Method for identifying road surface type, computing device and storage medium

Abstract

The embodiment of the invention relates to a method and equipment for identifying the type of a road surface, and a medium. The method comprises the following steps: acquiring a speed signal about an unsprung position of a target vehicle and an angular speed signal at the center of mass, wherein the angular speed signal comprises a pitch angular speed signal and a roll angle speed signal; determining a vibration signal of the unsprung position and an angular vibration signal at the center of mass of the target vehicle based on the acquired acceleration signal of the unsprung position and the angular velocity signal at the center of mass; and generating a prediction result about the type of the road surface on which the target vehicle travels through a trained neural network model at least based on the determined vibration signal of the unsprung position and the angular vibration signal at the center of mass. Therefore, the prediction of the type of the road surface on which the target vehicle runs can be effectively realized with relatively low hardware cost and computing power cost.

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to the field of vehicle automatic control, and more particularly to a method, computing device, and storage medium for road surface type identification. Background Art

[0002] Traditional methods for road recognition, such as computer vision solutions, usually require high-precision and high-frame-rate on-board cameras; solutions such as lidar sensors and point cloud maps usually also require image acquisition equipment or on-board radar equipment; the signals obtained by these solutions usually require a lot of computing power when processing, which leads to higher hardware and software costs to obtain the real-time status of the road in order to meet the needs of real-time vehicle control.

[0003] In addition, traditional road surface recognition methods do not specifically distinguish between road types, resulting in an inability to adjust control strategies more specifically based on road type differences when controlling the vehicle based on the recognition results.

[0004] In summary, the shortcomings of the traditional method for road surface recognition are: the hardware cost and computing power cost are high, and the recognition accuracy is insufficient, and the recognition results are difficult to meet the needs of precise vehicle control. Summary of the invention

[0005] In response to the above problems, the present disclosure provides a method, a computing device and a storage medium for identifying road surface types, which can effectively improve the recognition accuracy of road surface types and save hardware and computing power costs.

[0006] According to a first aspect of the present disclosure, there is provided a method for identifying a road type, comprising: acquiring a velocity signal about an unsprung position of a target vehicle and an angular velocity signal at a center of mass, the angular velocity signal comprising a pitch angular velocity signal and a roll angular velocity signal; determining a vibration signal about the unsprung position of the target vehicle and an angular vibration signal at the center of mass based on the acquired acceleration signal of the unsprung position and the angular velocity signal at the center of mass; and

[0007] Based on at least the determined vibration signal at the unsprung position and the angular vibration signal at the center of mass, a prediction result regarding the type of road surface on which the target vehicle is traveling is generated through the trained neural network model.

[0008] According to a second aspect of the present disclosure, a computing device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method of the first aspect of the present disclosure.

[0009] In a third aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program code is stored. When the computer program code is executed, the method of the first aspect of the present disclosure is executed.

[0010] In some embodiments, obtaining an acceleration signal about an unsprung position of a target vehicle includes: obtaining a velocity signal of the unsprung position of the target vehicle based on a predetermined sampling period and sampling frequency, the unsprung position including at least a left front wheel unsprung position and a right front wheel unsprung position of the target vehicle; and integrating the obtained acceleration signal of the unsprung position based on a predetermined time step and then removing a trend term to obtain a velocity signal of the unsprung position of the target vehicle.

[0011] In some embodiments, obtaining a velocity signal about an unsprung position and an angular velocity signal at the center of mass of a target vehicle includes: controlling an inertial measurement unit installed at the center of mass of the target vehicle to collect a pitch angular velocity signal and a roll angular velocity signal about the target vehicle; controlling a first acceleration sensor installed at an unsprung position of a left front wheel of the target vehicle to collect a left front unsprung acceleration signal; controlling a second acceleration sensor installed at an unsprung position of a right front wheel of the target vehicle to collect a right front unsprung acceleration signal; and, the first acceleration sensor and the second acceleration sensor are both zero-frequency sensors, and the predetermined sampling period and sampling frequency of the first acceleration sensor, the second acceleration sensor, and the inertial measurement unit are all the same.

[0012] In some embodiments, determining a vibration signal regarding an unsprung position and an angular vibration signal at the center of mass of a target vehicle includes: extracting a predetermined number of sampling data from the acquired velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position based on a predetermined delay time via a tap delay device to generate sampling series corresponding to the velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position, respectively; and performing fast Fourier transform on the generated sampling series to obtain corresponding frequency spectrum information; and determining the vibration signal regarding the unsprung position and the angular vibration signal at the center of mass of the target vehicle based on the acquired frequency spectrum information.

