Road surface grade identification method and suspension system embedded with friction nanometer generator

Abstract

The invention discloses a road surface grade identification method and a suspension system embedded with a friction nanometer generator. The method comprises the following steps: acquiring a friction electric signal; the friction electric signal is output by a friction nanometer generator synchronously acting with the suspension system; adopting wavelet transform to convert the friction electric signal into image data; constructing a pavement grade identification model based on a deep learning network; according to the method, the friction nano-generator is used as a self-powered sensor, and the friction electric signal output by the friction nano-generator is input into the convolutional neural network after being subjected to wavelet transformation to perform pavement grade identification; the problem that a suspension stroke sensor in the prior art needs to be additionally provided with an energy supply device such as a power supply is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent vehicles, and in particular to a road surface grade identification method and a suspension system embedded with a friction nanogenerator. Background Art

[0002] To ensure driver stability and passenger comfort, active suspension is gaining increasing attention as a means of improving suspension system performance and becoming a key component of intelligent vehicles. Compared to passive suspension, active suspension proactively generates a force to counteract inertial forces when the suspension system experiences spring deformation due to factors such as road undulations, thereby reducing vehicle body position changes. To achieve this, it is often necessary to accurately and promptly identify road surface irregularities and convert them into appropriate parameters for input into the control system to achieve active suspension control. Furthermore, road surface irregularities are crucial for vehicle safety. Poor road conditions can accelerate the damage and aging of vehicle components, impair the driver's perception of the road environment, and pose a safety hazard. Therefore, accurate identification of road surface irregularities is crucial for optimizing suspension control strategies and improving vehicle performance and safety.

[0003] Among the methods known to the inventors, common methods for identifying road surface roughness include direct measurement methods and response-based identification methods. Direct measurement methods are further divided into traditional measurement methods and methods based on vehicle-mounted instruments. The former mainly uses the three-meter ruler method and the level method, which are relatively complex to operate and more subjective; the latter mainly uses a laser leveler, which has fast detection and good reproducibility, but is expensive. The method based on vehicle dynamic response uses sensors installed on the vehicle to identify road surface roughness by inverse analysis of the vehicle dynamic response. It is more economical and convenient and has broad application prospects.

[0004] Among the methods known to the inventors, existing solutions primarily identify road surface grade based on suspension travel. This approach first uses a suspension travel sensor to measure vehicle suspension travel. Wavelet noise reduction is then used to preprocess the travel signal. An estimated road roughness coefficient is calculated from the preprocessed signal and a feature matrix of the estimated value is constructed. Finally, a judgment matrix for different road surface grades is constructed. The road surface grade at a given distance is identified by comparing the similarity between the estimated feature matrix and the judgment matrix for different road surface grades.

[0005] In order to realize the recognition of road surface unevenness, among the methods known to the inventors, all of them use sensors without exception, either directly identifying road surface elevation information, or sensing and monitoring one or more dynamic parameters of the vehicle in combination with dynamic response. Most of the sensors used in these technologies are conventional sensors, such as lidar, displacement sensors, acceleration sensors, etc., which usually need to be installed on the vehicle body or suspension during normal operation, and require additional power supply devices for energy supply. Therefore, how to reasonably design embedded intelligent sensors that do not require external power supply devices and realize road surface grade recognition on vehicles where space resources are relatively precious has become an unavoidable problem.

[0006] Vehicles inevitably encounter poor road conditions during driving, and after extended periods of continuous driving, components heat up and vibrate more intensely, presenting the sensors installed in the vehicle with the challenge of operating in such a harsh environment. This means that the widespread adoption of road roughness recognition technology requires ensuring the sensor's operating life and reliability in extremely harsh conditions. Furthermore, sensor cost must be taken into consideration. Maintaining high accuracy while reducing operating and maintenance costs is crucial to promoting the widespread adoption of road roughness recognition technology. These are all issues that existing technologies fail to address.

