acceleration of the movement speed, the motor current of the drive motor, the current change value of the motor current, and the anti-pinch monitoring result signal, and the anti-pinch monitoring result signal is used to indicate the real-time anti-pinch monitoring result determined based on at least one of the movement speed, acceleration, motor current and current change value;

[0102] A data preprocessing module 402, used for preprocessing the state data;

[0103] The state analysis module 403 is used to input the normalized state data into the driving state analysis model, obtain the state probability corresponding to each driving state output by the driving state analysis model, and use the driving state corresponding to the maximum state probability as the real-time driving state; the driving state includes a normal driving state, a correctly triggered anti-pinch state, and an incorrectly triggered anti-pinch state;

[0104] The control module 404 is used to control the tailgate driving system based on the real-time determination of the driving state.

[0105] The device of the embodiment of the present application obtains various status data based on the drive motor and the tailgate, pre-processes the status data and then inputs them into the drive status analysis model to obtain the state probability of each drive state corresponding to the current state data, and uses the drive state with the largest state probability as the real-time determination drive state for controlling the tailgate drive system. The drive states with state probabilities include a normal drive state, a correctly triggered anti-pinch state and an incorrectly triggered anti-pinch state. The real-time determination of the drive state can accurately and reliably reflect whether the current tailgate drive system is normal or needs adjustment, so as to control the tailgate drive system, while ensuring the safety and accuracy of the tailgate drive system and improving user experience.

[0106] In an exemplary embodiment, the data preprocessing module includes:

[0107] A noise filtering unit, used for performing noise filtering on the state data;

[0108] An outlier removal unit is used to remove outliers from the state data after noise filtering;

[0109] The normalization unit is used to normalize the state data after removing outliers.

[0110] In an exemplary embodiment, the noise filtering unit is specifically used for:

[0111] Perform moving average filtering on the motion speed;

[0112] Perform Kalman filtering on acceleration;

[0113] Low-pass filtering of the motor current;

[0114] Perform median filtering on the current change value;

[0115] Perform exponential smoothing filtering on the anti-pinch monitoring result signal.

[0116] In an exemplary embodiment, the outlier removal unit is specifically used for:

[0117] Based on the Z-score analysis method, outliers are removed from each state data after noise filtering.

[0118] In an exemplary embodiment, the neural network structure of the driving state analysis model includes an input layer, at least one hidden layer, and an output layer connected in sequence;

[0119] The input layer includes five input layer neurons, each of which is used to receive a corresponding state data;

[0120] The output layer includes three output layer neurons, and each output layer neuron is used to output the state probability of a corresponding driving state.

[0121] In an exemplary embodiment, each hidden layer includes a plurality of hidden layer neurons;

[0122] Each hidden layer neuron and each output layer neuron includes a linear transformation unit and a nonlinear transformation unit connected in sequence.

[0123] In an exemplary embodiment, the nonlinear transformation unit of the hidden layer neurons is used to perform nonlinear transformation based on the ReLU activation function;

[0124] The nonlinear transformation unit of the output layer neurons is used to perform nonlinear transformation based on the softmax activation function.

[0125] In an exemplary embodiment, the control module is specifically configured to:

[0126] When it is determined in real time that the driving state is a normal driving state or the anti-pinch state is correctly triggered, the configuration of the tailgate driving system is set to enable the anti-pinch function;

[0127] When it is determined in real time that the driving state is an erroneously triggered anti-pinch state, the configuration of the tailgate driving system is set to prohibit enabling the anti-pinch function at the current moment.

[0128] Figure 5 Schematic diagram of an electronic device 5 provided in an embodiment of the present application. Figure 5As shown, the electronic device 5 of this embodiment includes: a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501. When the processor 501 executes the computer program 503, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor 501 executes the computer program 503, the functions of the modules / units in the above-mentioned device embodiments are implemented.

[0129] The electronic device 5 may be a desktop computer, a notebook, a PDA, a cloud server, or other electronic device. The electronic device 5 may include, but is not limited to, a processor 501 and a memory 502. Those skilled in the art will appreciate that Figure 5 The electronic device 5 is merely an example and does not limit the electronic device 5 , and may include more or less components than those shown in the figure, or different components.