[0013] In some embodiments, generating a prediction result about the type of road surface on which the target vehicle is traveling by a trained neural network model includes: based on the acquired multiple groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass, acquiring amplitude-frequency data about each group of vibration signals at the unsprung position and the angular vibration signals at the center of mass, so as to determine multiple groups of maximum amplitudes and frequencies corresponding to the maximum amplitudes corresponding to the multiple groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass, respectively; generating input feature data based on the maximum amplitudes and the frequencies corresponding to the maximum amplitudes, the velocity signal of the unsprung position of the target vehicle, and the angular velocity signal at the center of mass; and inputting the generated input feature data into the trained neural network model, so as to predict the type of road surface on which the target vehicle is currently traveling.

[0014] In some embodiments, the trained neural network model is obtained by training through the following steps: obtaining multiple sets of training vibration data about the target vehicle, vehicle speed data corresponding to the training vibration data, and road surface types corresponding to the training vibration data, the multiple sets of training vibration data are generated based on the left front unsprung vibration signal, the right front unsprung vibration signal, the pitch angle vibration signal at the center of mass, and the roll angle vibration signal at the center of mass; determining the road surface type corresponding to the training vibration data as verification data, determining the training vibration data and the vehicle speed data corresponding to the training vibration data as training data of the neural network model, so as to obtain the road surface type predicted by the neural network model; and based on the training vibration data, the vehicle speed data corresponding to the training vibration data, the road surface type corresponding to the training vibration data, and the predicted road surface type, training the neural network model with a predetermined loss function until the model converges, so as to obtain a trained neural network model.

[0015] In some embodiments, the prediction result about the type of road surface on which the target vehicle is traveling includes at least one or more of ordinary roads, wavy roads, impact roads, cross-axle roads, and bumpy roads.

[0016] In some embodiments, the road surface type is related to the flatness, slope and / or unevenness type of the road on which the vehicle is traveling, and the road surface type corresponding to the training vibration data is associated with at least one of the amplitude characteristics and frequency characteristics of the left front unsprung vibration signal, the right front unsprung vibration signal, the pitch angle vibration signal at the center of mass, and the roll angle vibration signal at the center of mass of the target vehicle.

[0017] In some embodiments, the method for identifying a road type further includes: matching a vibration reduction strategy for the target vehicle based on the determined road type on which the target vehicle is currently traveling, so as to control a vibration reduction system carried by the target vehicle based on the matched vibration reduction strategy.

[0018] In some embodiments, the vibration reduction strategy includes at least two of the following: a standard vibration reduction control strategy, a vibration isolation control strategy, an impact vibration reduction control strategy, and a resonance control strategy; and

[0019] Controlling the vibration reduction system carried by the target vehicle at least includes: controlling the current of the magnetorheological vibration reducer carried by the target vehicle.

[0020] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements.

[0022] Figure 1 A schematic diagram of a scenario for implementing a method for identifying a road surface type according to an embodiment of the present invention is shown.

[0023] Figure 2 A flow chart of a method 200 for identifying a road surface type according to an embodiment of the present disclosure is shown.

[0024] Figure 3 A flowchart of a method 300 for generating a prediction result of a road surface type according to an embodiment of the present disclosure is shown.

[0025] Figure 4 A flowchart of a method 400 for training a neural network model according to an embodiment of the present disclosure is shown.

[0026] Figure 5 A schematic diagram of a neural network model input data processing process according to an embodiment of the present disclosure is shown.

[0027] Figure 6 A block diagram of an electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0028] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0029] As used herein, the term "including" and its variations mean open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "based at least in part on". The terms "an example embodiment" and "an embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0030] As described above, in the case of identifying the road type of the road on which the vehicle is traveling, the traditional methods for road type identification, such as computer vision solutions, such as laser radar sensors, point cloud maps, etc., usually require a large amount of computing power when processing the signals obtained by these solutions, resulting in high hardware and software costs to obtain the real-time status of the road surface in order to meet the needs of real-time vehicle control. In summary, the shortcomings of the traditional methods for identifying road types are: the hardware cost and computing power cost are both high, and the recognition accuracy is insufficient, and the recognition results are difficult to meet the needs of precise vehicle control.