[0007] To the best of the inventor's knowledge, existing research focuses primarily on analyzing and processing collected signals and utilizing road surface grade identification results to design vehicle suspension control systems. However, the raw collected signals often cannot be directly used as input parameters for road surface grade identification. Furthermore, the collected signals typically undergo a series of preprocessing steps before road surface grade identification is performed, which complicates the entire identification process and increases the cost and barrier to entry for practical use. Summary of the Invention

[0008] The purpose of the present invention is to provide a road surface grade identification method and a suspension system embedded with a friction nanogenerator. The friction nanogenerator is used as a self-powered sensor, and its output triboelectric signal is input into a convolutional neural network after wavelet transformation for road surface grade identification, so as to solve the problem that the suspension dynamic stroke sensor used in the prior art must be equipped with additional power supply devices such as a power supply.

[0009] To achieve the above object, the present invention provides the following solution: The present invention provides a road surface grade identification method, the road surface grade identification method comprising: Acquiring a triboelectric signal; the triboelectric signal is output by a triboelectric nanogenerator that operates synchronously with the suspension system; Using wavelet transform to convert the triboelectric signal into image data; Construct a road surface grade recognition model based on deep learning network; The image data is input into the road surface grade recognition model to obtain a road surface grade recognition result.

[0010] Preferably, a road surface grade recognition model is constructed based on a deep learning network, including: Obtain deep learning networks and datasets; The data set is divided into a training set and a test set; the training set and the test set each include a plurality of training sample pairs; each training sample pair consists of historical image sample data and a sample data label; the historical image sample data is obtained by converting the historical triboelectric signal output by the triboelectric nanogenerator; The deep learning network is trained with the historical image sample data in the training set as input and the sample data label corresponding to the historical image sample data as output, and the trained deep learning network is tested with the samples in the test set until the output error of the trained deep learning network reaches a set threshold, and the trained deep learning network is used as a road surface grade recognition model.

[0011] Preferably, the deep learning network is a convolutional neural network.

[0012] Also provided is a suspension system embedded with a friction nanogenerator, comprising an upper shock absorber seat, a lower shock absorber seat, a shock absorber installed between the upper shock absorber seat and the lower shock absorber seat, and a friction nanogenerator matched with the shock absorber; The triboelectric nanogenerator includes a generator body distributed in parallel with the shock absorber, the generator body includes a first connecting portion fixed at a position above the shock absorber and a second connecting portion fixed at a position below the shock absorber, and the first connecting portion and the second connecting portion move relative to each other along the shock absorption direction of the shock absorber; The first connecting portion is provided with an interdigital electrode that moves synchronously therewith, and the interdigital electrode includes a main portion and a contact portion for outputting a triboelectric signal to the outside. The second connecting portion is provided with a friction portion that is in close contact with the main portion, and the friction portion moves synchronously with the second connecting portion and rubs against the main portion. The first connecting portion of the generator body is fixedly connected to the upper seat of the shock absorber, and the second connecting portion of the generator body is fixedly connected to the lower seat of the shock absorber.

[0013] Preferably, the friction portion includes a plurality of bosses arranged in sequence and at intervals along the movement direction of the second connecting portion, and each of the bosses moves synchronously with the second connecting portion and rubs against the main body.

[0014] Preferably, the surface of the boss in friction with the main body is coated with polytetrafluoroethylene.

[0015] Preferably, the width of each of the bosses is equal to the width of a single electrode on the interdigitated electrodes, and the width of the groove between two adjacent bosses is equal to the interval between the electrodes.

[0016] Preferably, the first connecting part is an insulating hollow cylindrical sleeve, the axis of the first connecting part extends along the shock absorbing direction of the shock absorber, and the top of the first connecting part is fixed at a position above the shock absorber, and the bottom is provided with an opening for the second connecting part to extend into, and the interdigitated electrode is arranged on the inner wall of the first connecting part.

[0017] Preferably, the interdigital electrodes include an interdigital electrode substrate made of a flexible circuit board, and the substrate is tightly fitted to the inner wall of the first connecting portion.

