CN115830368A - Vehicle shock absorber defect diagnosis method, system, equipment and medium - Google Patents

Abstract

The invention belongs to the technical field of defect diagnosis, and particularly discloses a method, a system, equipment and a medium for diagnosing defects of a vehicle shock absorber, which comprise the following steps: collecting various data of the vehicle and preprocessing the data; extracting the characteristics of fault diagnosis through a deep learning framework, and outputting a fault classification result; the deep learning architecture comprises: convolutional neural networks, pooling layers, full connectivity layers, and relus for the SE module were introduced. The method comprises the steps of inputting preprocessed data into a deep learning framework, convolving the data through a convolutional neural network introduced into an SE module to obtain a feature map, and judging which feature maps have the largest contribution to the optimal solution of a classification problem through the SE module; extracting and classifying the features through the pooling layer and the full-link layer, and then performing fault classification diagnosis through the ReLU; the invention can detect the defects of the shock absorber of the chassis of the unmanned vehicle without a plurality of sensors, and can ensure the accuracy of the detection result and improve the safety of the vehicle by using the algorithm of basic data.

Description

A vehicle shock absorber defect diagnosis method, system, equipment and medium

technical field

The invention belongs to the technical field of defect diagnosis, and in particular relates to a defect diagnosis method, system, equipment and medium of a vehicle shock absorber.

Background technique

Monitoring is important for autonomous vehicles. A major challenge in autonomous driving is to shift the responsibility for state detection from the driver to the computer system. Since chassis components have a significant impact on vehicle stability, it is important to monitor them and diagnose any defects. It is also a prerequisite for achieving a high level of automation. The high availability of data in modern cars and current trends in the field of machine learning enable data-based fault detection of chassis components.

Although shock absorber defects account for a large percentage of vehicle defects, there are only a few systems that monitor shock absorbers. For fault detection and isolation systems, no additional sensors are required in the state of the art, but algorithms using underlying data, cost-driven development of mass-produced cars in any case leads to functions increasingly in software Realize that information technology plays an important role in the automotive industry.

Contents of the invention

In order to overcome the shortcomings of the prior art, the object of the present invention is to provide a vehicle shock absorber defect diagnosis method, system, equipment and medium to solve the problem of using multiple sensors for fault detection in the prior art, and for unmanned driving The detection of shock absorbers in vehicles is not comprehensive enough and the detection results are not accurate enough.

In order to achieve the above object, the present invention adopts the following technical solutions to realize:

In a first aspect, the present invention provides a method for diagnosing a vehicle shock absorber defect, comprising:

S1: Collect various data of the vehicle and perform preprocessing;

S2: Input the preprocessed data into the deep learning architecture, and convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; extract and classify the features through the pooling layer and the fully connected layer, and then pass ReLU Carry out fault classification diagnosis and output fault classification results.

Further, the defect behavior of the damper is simulated by changing the valve current before collecting data; test runs with different wheel speeds are carried out on different surfaces; the current supply to a single and all dampers is varied, simulating a single defect and multiple defects .

Further, the various data of the vehicle operation include yaw rate, lateral and longitudinal acceleration, steering angle and wheel speed.

Further, the yaw rate and steering angle are obtained by accessing the bus communication to obtain the data of the anti-lock braking system and the vehicle body electronic stability system, and the yaw rate and the steering angle measured by the yaw rate sensor are collected.

Further, the preprocessing specifically includes: cutting and filtering the collected data, a measurement file is divided into multiple sequences, the sequence length is fixed, the number of data points is the same, and standardized, and the discrete Fourier transform signal processing method is used:

Where X(k) represents the data after DFT transformation, N represents the number of Fourier transform points, k represents the kth frequency spectrum of Fourier transform, x(n) is the sampled analog signal, and x(n) is the complex signal Or a real signal, the imaginary part is 0, and the formula is expanded as:

The short-time Fourier transform is used for processing:

R _ij ＝θ(ε-||X _i -X _j ||)

For the time series signal u _k (k=1,2,...,n), determine its sampling time interval as Δt, determine the embedding dimension m and delay time τ, and then reconstruct the event sequence, and the reconstructed dynamic The system is:

x _i =[u _i , u _i+τ ,..., u _i+(m-1)τ ]

The distance between i point x _i and j point x _j in the reconstructed phase space:

S _ij ＝||X _i -X _j ||

where R _ij is the recursive value, delay coefficient τ, embedding dimension m and threshold ε.