[0031] In order to at least partially solve one or more of the above problems and other potential problems, an exemplary embodiment of the present disclosure proposes a solution for identifying the type of road surface. In the solution of the present disclosure, by acquiring a velocity signal about the unsprung position of the target vehicle and an angular velocity signal at the center of mass, the angular velocity signal includes a pitch angular velocity signal and a roll angular velocity signal; based on the acquired acceleration signal of the unsprung position and the angular velocity signal at the center of mass, a vibration signal about the unsprung position of the target vehicle and an angular vibration signal at the center of mass are determined; and based on at least the determined vibration signal of the unsprung position and the angular vibration signal at the center of mass, a prediction result about the type of road surface on which the target vehicle is traveling is generated through a trained neural network model. The present disclosure can acquire multiple vibration signals about the target vehicle while driving, and based on the obtained vibration signals, predict the type of road surface on which the target vehicle is traveling through a trained neural network, so that the target vehicle can be controlled accordingly based on the predicted road surface type.

[0032] Figure 1 FIG. 1 is a schematic diagram showing an environment 100 for implementing a method for identifying a road surface type according to an embodiment of the present invention. Figure 1As shown in , environment 100 includes computing device 110, target vehicle 130, multiple sensors 130-x carried on the target vehicle, and network 140. Computing device 110, target vehicle 130, and multiple sensors 130-3 carried on the target vehicle can communicate with each other. Computing device 110, target vehicle 130, and multiple sensors 130-x carried on the target vehicle can exchange data through network 140 (e.g., Internet, communication network), or directly exchange data through communication connection. Computing device 110 can be independently carried on target vehicle 130, or it can be integrated with the control system of target vehicle 130 and carried on target vehicle 130.

[0033] The target vehicle 130, which may, for example, be equipped with a plurality of sensors for collecting initial signals of a velocity signal of an unsprung position of the target vehicle and an angular velocity signal at the center of mass, communicates with the computing device 110 to send the initial signals collected by the plurality of sensors carried on the target vehicle 130 to the computing device 110 in real time.

[0034] For more information about the multiple sensors installed on the target vehicle 130, please refer to Figure 1 The target vehicle 130 is provided with acceleration sensors 130 - 1 and 130 - 2 at unsprung positions, which are respectively arranged at unsprung positions of the left front wheel and the right front wheel, and an inertial measurement unit 130 - 3 is provided at the center of mass of the target vehicle 130 .

[0035] Regarding the computing device 110, it is used, for example, to communicate with the target vehicle 130 and multiple sensors carried on the target vehicle 130 to obtain initial signals collected by the multiple sensors, and based on the obtained initial signals, process to obtain the velocity signal of the unsprung position of the target vehicle and the angular velocity signal at the center of mass to determine the vibration signal of the unsprung position and the angular vibration signal at the center of mass, so as to generate a prediction result about the type of road surface on which the target vehicle is traveling through a trained neural network model.

[0036] The computing device 110 may have one or more processing units, including dedicated processing units such as GPU, FPGA and ASIC, and general processing units such as CPU. In addition, one or more virtual machines may also be running on each computing device 110. In some embodiments, the computing device 110 and the control system of the target vehicle 130 may be integrated together or may be separately arranged. In some embodiments, the computing device 110 includes, for example, a speed signal acquisition module 112, a vibration signal acquisition module 114, and a road surface type prediction module 116.

[0037] The speed signal acquisition module 112 is used to acquire a speed signal of an unsprung position of the target vehicle and an angular speed signal at the center of mass, wherein the angular speed signal includes a pitch angular speed signal and a roll angular speed signal.

[0038] The vibration signal acquisition module 114 is used to determine a vibration signal of the unsprung position and an angular vibration signal at the center of mass of the target vehicle based on the acquired acceleration signal of the unsprung position and the angular velocity signal at the center of mass.

[0039] The road type prediction module 116 is configured to generate a prediction result on the road type on which the target vehicle is traveling based on at least the determined vibration signal at the unsprung position and the angular vibration signal at the center of mass through a trained neural network model.

[0040] Figure 2 FIG. 2 is a flow chart of a method 200 for identifying a road surface type according to an embodiment of the present disclosure. The method 200 may be performed as follows: Figure 1 The computing device 110 shown may also be executed in Figure 6 The method 200 is executed at the electronic device 600. It should be understood that the method 200 may further include additional steps not shown and / or may omit the steps shown, and the scope of the present disclosure is not limited in this respect.