[0018] Compared with the prior art, the present invention has achieved the following technical effects: This application uses a friction nanogenerator as a self-powered sensor, and its output triboelectric signal is input into a convolutional neural network after wavelet transformation for road surface grade identification, so as to solve the problem that the suspension dynamic stroke sensor used in the prior art must be equipped with additional power supply devices such as a power supply. And the collected triboelectric signal data is pre-processed using wavelet transformation to solve the problem of cumbersome preprocessing links. In addition, the wavelet transform can effectively extract the time-frequency information in the triboelectric signal, and the transformed data can be directly used as the input of the convolutional neural network, without the need for additional preprocessing operations on the triboelectric signal, which simplifies the operation of the preprocessing link. Further, a deep learning network is used to identify the road surface grade to solve the problem of complex recognition process. The deep learning network is simple to build, and has strong feature extraction and generalization capabilities. It can greatly simplify the recognition process while ensuring high recognition accuracy, effectively reducing the cost and threshold of recognition process maintenance and iteration, and facilitating secondary development and expansion and extension. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A flow chart of an implementation method of a road surface grade identification method provided by one embodiment of the present invention; Figure 2 A diagram showing a theoretical model of a linear suspension system according to an embodiment of the present invention; Figure 3 Schematic diagram of the interdigitated electrode structure in one embodiment of the present invention; Figure 4This is a schematic structural diagram of the friction portion in one embodiment of the present invention; Figure 5 The basic structure of a suspension system embedded with a triboelectric nanogenerator in one embodiment of the present invention; Figure 6 for Figure 5 A magnified view of the friction nanogenerator in the middle; Figure 7 A schematic diagram of a time-frequency image obtained by wavelet transformation according to an embodiment of the present invention; Figure 8 A schematic diagram of the structure of a convolutional neural network provided by one embodiment of the present invention; Figure 9 A schematic diagram of an array of real triboelectric signal data collected after the triboelectric nanogenerator prototype provided by one embodiment of the present application was installed on a vehicle suspension test bench under excitation at different road surface levels; Figure 10 This is a schematic diagram of triboelectric signal image data under different road surface grade excitations after wavelet transform preprocessing provided in one embodiment of the present application; Figure 11 A schematic diagram of a confusion matrix of test results provided in one embodiment of the present application.

[0021] Among them, 1-first connecting part, 2-second connecting part, 3-shock absorber, 4-main part, 5-contact part, 6-boss, 7-mounting hole, 8-screw, 9-shock absorber upper seat, 10-shock absorber lower seat, 11-damper, 12-shock absorber spring. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] The purpose of the present invention is to provide a road surface grade identification method and a suspension system embedded with a friction nanogenerator. The friction nanogenerator is used as a self-powered sensor, and its output triboelectric signal is input into a convolutional neural network after wavelet transformation for road surface grade identification, so as to solve the problem that the suspension dynamic stroke sensor used in the prior art must be equipped with additional power supply devices such as a power supply.

[0024] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] like Figures 1 to 11As shown, in order to ensure recognition accuracy while significantly simplifying the complexity of the road surface recognition process, the present application provides a road surface grade recognition method, which includes: Step 1: Obtaining triboelectric signals: When road excitation is input into the vehicle's suspension system, vibrations are generated, and the triboelectric nanogenerator, which operates synchronously with the suspension system, outputs triboelectric signals.

[0026] Step 2: Use wavelet transform to convert the triboelectric signal into image data.

[0027] Step 3: Build a road surface grade recognition model based on deep learning network.

[0028] Step 4: Input the image data into the road surface grade recognition model to obtain the road surface grade recognition result.

[0029] In actual application, the implementation process of the road surface grade identification method provided by this application can be as follows: Figure 1 As shown in the figure, the pre-trained convolutional neural network is the road surface grade recognition model mentioned above.

[0030] Furthermore, in order to further improve the accuracy of road surface recognition, the implementation process of the above step 3 may include: Step 31: Get the deep learning network and dataset.

[0031] Step 32: Divide the dataset into a training set and a test set. Each training set and test set includes multiple training sample pairs. Each training sample pair consists of historical image sample data and sample data labels. The historical image sample data is converted from the historical triboelectric signals output by the triboelectric nanogenerator.