Further, the preprocessed data is input into the deep learning framework, and the data is convoluted by introducing the convolutional neural network of the SE module to obtain a feature map; the features are extracted and classified through the pooling layer and the fully connected layer, Then carry out fault classification diagnosis through ReLU, and output the steps of fault classification results, including:

The processed data is input into the convolutional neural network. First, features are extracted by using convolution filters of different sizes and depth-separable convolutions in parallel, and the SE module is introduced into the convolutional neural network; convolutional filters of different sizes The depth separable convolution is used in several branches, and the depth separable convolution is used in one branch; the depth separable convolution first performs channel-by-channel convolution on the channel depth respectively, and concatenates the output, and then uses the unit convolution kernel to perform Channel convolution to obtain a feature map; input the feature map to the pooling layer to select the features extracted in the convolutional layer, extract graphic features, input to the fully connected layer, and the fully connected layer converts all feature matrices of the pooling layer into a large one-dimensional feature vector; through ReLU for fault classification and diagnosis; finally output the fault classification results.

Further, the fault classification diagnosis includes whether the left front, right front, left rear, and right rear dampers are defective.

In a second aspect, the present invention provides a vehicle shock absorber defect diagnosis system, comprising:

The data collection and preprocessing module is used to collect various data of vehicle operation and perform preprocessing;

The feature extraction and fault classification output module is used to input the preprocessed data to the deep learning framework, and to convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; through the pooling layer and the fully connected layer to The features are extracted and classified, and then the fault classification and diagnosis are carried out through ReLU, and the fault classification results are output.

In a third aspect, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and operable on the processor, wherein the processor executes the computer program At the same time, the vehicle shock absorber defect diagnosis method described in any one of the above is realized.

In a fourth aspect, the present invention provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and it is characterized in that, when the computer program is executed by a processor, one of the above-mentioned ones is implemented. A vehicle shock absorber defect diagnosis method.

The present invention has at least the following beneficial effects:

In the present invention, the preprocessed data is input into the deep learning framework, and the data is convolved by introducing the convolutional neural network of the SE module to obtain the feature map, and the SE module can judge which feature map contributes the most to the optimal solution of the classification problem ; The feature is extracted and classified through the pooling layer and the fully connected layer, and then the fault classification and diagnosis is performed through ReLU; the present invention can detect the defect of the shock absorber of the unmanned vehicle chassis, and it can be carried out without multiple sensors. Algorithms using basic data can ensure the accuracy of detection results and improve vehicle safety.

Description of drawings

The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention, and the schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations to the present invention. In the attached picture:

Figure 1 is a schematic diagram of convolution;

Fig. 2 is a block diagram of a vehicle shock absorber defect diagnosis system.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and examples. It should be noted that, in the case of no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

The following detailed descriptions are all exemplary descriptions, and are intended to provide further detailed descriptions of the present invention. Unless otherwise specified, all technical terms used in the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention belongs. Terms used in the present invention are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention.

Example 1

A method for diagnosing defects of a vehicle shock absorber, comprising:

S1: Collect various data of vehicle operation and perform preprocessing;

Various data of vehicle operation include yaw rate, lateral and longitudinal acceleration, steering angle and wheel speed.

The data of the anti-lock braking system (ABS) and the electronic stability system (ESP) of the body are obtained by accessing the bus communication, and the yaw rate and steering angle measured by the yaw rate sensor are collected.

The adjustment of the damper characteristics is carried out through the current of the damper valve, and by changing the valve current, the defective behavior of the damper is simulated. Test runs with different wheel speeds were carried out on different surfaces. The current supply to individual and all dampers was varied to simulate single and multiple defects.

Defect diagnosis is performed using controller area network signals (CAN) for all wheel speeds, and calculated longitudinal and lateral accelerations at the vehicle's center of gravity. All signals are sampled at the same frequency. Due to the particularity of bus load fluctuations in CAN communication, the signals cannot be provided at exactly the same time, so sampling rate conversion (resampling) is required.

The collected data needs to be pre-processed before passing into the network. In the preprocessing part, the collected data is cut and filtered, and a measurement file is divided into multiple sequences with a fixed length, that is, the same number of data points; and standardized to prevent some inputs from being too heavy or too light. Time-based data tend to contain high noise components and have high dimensions. To deal with these problems, discrete Fourier transform (DFT) signal processing methods can be used:

Where X(k) represents the data after DFT transformation, N represents the number of Fourier transform points, k represents the kth frequency spectrum of Fourier transform, x(n) is the sampled analog signal, x(n) in the formula It can be a complex signal. In practice, x(n) is a real signal, that is, the imaginary part is 0. At this time, the formula can be expanded as:

The Fast Fourier Transform (FFT) is implemented using the Discrete Fourier Transform. The discrete Fourier transform can transform the signal from the time domain to the frequency domain, and both the time domain and the frequency domain are discrete. Fast Fourier transform is a fast algorithm of discrete Fourier transform, which can transform a signal into the frequency domain. It is difficult to see the characteristics of some signals in the time domain, but if they are transformed into the frequency domain, the characteristics are easy to see. This is why many signal analyzes employ the Fast Fourier Transform transformation. In addition, the fast Fourier transform can extract the spectrum of a signal, which is often used in spectrum analysis. The data samples transformed with the fast Fourier transform lose the change information in the time domain. To overcome this problem, the short-time Fourier transform (STFT) can be used. In the short-time Fourier transform, a time-shifted analysis window is used to extract local frequencies by Fourier transform. The Gramian angle field is a two-dimensional representation of a time signal. A time signal is scaled to a fixed interval and then converted to polar coordinates. Values are expressed as angle cosines and time as radius. Recursion graph is an important method to analyze the periodicity, chaos and non-stationarity of time series. It can reveal the internal structure of time series and give prior knowledge about similarity, information and predictability. The recurrence graph is especially suitable for short time series data, and can test the stationarity and internal similarity of the time series. The short-time Fourier transform is used for processing, which is expressed by the following formula:

R _ij ＝θ(ε-||X _i -X _j ||)

For the time series signal u _k (k=1,2,...,n), determine its sampling time interval as Δt, determine the appropriate embedding dimension m and delay time τ through relevant theoretical calculations, and then reconstruct the event sequence , the reconstructed dynamical system is:

x _i =[u _i , u _i+τ ,..., u _i+(m-1)τ ]

The distance between i point x _i and j point x _j in the reconstructed phase space:

S _ij ＝||X _i -X _j ||

where R _ij is the recursive value, delay coefficient τ, embedding dimension m and threshold ε. The commonly used embedding dimension selection method is the pseudo-neighborhood method, and the delay coefficient selection is the average mutual information method. There is currently no better method for the optimal recursion threshold, and generally 10% of the peak value is selected. The results were evaluated using a confusion matrix.

The deep learning architecture is shown in Table 1.

Table 1 Deep Learning Architecture

S2: Extract the features of the fault diagnosis through the deep learning architecture, and output the fault classification results; the deep learning architecture includes: the convolutional neural network introduced with the SE module, the pooling layer, the fully connected layer and the ReLU. Input the preprocessed data into the deep learning architecture, and convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; extract and classify the features through the pooling layer and the fully connected layer, and then perform fault detection through the ReLU Classification diagnosis, output fault classification results.

In order to cover the impact of different filter sizes, the processed data is input into the convolutional neural network, and the features are first extracted by using different sizes of convolutional filters and depthwise separable convolutions. The SE module is introduced into the convolutional neural network. In the neural network; the SE module is used to judge which feature maps contribute the most to the optimal solution of the classification problem; convolution filters of different sizes are used in several branches, and depthwise separable convolutions are used in one branch, such as Figure 1 shows. Depth separable convolution first performs channel-by-channel convolution (depthwise convolution, DC) on the channel (depth), and concatenates the output, and then uses the unit convolution kernel to perform channel convolution (pointwise convolution, PC) to obtain the feature map . Input the feature map into the pooling layer to select the features extracted in the convolutional layer, extract the graphic features, and then input it to the fully connected layer. The fully connected layer converts all the feature matrices of the pooling layer into a one-dimensional large feature vector. The fully-connected layer is generally placed at the end of the convolutional neural network structure for classification; then fault classification and diagnosis is performed through ReLU (modified linear unit), ReLU is an activation function, generally used in classification problems; finally output fault classification results ; Fault classification diagnosis includes: whether the left front, right front, left rear, right rear four dampers are defective.

Compared with conventional convolution operations, the number of parameters and computational cost are relatively low. One convolution kernel of channel-by-channel convolution is responsible for one channel, and one channel is only convolved by one convolution kernel. The number of feature map channels generated by this process is exactly the same as the number of input channels. The point-by-point convolution operation will weight the feature map of the previous step in the depth direction to generate a new feature map. This results in different feature maps for each branch, which on the one hand can be produced by ordinary convolution operations with filter masks of different sizes, and on the other hand can be produced by extracting features. By extracting features from each data channel individually and then combining them. and their subsequent combinations.

The SE module realizes the scaling of a single feature map through the Network-in-Network (NIN) method. An internal multi-layer perceptron (MLP) allows the network to autonomously learn which feature maps contribute most to the optimal solution of the classification problem. Therefore, SE blocks are additionally introduced into the network. The network with SE blocks is followed by a max pooling layer. Subsampling layers improve performance while reducing the number of parameters. Dropout is used to reduce overfitting.

Example 2

A vehicle shock absorber defect diagnosis system, comprising:

The data collection and preprocessing module is used to collect various data of vehicle operation and perform preprocessing;

The feature extraction and fault classification output module is used to input the preprocessed data to the deep learning framework, and to convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; through the pooling layer and the fully connected layer to The features are extracted and classified, and then the fault classification and diagnosis are carried out through ReLU, and the fault classification results are output.