[0041] In step 202 , the computing device 110 acquires a velocity signal about an unsprung position of the target vehicle and an angular velocity signal at the center of mass, where the angular velocity signal includes a pitch angular velocity signal and a roll angular velocity signal.

[0042] Regarding the pitch angle, it is, for example, the angle at which the front or rear of the vehicle is lifted relative to the horizontal plane of the center of mass, such as the front of the vehicle is lifted when the vehicle is going uphill, the rear of the vehicle is lifted when the vehicle is going downhill, and the rear of the vehicle is lifted during emergency braking, etc. Regarding the roll angle, also known as the roll angle, it is, for example, the swing amplitude of the left and right sides of the vehicle body relative to the horizontal plane of the center of mass. When the vehicle turns to one side, or when one side of the vehicle's wheels is higher or lower than the other side, the vehicle body rolls, which will cause a change in the roll angle.

[0043] Regarding the speed signal of the unsprung position of the target vehicle, the speed signal of the unsprung position can be obtained by using an acceleration sensor installed at the unsprung position of the target vehicle to obtain the acceleration signal of the unsprung position, and then processing the acceleration signal.

[0044] In some embodiments, obtaining an acceleration signal about an unsprung position of a target vehicle includes: obtaining a velocity signal of the unsprung position of the target vehicle based on a predetermined sampling period and sampling frequency; and the unsprung position includes at least an unsprung position of a left front wheel and an unsprung position of a right front wheel of the target vehicle. For example, by setting the sampling period and sampling frequency of sensors of multiple unsprung positions to the same value, multiple groups of acceleration signals of unsprung positions obtained by multiple sensors, and the obtained source signals achieve a synchronization effect, and no additional synchronization operation is required after obtaining multiple groups of signals.

[0045] In addition, in the method for identifying the road type provided in the embodiment of the present disclosure, the unsprung position may only require the left front wheel unsprung position and the right front wheel unsprung position of the target vehicle, that is, the front wheel unsprung position of the target vehicle. It should be understood that during normal driving, the road surface that the front wheels of the target vehicle pass through will also be passed by its rear wheels. By detecting the acceleration signal under the front wheel sprung wheel, the road surface that the target vehicle has passed through can be fully reflected, and the computing power requirement is relatively low, which can meet the lower latency road type prediction needs and vehicle control needs.

[0046] In some embodiments, the acceleration signal of the unsprung position is integrated based on a predetermined time step and then the trend term is removed to obtain the velocity signal of the unsprung position of the target vehicle. For example, the acceleration signal of the unsprung position is integrated based on a predetermined step by using a reverse Euler time integration method.

[0047] In the method for identifying the road type provided in the embodiment of the present disclosure, for the acquired acceleration signal of the unsprung position, a method of first integrating and then removing the trend term is adopted, so that a more accurate integration result can be obtained. This is determined by the vibration signal characteristics of the target vehicle. The vibration caused by the road surface during the driving of the target vehicle is a low-frequency signal on most roads, and the acceleration signal of the unsprung position is affected by gravity when it is collected. The obtained result tends to be that the gravity acceleration g will fluctuate at a low frequency near g. Since the trend term is close to the constant gravity acceleration g, a more accurate result can be obtained by integrating after removing the trend term.

[0048] In some embodiments, obtaining a velocity signal about an unsprung position of a target vehicle and an angular velocity signal at the center of mass includes: controlling an inertial measurement unit installed at the center of mass of the target vehicle to collect a pitch angular velocity signal and a roll angular velocity signal about the target vehicle; controlling a first acceleration sensor installed at an unsprung position of a left front wheel of the target vehicle to collect a left front unsprung acceleration signal; and controlling a second acceleration sensor installed at an unsprung position of a right front wheel of the target vehicle to collect a right front unsprung acceleration signal.

[0049] In some embodiments, the first acceleration sensor and the second acceleration sensor are both zero-frequency sensors. Regarding the use of a zero-frequency sensor to collect acceleration signals, since what is to be obtained is a vibration signal about the vehicle body, and the vibration frequency of the vehicle body during driving is relatively low, usually about 1 Hz under the sprung position, which is a low-frequency signal, a zero-frequency sensor can be used to obtain a more accurate measurement result. In addition, when a zero-frequency sensor is used to measure the acceleration signal of the unsprung position, the data measured by the zero-frequency sensor will fluctuate above and below the gravity acceleration g. Therefore, in combination with the method of integrating the unsprung acceleration signal of the target vehicle obtained in the embodiment of the present disclosure first and then removing the trend term, a more accurate unsprung velocity signal can be obtained.