[0032] Step 33: The deep learning network is trained using the historical image sample data in the training set as input and the sample data labels corresponding to these historical image sample data as output. The trained deep learning network is then tested using samples from the test set until the output error of the trained deep learning network reaches a set threshold. The trained deep learning network is then used as the road surface grade recognition model. The output error of the deep learning network is determined based on the prediction results of the deep learning network and the corresponding sample data labels.

[0033] Furthermore, in order to enhance the signal feature representation and facilitate the recognition of road surface grade, in this application, the deep learning network used is as follows: Figure 8 Taking the convolutional neural network shown as an example, the implementation process of the road surface grade recognition method provided above in this application is explained.

[0034] The collected triboelectric signal data are preprocessed using wavelet transform to convert the one-dimensional time series into a two-dimensional time-frequency image. For example, the converted time-frequency image is as follows: Figure 7 As shown, its size is 256×256×3, which is a three-channel color image data.

[0035] The convolutional neural network is trained using several time-frequency image data. The structure and parameter settings of the convolutional neural network are as follows: (1) Input layer: The input size is 256×256×3.

[0036] (2) Convolutional layer 1: The input channels are 3, the output channels are 32, the convolution kernel size is 3, the padding is 1, the stride is 1, batch normalization and LeakyReLU activation function are set, and the output size is 256×256×32.

[0037] (3) Pooling layer 1: The pooling kernel size is 2, the stride is 2, and the output size is 128×128×32.

[0038] (4) Convolutional layer 2: The input channel is 32, the output channel is 64, the convolution kernel size is 3, the padding is 1, the stride is 1, batch normalization and LeakyReLU activation function are set, and the output size is 128×128×64.

[0039] (5) Pooling layer 2: The pooling kernel size is 2, the stride is 2, and the output size is 64×64×64.

[0040] (6) Convolutional layer 3: The input channel is 64, the output channel is 128, the convolution kernel size is 3, the padding is 1, the stride is 1, batch normalization and LeakyReLU activation function are set, and the output size is 64×64×128.

[0041] (7) Pooling layer 3: The pooling kernel size is 2, the stride is 2, and the output size is 32×32×128.

[0042] (8) Convolutional layer 4: The input channel is 128, the output channel is 256, the convolution kernel size is 3, the padding is 1, the stride is 1, batch normalization and LeakyReLU activation function are set, and the output size is 32×32×256.

[0043] (9) Pooling layer 4: The pooling kernel size is 2, the stride is 2, and the output size is 16×16×256.

[0044] (10) Convolutional layer 5: The input channel is 256, the output channel is 512, the convolution kernel size is 3, the padding is 1, the stride is 1, batch normalization and LeakyReLU activation function are set, and the output size is 16×16×512.

[0045] (11) Pooling layer 5: The pooling kernel size is 2, the stride is 2, and the output size is 8×8×512.

[0046] (12) Fully connected layer: The input is Flattened to 8×8×512.

[0047] (13) Hidden layer 1: Input is 100, output is 4.

[0048] Among them, the input layer size of the convolutional neural network is determined by the image size obtained after wavelet transform preprocessing, and the final output category is determined by the number of road surface grades contained in the image data.

[0049] Based on the above description, this application uses wavelet transform to preprocess the collected triboelectric signal data to solve the problem of cumbersome preprocessing. In addition, wavelet transform can effectively extract time-frequency information from triboelectric signals. The transformed data can be used directly as input for convolutional neural networks, eliminating the need for additional preprocessing of the triboelectric signals and simplifying the preprocessing process.

[0050] Figure 9 The present application shows the real triboelectric signal data collected after the friction nanogenerator prototype prepared was installed on a vehicle suspension test bench under the stimulation of four road surface grades A, B, C, and D. Two groups of triboelectric signals with a length of 150s were collected for each road surface grade. It can be seen that as the road surface grade increases, the amplitude of the triboelectric signal also increases, indicating that the triboelectric signal output by the proposed triboelectric nanogenerator has the ability to reflect the road surface grade. The collected data was divided into a format with a length of 4s and a time interval of 2s, and 150 groups of sample data for each of the four road surface grades were obtained. The sample data was preprocessed using wavelet transform to obtain 150 image data for each of the four road surface grades. Figure 10 The triboelectric signal image data under different road surface grades after wavelet transform preprocessing is shown. The preprocessed image data is divided into training set, validation set and test set in a ratio of 3:1:1, and input into the constructed deep learning network for training and testing. The confusion matrix of the test results is as follows Figure 11 As shown in the figure, it can be seen that the recognition accuracy of the trained network for road surface grade reaches 98.33%, which proves the feasibility and effectiveness of the proposed road surface grade recognition method.