[0130] The processor 501 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0131] The memory 502 may be an internal storage unit of the electronic device 5, for example, a hard disk or memory of the electronic device 5. The memory 502 may also be an external storage device of the electronic device 5, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 5. The memory 502 may also include both an internal storage unit of the electronic device 5 and an external storage device. The memory 502 is used to store computer programs and other programs and data required by the electronic device.

[0132] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0133] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. The computer program may include computer program code, which may be in source code form, object code form, executable file or some intermediate form. The computer-readable storage medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0134] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for controlling a tailgate driving state, characterized in that: include: Acquiring status data of a tailgate drive system; the tailgate drive system comprises a drive motor and a tailgate controlled by the drive motor, the status data comprising a movement speed of the tailgate, an acceleration of the movement speed, a motor current of the drive motor, a current change value of the motor current, and an anti-pinch monitoring result signal, the anti-pinch monitoring result signal being used to indicate a real-time anti-pinch monitoring result determined based on at least one of the movement speed, the acceleration, the motor current, and the current change value; Preprocessing the status data; The pre-processed state data is input into a driving state analysis model to obtain the state probabilities corresponding to each driving state output by the driving state analysis model, and the driving state corresponding to the maximum state probability is used as the real-time driving state; the driving state includes a normal driving state, a correctly triggered anti-pinch state, and an incorrectly triggered anti-pinch state; Based on the real-time determination of the driving state, the tailgate driving system is controlled.

2. The method according to claim 1, characterized in that The process of preprocessing the state data includes: performing noise filtering on the state data; removing outliers from the state data after noise filtering; The state data after removing the outliers is normalized.

3. The method according to claim 2, characterized in that The process of performing noise filtering on the state data includes: Performing moving average filtering on the motion speed; performing Kalman filtering on the acceleration; Performing low-pass filtering on the motor current; Performing median filtering on the current change value; The anti-pinch monitoring result signal is subjected to exponential smoothing filtering.

4. The method according to claim 2, characterized in that: The process of removing outliers from the state data after noise filtering includes: Based on the Z-score analysis method, outliers are removed from each state data after noise filtering.

5. The method according to claim 1, characterized in that The neural network structure of the driving state analysis model includes an input layer, at least one hidden layer and an output layer connected in sequence; The input layer includes five input layer neurons, each of which is used to receive a corresponding type of state data; The output layer includes three output layer neurons, and each of the output layer neurons is used to output the state probability of a corresponding driving state.

6. The method according to claim 5, characterized in that Each of the hidden layers includes a plurality of hidden layer neurons; Each of the hidden layer neurons and each of the output layer neurons includes a linear transformation unit and a nonlinear transformation unit connected in sequence.

7. The method according to claim 6, characterized in that The nonlinear transformation unit of the hidden layer neurons is used to perform nonlinear transformation based on a ReLU activation function; The nonlinear transformation unit of the output layer neurons is used to perform nonlinear transformation based on a softmax activation function.

8. The method according to any one of claims 1 to 7, characterized in that The process of controlling the tailgate driving system based on the real-time determination of the driving state includes: When the real-time determination of the driving state is the normal driving state or the correctly triggered anti-pinch state, setting the configuration of the tailgate driving system to enable the anti-pinch function; When the real-time determination driving state is the error-triggered anti-pinch state, the configuration of the tailgate driving system is set to prohibit enabling the anti-pinch function at the current moment.

9. A control device for a tailgate driving state, characterized in that: include: a data acquisition module, for acquiring status data of a tailgate drive system; the tailgate drive system comprises a drive motor and a tailgate controlled by the drive motor, the status data comprises a movement speed of the tailgate, an acceleration of the movement speed, a motor current of the drive motor, a current change value of the motor current, and an anti-pinch monitoring result signal, the anti-pinch monitoring result signal being used to indicate a real-time anti-pinch monitoring result determined based on at least one of the movement speed, the acceleration, the motor current, and the current change value; A data preprocessing module, used for preprocessing the state data; A state analysis module is used to input the preprocessed state data into a driving state analysis model, obtain the state probability corresponding to each driving state output by the driving state analysis model, and use the driving state corresponding to the maximum state probability as the real-time driving state; the driving state includes a normal driving state, a correctly triggered anti-pinch state, and an erroneously triggered anti-pinch state; A control module is used to control the tailgate driving system based on the real-time determination of the driving state.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images