[0050] In some embodiments, the predetermined sampling period and sampling frequency of the first acceleration sensor, the second acceleration sensor, and the inertial measurement unit are the same. Specifically, the sampling period and sampling frequency of the sensor used to obtain the vehicle vibration signal are set to the same value, so that the obtained source signals can be synchronized, so that the vibration signal of the target vehicle can be obtained without additional synchronization.

[0051] The predetermined sampling period and sampling frequency are set based on parameters such as the actual application scenario, computing power conditions, and target vehicle configuration, so that the collected unsprung acceleration signal, pitch angular velocity signal, and roll angular velocity signal of the target vehicle can be processed to obtain more accurate vibration signals of the unsprung position of the target vehicle and angular vibration signals at the center of mass, and more accurate vibration signals of multiple target vehicles associated with road surface type characteristics can be obtained, so as to achieve more accurate road surface type identification.

[0052] At step 204 , the computing device 110 determines a vibration signal at the unsprung position and an angular vibration signal at the center of mass of the target vehicle based on the acquired acceleration signal at the unsprung position and the angular velocity signal at the center of mass.

[0053] In some embodiments, determining a vibration signal regarding an unsprung position and an angular vibration signal at the center of mass of a target vehicle includes: extracting a predetermined number of sampling data from the acquired velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position based on a predetermined delay time via a tap delay device to generate sampling series corresponding to the velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position, respectively; and performing fast Fourier transform on the generated sampling series to obtain corresponding frequency spectrum information; and determining the vibration signal regarding the unsprung position and the angular vibration signal at the center of mass of the target vehicle based on the acquired frequency spectrum information.

[0054] Please refer to Figure 5, which shows that after the unsprung velocity signal and the angular velocity signal at the center of mass of the target vehicle pass through the tap delay device and the fast Fourier transform, the data obtained are the pitch angle vibration signal, the roll angle vibration signal, the left front unsprung vibration signal and the right front unsprung vibration signal.

[0055] Regarding the tap delay device, in the solution provided by the embodiment of the present disclosure, the processing of the acceleration signal of the unsprung position obtained is not processed by the traditional high-pass filter, but a tap delay device is selected, which can delay less than 2ms when obtaining the sampled data (theoretically, there can even be no delay); while the traditional solution using a filter usually causes the delay time to reach more than 10ms due to the characteristics of the filter itself. Therefore, the above solution has a smaller delay, can obtain road type identification data faster, and can more real-timely match the corresponding vibration reduction strategy for the target vehicle according to the road type on which the target vehicle is traveling, so as to control the vibration reduction system of the target vehicle faster.

[0056] For example, the predetermined delay time is T, the predetermined sampling frequency is M, and the tap delay device continuously extracts sample data from the predetermined delay time T at T / M as the time point, for example, T / M=t, and sampling is performed every t time, and the M data collected at the time t within the T time length are used as a set of sampling series. As the time is updated, the velocity signal, pitch angular velocity signal, and roll angular velocity signal of the acquired unsprung position are continuously sampled to generate sampling series containing M sampling data respectively. For example, T=1 second, M=200, and the generated sampling series of the left front unsprung velocity signal, the right front unsprung velocity signal, the pitch angular velocity signal, and the roll angular velocity signal of the target vehicle respectively contain 200 sampling data.

[0057] Regarding the spectrum information, after fast Fourier transform is performed on the sampling series of the left front unsprung velocity signal, the right front unsprung velocity signal, the pitch angular velocity signal and the roll angular velocity signal, the spectrum information of the left front unsprung velocity signal, the right front unsprung velocity signal, the pitch angular velocity signal and the roll angular velocity signal can be obtained respectively. These spectrum information contain information such as the amplitude, frequency, and period of the signal, and can reflect the vibration information of the corresponding position. Specifically, according to the obtained spectrum information, the time domain, frequency domain information, amplitude spectrum and phase spectrum of the corresponding signal can be obtained to determine the vibration conditions of the left front unsprung, the right front unsprung, the pitch angle and the roll angle of the target vehicle.

[0058] At step 206 , the computing device 110 generates a prediction result regarding the type of road surface on which the target vehicle is traveling based on at least the determined vibration signal at the unsprung position and the angular vibration signal at the center of mass through the trained neural network model.