[0051] in, Figure 9 In the first column, A-1 and A-2 correspond to the real triboelectric signal data collected under the excitation of Class A road surface. Figure 9 In the second column, B-1 and B-2 correspond to the real triboelectric signal data collected under the B-level road excitation. Figure 9In the third column, C-1 and C-2 correspond to the real triboelectric signal data collected under the C-level road excitation. Figure 9 D-1 and D-2 in the fourth column correspond to the actual triboelectric signal data collected under D-level road excitation. Figure 10 Part (a) shows the image data of Class A road surface. Figure 10 Part (b) shows the image data of a Class B road surface. Figure 10 Part (c) shows the image data of a Class C road surface. Figure 10 Part (d) shows the image data of a Class D road surface.

[0052] This application uses a deep learning network to identify road surface grades, addressing the complexity of the recognition process. Deep learning networks are simple to build, with strong feature extraction and generalization capabilities. They significantly simplify the recognition process while maintaining high recognition accuracy, effectively reducing the cost and barrier to maintenance and iteration, and facilitating secondary development and expansion.

[0053] In addition, installing a pre-trained convolutional deep learning network on a vehicle can achieve real-time recognition of road surface grade while the vehicle is driving, which can save computing resources while improving response speed.

[0054] like Figures 2 to 6 As shown, the present invention also provides a suspension system embedded with a friction nanogenerator, comprising a shock absorber upper seat 9, a shock absorber lower seat 10, a shock absorber 3 installed between the shock absorber upper seat 9 and the shock absorber lower seat 10, and a friction nanogenerator matched with the shock absorber 3; the friction nanogenerator comprises a generator body distributed in parallel with the shock absorber 3, the generator body comprises a first connecting portion 1 fixed at a position above the shock absorber 3, and a second connecting portion 2 fixed at a position below the shock absorber, the first connecting portion 1 and the second connecting portion 2 move relative to each other along the shock absorption direction of the shock absorber 3; specifically, it should be noted that the first connecting portion 1 and the second connecting portion 2 are both fixed on the suspension system structure outside the shock absorber 3, and their installation positions correspond to the top and bottom ends of the shock absorber 3, respectively, so as to achieve parallel distribution with the shock absorber 3. The first connecting part 1 and the second connecting part 2 move relative to each other along the damping direction of the shock absorber 3; the first connecting part 1 is provided with an interdigital electrode that moves synchronously with it, and the interdigital electrode includes a main body part 4 and a contact part 5 for outputting the friction electric signal to the outside, and the contact part 5 extends to the outside of the main body part 4 to facilitate the connection and arrangement of the wires. The second connecting part 2 is provided with a friction part that is close to the main body part 4, and the friction part moves synchronously with the second connecting part 2 and rubs with the main body part 4. The preferred second connecting part 2 is an L-shaped structure, and its bottom is fixedly connected to the position below the shock absorber 3 to increase the connection structure with the suspension system and improve the connection strength. The top of the second connecting part 2 is used to connect the friction part to make full use of the space of the suspension system and achieve a reasonable and reliable fixed connection.

[0055] The first connection portion 1 of the generator body is fixedly connected to the upper shock absorber seat 9, and the second connection portion 2 of the generator body is fixedly connected to the lower shock absorber seat 10. Specifically, the first connection portion 1 is an insulated hollow cylindrical sleeve, which is fixed to the upper shock absorber seat by screws 8. The bottom position of the second connection portion 2 is fixed to the lower shock absorber seat by screws 8. The upper shock absorber seat and the lower shock absorber seat are respectively fixed to the vehicle body and the wheel. Specifically, the upper shock absorber seat 9 and the lower shock absorber seat 10 are provided with mounting holes 7 for connecting the vehicle body and the wheel, respectively. The friction portion of the second connection portion 2 is in close contact with the interdigital electrodes attached to the inner wall of the first connection portion 1. The triboelectric signal generated by the vibration of the suspension system is output to the outside through the contact portion 5 of the interdigital electrodes.