Example 3

The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and operable on the processor, wherein the processor implements the embodiment when executing the computer program A vehicle shock absorber defect diagnosis method described in 1.

Example 4

The present invention provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and it is characterized in that, when the computer program is executed by a processor, the defect of a vehicle shock absorber described in Embodiment 1 is realized diagnosis method.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: the present invention can still be Any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention shall fall within the protection scope of the claims of the present invention.

Claims (10)

1. A vehicle shock absorber defect diagnosis method, characterized in that, comprising: S1: Collect various data of vehicle operation and perform preprocessing; S2: Input the preprocessed data into the deep learning architecture, and convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; extract and classify the features through the pooling layer and the fully connected layer, and then pass ReLU Carry out fault classification diagnosis and output fault classification results. 2. A method for diagnosing defects of a vehicle shock absorber according to claim 1, characterized in that, before the data is collected, the defect behavior of the damper is simulated by changing the valve current; the test runs of different wheel speeds are carried out on different surfaces ; Current supply to individual and all dampers is varied, simulating single and multiple defects. 3. A method for diagnosing vehicle shock absorber defects according to claim 1, wherein the various data of the vehicle operation include yaw rate, transverse and longitudinal acceleration, steering angle and wheel speed. 4. A kind of vehicle shock absorber defect diagnosis method according to claim 3, is characterized in that, described yaw rate and steering angle obtain the data of anti-lock braking system and vehicle body electronic stability system by accessing bus communication, Collect the yaw rate and steering angle measured by the yaw rate sensor. 5. A method for diagnosing vehicle shock absorber defects according to claim 1, wherein the preprocessing specifically includes: cutting and filtering the collected data, one measurement file is divided into multiple sequences, and the sequence length Fixed, the number of data points is the same, standardized, using the discrete Fourier transform signal processing method:

Where X(k) represents the data after DFT transformation, N represents the number of Fourier transform points, k represents the kth frequency spectrum of Fourier transform, x(n) is the sampled analog signal, and x(n) is the complex signal Or a real signal, the imaginary part is 0, and the formula is expanded as:

The short-time Fourier transform is used for processing: R _ij ＝θ(ε-||X _i -X _j ||) For the time series signal u _k (k=1,2,...,n), determine its sampling time interval as Δt, determine the embedding dimension m and delay time τ, reconstruct the event sequence, and the reconstructed dynamical system for: x _i ＝[u _i ,u _i+τ ,...,u _i+(m-1)τ ] The distance between i point x _i and j point x _j in the reconstructed phase space: S _ij ＝||X _i -X _j || where R _ij is the recursive value, delay coefficient τ, embedding dimension m and threshold ε. 6. A method for diagnosing defects of a vehicle shock absorber according to claim 1, wherein the preprocessed data is input to the deep learning framework, and the data is convoluted by introducing a convolutional neural network of the SE module. The feature map is obtained; the feature is extracted and classified through the pooling layer and the fully connected layer, and then the fault classification diagnosis is performed through the ReLU, and the steps of outputting the fault classification result include: The processed data is input into the convolutional neural network. First, features are extracted by using convolution filters of different sizes and depth-separable convolutions in parallel, and the SE module is introduced into the convolutional neural network; convolutional filters of different sizes The depth separable convolution is used in several branches, and the depth separable convolution is used in one branch; the depth separable convolution first performs channel-by-channel convolution on the channel depth respectively, and concatenates the output, and then uses the unit convolution kernel to perform Channel convolution to obtain a feature map; input the feature map to the pooling layer to select the features extracted in the convolutional layer, extract graphic features, input to the fully connected layer, and the fully connected layer converts all feature matrices of the pooling layer into a large one-dimensional feature vector; through ReLU for fault classification and diagnosis; finally output the fault classification results. 7 . The method for diagnosing defects of a vehicle shock absorber according to claim 6 , wherein the fault classification diagnosis includes whether the left front, right front, left rear, and right rear dampers are defective. 8. A vehicle shock absorber defect diagnosis system, characterized in that it comprises: The data collection and preprocessing module is used to collect various data of vehicle operation and perform preprocessing; The feature extraction and fault classification output module is used to input the preprocessed data to the deep learning framework, and to convolve the data by introducing the convolutional neural network of the SE module to obtain the feature map; through the pooling layer and the fully connected layer to The features are extracted and classified, and then the fault classification and diagnosis are carried out through ReLU, and the fault classification results are output. 9. A computer device comprising a memory, a processor, and a computer program stored in the memory and operable on the processor, wherein claim 1 is realized when the processor executes the computer program - A method for diagnosing a defect in a vehicle shock absorber according to any one of 7. 10. A computer-readable storage medium, the computer-readable storage medium stores a computer program, characterized in that, when the computer program is executed by a processor, the computer program according to any one of claims 1-7 is implemented. Vehicle shock absorber defect diagnosis method.

Images