[0059] In some embodiments, the prediction result about the type of road surface on which the target vehicle is traveling includes at least one or more of ordinary roads, wavy roads, impact roads, cross-axle roads, and bumpy roads.

[0060] In some embodiments, the road surface type is related to the flatness, undulation type and body posture of the road on which the vehicle is traveling, and the road surface type corresponding to the training vibration data is associated with multiple amplitude characteristics and frequency characteristics of the left front spring under-vibration signal, the right front spring under-vibration signal, the pitch angle vibration signal at the center of mass and the roll angle vibration signal at the center of mass of the target vehicle. It is worth noting that the road surface type is essentially a characteristic of the road surface itself. Since it is difficult to directly collect road surface morphological characteristics in actual scenarios, this solution adopts a method of obtaining body posture information through sensors to assist in determining the road surface type.

[0061] For example, straight paved roads are the most common type of urban roads, such as asphalt and cement roads. The road surface is flat, the amplitude is very small, and the body vibration frequency is random between 1-10Hz, and the bumps of the road surface and the shaking of the body are almost unnoticeable; on wavy roads, the road surface is undulating, and passengers feel weightless / overweight during vehicle driving, and the body vertical vibration amplitude is large and the frequency is low, between 1-2Hz; on impact roads, it is often in the shape of a raised circle, square, or trapezoid, and the body vibration time is short, but the amplitude is very large. The main frequency corresponding to the amplitude is usually equal to the body resonance frequency, and the body generally shakes slightly; on cross-axis roads, it is a left and right slope, the amplitude and frequency are constant, and it is in the shape of two sine waves, and the two waves are staggered, with corresponding peaks and valleys. When the vehicle passes, the occupants feel a strong shaking, the body rolls significantly, and the roll angle vibration frequency is close to the main frequency; on bumpy roads (such as gravel roads, slate roads, gravel roads, and damaged cement roads), the body is stable as a whole, but bumpy, the unsprung vibration amplitude is large, and the body is generally relatively stable or shakes slightly.

[0062] Therefore, by analyzing the characteristics of the vibration signal at the unsprung position and the angular vibration signal at the center of mass of the target vehicle, the amplitude and frequency characteristics of the vibration signal can be matched with the specific road surface type to obtain an accurate road surface type prediction result.

[0063] In the above scheme, by controlling multiple sensors at the unsprung position and the center of mass position of the target vehicle body, the unsprung acceleration signal and the center of mass angular velocity signal of the target vehicle can be synchronously collected, and the source signal obtained by the sensor is processed, and the sampling data is extracted by a tap delay device to generate a sampling sequence corresponding to each signal, and the sampling sequence corresponding to the unsprung acceleration signal and the center of mass angular velocity signal is fast Fourier transformed to obtain the frequency spectrum information of the left front unsprung velocity signal, the right front unsprung velocity signal, the pitch angular velocity signal and the roll angular velocity signal corresponding to the target vehicle respectively, thereby obtaining the unsprung vibration signal and the angular vibration signal at the center of mass of the target vehicle, and based on the obtained multiple vibration signals, the type of road surface on which the target vehicle is traveling is predicted through a neural network model, and various road surface types can be identified in real time and accurately through the above vibration signals, and the hardware requirements and computing power costs are greatly saved.

[0064] Figure 3 FIG. 3 is a flowchart of a method 300 for generating a prediction result of a road surface type according to an embodiment of the present disclosure. The method 300 may be performed as follows: Figure 1 The computing device 110 shown may also be executed in Figure 6 The method 300 is executed at the electronic device 600. It should be understood that the method 300 may further include additional steps not shown and / or may omit the steps shown, and the scope of the present disclosure is not limited in this respect.

[0065] In step 302, the computing device 110 obtains amplitude-frequency data about each group of vibration signals at the unsprung position and the angular vibration signals at the center of mass based on the obtained multiple groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass, so as to determine multiple groups of maximum amplitudes and frequencies corresponding to the maximum amplitudes respectively corresponding to the multiple groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass.