[0056] The friction part is pressed against the main part 4 of the forked electrode to ensure stable contact between the two. Then, during the vibration of the suspension system, the friction part and the main part 4 of the forked electrode can continue to rub against each other, thereby generating a friction electric signal for road grade identification and outputting it through the contact part 5. As a self-powered sensor, the generator helps to reduce dependence on external power supply and reduce the complexity of the system. At the same time, the structure of the entire generator body is simple, which fully reduces the processing and maintenance costs and extends its service life.

[0057] Furthermore, the shock absorber 3 includes a damper 11 and a shock-absorbing spring 12. The damper 11 is connected at the middle position between the shock absorber upper seat 9 and the shock absorber lower seat 10. The first connecting part 1 adopts an insulating hollow cylindrical sleeve, the top of which is connected to the shock absorber upper seat 9, and the whole is surrounded by the outer peripheral side of the damper 11, and there is an annular gap between the first connecting part 1 and the damper 11 for the second connecting part 2 to extend into. The shock-absorbing spring 12 is surrounded by the outer peripheral side of the first connecting part 1 and has an annular gap between it and the first connecting part 1.

[0058] In a specific embodiment, the friction portion includes a plurality of bosses 6 arranged in sequence along the direction of movement of the second connecting portion 2. Each boss 6 moves synchronously with the second connecting portion 2 and rubs against the main portion 4. The distribution of each boss 6 makes the entire friction portion have a grid-like structure to optimize charge distribution: the grid-like design can increase the contact area of ​​the friction surface, making the charge distribution more uniform. When the main portion 4 of the interdigitated electrode and the friction portion rub against each other, the charge will be evenly distributed on the surface, thereby reducing the unevenness of charge accumulation and improving power generation efficiency. It also improves energy conversion efficiency: the grid-like design increases the number of contacts, allowing mechanical motion to be more efficiently converted into electrical energy. The multiple bosses 6 form multiple friction contacts, significantly improving power generation efficiency, and can exhibit high energy conversion efficiency even at lower vibration frequencies. Furthermore, it enhances mechanical stability: the grid-like design can enhance the stability of the mechanical structure while ensuring power generation efficiency. The grid-like structure can disperse stress and reduce stress concentration caused by mechanical motion, thereby extending the service life of the equipment.

[0059] In a specific embodiment, the surface of the boss 6 that rubs against the main body 4 is coated with polytetrafluoroethylene. The surface of the boss 6 is coated with polytetrafluoroethylene so that the friction resistance of the boss 6 is small and the wear resistance is high. The main body 4 of the interdigitated electrode is preferably made of copper material so that it can output a more characteristic friction electric signal when rubbing against the boss 6.

[0060] In a specific embodiment, the width of each boss 6 is equivalent to the width of a single electrode on the interdigitated electrode, and the groove width between two adjacent bosses 6 is equivalent to the spacing between the electrodes. When relative displacement occurs, the area of ​​friction between different electrodes and the outer surfaces of different bosses 6 remains consistent to enhance the characteristics of the output electrical signal.

[0061] In a specific embodiment, the first connecting part 1 is an insulating hollow cylindrical sleeve, the axis of the first connecting part 1 extends along the shock absorption direction of the shock absorber 3, and its top is fixed at a position above the shock absorber 3, and its bottom is provided with an opening for the second connecting part 2 to extend into, and the interdigital electrode is arranged on the inner wall of the first connecting part 1. By providing a hollow cylindrical sleeve, protection is formed for the friction part and the interdigital electrode to prevent the friction between the two from being interfered with by the outside world, so as to fully ensure the output of the electrical signal.

[0062] In a specific embodiment, the interdigitated electrode includes an interdigitated electrode substrate made of a flexible circuit board, and the substrate is tightly fitted to the inner wall of the first connecting part 1, so that the interdigitated electrode can conform to the inner wall structure of the first connecting part 1, thereby enhancing the connection strength between the two.