[0066] Regarding amplitude-frequency data, a set of amplitude-frequency data of vibration signals includes a variety of amplitudes, frequencies and vibration period information. In this solution, since the body vibration is a low-frequency signal, the vibration signal with the largest amplitude has the lowest frequency, and the vibration information with the largest amplitude can more directly reflect the road conditions, this solution generates input feature data about the set of vibration signals based on the maximum amplitude of the vibration signal and the frequency corresponding to the maximum amplitude. For example, four sets of input data are generated based on the maximum amplitude and the frequency corresponding to the maximum amplitude of the left front spring under-vibration signal, the right front spring under-vibration signal, the pitch angle vibration signal and the roll angle vibration signal.

[0067] In addition, by taking the maximum amplitude and the frequency corresponding to the maximum amplitude as the input characteristic data of the vibration signal in the above manner, limited computing power can be concentrated on the signal of interest to improve computing efficiency and speed up the response time of road type identification.

[0068] At step 304 , the computing device 110 generates input feature data based on the maximum amplitude and the frequency corresponding to the maximum amplitude, the velocity signal about the unsprung position of the target vehicle, and the angular velocity signal at the center of mass.

[0069] At step 306 , the computing device 110 inputs the generated input feature data into the trained neural network model to predict the type of road surface on which the target vehicle is currently traveling.

[0070] Please refer to Figure 5 , which illustrates the input data of the neural network model 510, the input data are the maximum amplitude and the frequency corresponding to the maximum amplitude (hereinafter referred to as the main frequency) corresponding to the pitch angle vibration signal, the maximum amplitude and the main frequency corresponding to the roll angle vibration signal, the maximum amplitude and the main frequency corresponding to the left front unsprung vibration signal, and the maximum amplitude and the main frequency corresponding to the right front unsprung vibration signal; that is, there are 4 groups and 8 input data in total, and the output result of the neural network model 510 is the predicted road surface type.

[0071] Therefore, there are four groups of data input into the trained neural network model, namely, the input feature data of the left front spring vibration signal, the right front spring vibration signal, the pitch angle vibration signal and the roll angle vibration signal, and each group of input feature data is composed of the maximum amplitude of the corresponding vibration signal and the frequency corresponding to the maximum amplitude.

[0072] Figure 4 400 is a flowchart of a method 400 for training a neural network model according to an embodiment of the present disclosure. The method 400 may be performed as follows: Figure 1 The computing device 110 shown may also be executed in Figure 6 The method 400 is executed at the electronic device 600 shown. It should be understood that the method 400 may further include additional steps not shown and / or may omit the steps shown, and the scope of the present disclosure is not limited in this respect.

[0073] In step 402, the computing device 110 obtains multiple sets of training vibration data about the target vehicle, vehicle speed data corresponding to the training vibration data, and a road surface type corresponding to the training vibration data, wherein the multiple sets of training vibration data are generated based on a left front unsprung vibration signal, a right front unsprung vibration signal, a pitch angle vibration signal at the center of mass, and a roll angle vibration signal at the center of mass.

[0074] For example, a set of training data a1 includes a1{vehicle speed a1 , Left front unsprung vibration a1 , right front unsprung vibration a1 , pitch angle vibration at the center of mass a1 , roll angle vibration at the center of mass a1}, such as vibration a1is the maximum amplitude of the corresponding vibration signal and the frequency corresponding to the maximum amplitude. The frequency corresponding to the maximum amplitude can be called the main vibration frequency, such as the left front spring under vibration a1 = {Maximum amplitude under left front spring a1 , left front spring vibration main frequency a1}, it should be understood that other training data can be deduced by analogy and will not be given one by one.

[0075] In step 404, the computing device 110 determines the road surface type corresponding to the training vibration data as verification data, and determines the training vibration data and the vehicle speed data corresponding to the training vibration data as training data of the neural network model to obtain the road surface type predicted by the neural network model.

[0076] Regarding verification data, for example, for a set of training data a1 obtained, the actual type of the road surface on which the target vehicle is traveling during actual collection is the verification data. The road surface type of the verification data can be classified by the tester based on the actual flatness, slope and / or unevenness of the road surface.

[0077] In step 406 , the computing device 110 trains the neural network model with a predetermined loss function based on the training vibration data, the vehicle speed data corresponding to the training vibration data, the road surface type corresponding to the training vibration data, and the predicted road surface type until the model converges to obtain a trained neural network model.

[0078] Therefore, training data can be generated based on the sensor data collected by the target vehicle when it is actually driving on the road, and verification data can be obtained based on the actual road conditions. The data can be put into the neural network model for training, which can enable the trained neural network model to have better prediction results for the road type in the real scene.