[0063] Adaptive changes based on actual needs are all within the scope of protection of the present invention.

[0064] It should be noted that it will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the present invention is defined by the appended claims, not the foregoing description, and it is intended that all variations within the meaning and range of equivalents of the claims be encompassed within the present invention. Any reference signs in the claims should not be construed as limiting the claim to which they relate.

[0065] The present invention uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A road surface grade identification method, characterized in that: The road surface grade identification method comprises: Acquiring a triboelectric signal; the triboelectric signal is output by a triboelectric nanogenerator that operates synchronously with the suspension system; Using wavelet transform to convert the triboelectric signal into image data; Construct a road surface grade recognition model based on deep learning network; The image data is input into the road surface grade recognition model to obtain a road surface grade recognition result.

2. The road surface grade identification method according to claim 1, characterized in that: Build a road surface grade recognition model based on a deep learning network, including: Obtain deep learning networks and datasets; The data set is divided into a training set and a test set; the training set and the test set each include a plurality of training sample pairs; each training sample pair consists of historical image sample data and a sample data label; the historical image sample data is obtained by converting the historical triboelectric signal output by the triboelectric nanogenerator; The deep learning network is trained with the historical image sample data in the training set as input and the sample data label corresponding to the historical image sample data as output, and the trained deep learning network is tested with the samples in the test set until the output error of the trained deep learning network reaches a set threshold, and the trained deep learning network is used as a road surface grade recognition model.

3. The road surface grade identification method according to claim 2, characterized in that: The deep learning network is a convolutional neural network.

4. A suspension system embedded with a triboelectric nanogenerator for use in the road surface grade identification method according to any one of claims 1 to 3, characterized in that: The device comprises a shock absorber upper seat, a shock absorber lower seat, a shock absorber installed between the shock absorber upper seat and the shock absorber lower seat, and a friction nanogenerator matched with the shock absorber; The triboelectric nanogenerator includes a generator body distributed in parallel with the shock absorber, the generator body includes a first connecting portion fixed at a position above the shock absorber and a second connecting portion fixed at a position below the shock absorber, and the first connecting portion and the second connecting portion move relative to each other along the shock absorption direction of the shock absorber; The first connecting portion is provided with an interdigital electrode that moves synchronously therewith, and the interdigital electrode includes a main portion and a contact portion for outputting a triboelectric signal to the outside. The second connecting portion is provided with a friction portion that is in close contact with the main portion, and the friction portion moves synchronously with the second connecting portion and rubs against the main portion. The first connecting portion of the generator body is fixedly connected to the upper seat of the shock absorber, and the second connecting portion of the generator body is fixedly connected to the lower seat of the shock absorber.

5. The suspension system embedded with a triboelectric nanogenerator according to claim 4, characterized in that: The friction portion includes a plurality of bosses arranged in sequence and at intervals along the movement direction of the second connecting portion. Each boss moves synchronously with the second connecting portion and rubs against the main body.

6. The suspension system embedded with a triboelectric nanogenerator according to claim 5, characterized in that: The surface of the boss that rubs against the main body is coated with polytetrafluoroethylene.

7. The suspension system embedded with a triboelectric nanogenerator according to claim 5 or 6, characterized in that: The width of each of the bosses is equal to the width of a single electrode on the interdigital electrodes, and the width of the groove between two adjacent bosses is equal to the interval between the electrodes.

8. The suspension system embedded with a triboelectric nanogenerator according to claim 7, characterized in that: The first connecting part is an insulating hollow cylindrical sleeve, the axis of the first connecting part extends along the shock absorbing direction of the shock absorber, and the top of the first connecting part is fixed at a position above the shock absorber, and the bottom of the first connecting part is provided with an opening for the second connecting part to extend into, and the interdigitated electrode is arranged on the inner wall of the first connecting part.

9. The suspension system embedded with a triboelectric nanogenerator according to claim 8, characterized in that: The interdigital electrode includes an interdigital electrode substrate made of a flexible circuit board, and the substrate is tightly attached to the inner wall of the first connecting portion.

Images