[0079] Regarding neural network models, such as convolutional neural networks (CNN), recurrent neural networks (RNN), deep belief networks (DBN), deep autoencoders (AutoEncoder), generative adversarial networks (GAN), long short-term memory networks (LSTM), and transformer models that can be used for classification, you can choose an adapted model and make adjustments based on the hardware and computing power support of the actual usage scenario.

[0080] In some embodiments, the method for identifying a road type further includes: matching a vibration reduction strategy for the target vehicle based on the determined road type on which the target vehicle is currently traveling, so as to control a vibration reduction system carried by the target vehicle based on the matched vibration reduction strategy.

[0081] For example, in response to determining that the type of the road currently traveled on has changed, the vibration reduction strategy is re-matched for the target vehicle so as to adjust the vibration reduction strategy for the target vehicle in real time based on the change in the road type. For example, if it is determined that the road type has changed from an ordinary road to a bumpy road, the vibration reduction strategy is switched from a standard vibration reduction control strategy to a vibration isolation control strategy.

[0082] In some embodiments, the vibration reduction strategy includes at least two of the following: a standard vibration reduction control strategy, a vibration isolation control strategy, an impact vibration reduction control strategy, and a resonance control strategy; and controlling the vibration reduction system carried by the target vehicle at least includes: controlling the current of a magnetorheological shock absorber carried by the target vehicle.

[0083] For example, on straight paved roads and the default mode, the standard vibration reduction control strategy is used, on bumpy roads, the vibration isolation control strategy is used, on impact roads, the impact control strategy is used, and on wavy roads and cross-axle roads, the resonance control strategy is used. The choice of the resonance strategy and the vibration type reflected by the road type. It should be understood that when other road types are added, the corresponding vibration reduction strategy can also be preset or matched.

[0084] Therefore, the above scheme can match a suitable vibration reduction strategy for the target vehicle based on the determined road type on which the target vehicle is currently traveling, and adjust the new vibration reduction strategy according to the change of the road type of the target vehicle, so that the target vehicle can adopt a vibration reduction strategy that matches its road type and vehicle body vibration state in real time, so as to improve the user's experience and comfort when riding.

[0085] Figure 6 Schematic diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. Figure 1 The computing device 110 shown can be implemented by an electronic device 600. As shown, the electronic device 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) 602 or computer program instructions loaded from a storage unit 608 to a random access memory (RAM) 603. In the random access memory 603, various programs and data required for the operation of the electronic device 600 can also be stored. The central processing unit 601, the read-only memory 602, and the random access memory 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0086] Multiple components in the electronic device 600 are connected to the input / output interface 605, including: an input unit 606, such as a keyboard, a mouse, a microphone, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0087] The various processes and processing described above, such as methods 200 to 400, may be performed by the central processing unit 601. For example, in some embodiments, methods 200 to 400 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via the read-only memory 602 and / or the communication unit 609. When the computer program is loaded into the random access memory 603 and executed by the central processing unit 601, one or more actions of the methods 200 to 400 described above may be performed.

[0088] The present disclosure relates to methods, apparatuses, systems, electronic devices, computer-readable storage media and / or computer program products. The computer program products may include computer-readable program instructions for executing various aspects of the present disclosure.

[0089] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0090] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge computing devices. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.

[0091] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0092] Various aspects of the present disclosure are described herein with reference to flowcharts and / or step diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each of the steps in the flowcharts and / or step diagrams and the combination of the steps in the flowcharts and / or step diagrams can be implemented by computer-readable program instructions.

[0093] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more steps in the flowchart and / or step diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more steps in the flowchart and / or step diagram.

[0094] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more steps in the flowchart and / or step diagram.

[0095] The flowcharts and step diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each step in the flowchart or step diagram can represent a module, a program segment or a part of an instruction, and a module, a program segment or a part of an instruction contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the steps can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each step in the step diagram and / or the flowchart, and the combination of the steps in the step diagram and / or the flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0096] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for identifying a road surface type, comprising: Acquire a velocity signal about an unsprung position of the target vehicle and an angular velocity signal at a center of mass, wherein the angular velocity signal includes a pitch angular velocity signal and a roll angular velocity signal; Determine a vibration signal of the unsprung position and an angular vibration signal at the center of mass of the target vehicle based on the acquired acceleration signal of the unsprung position and the angular velocity signal at the center of mass; as well as Based on at least the determined vibration signal at the unsprung position and the angular vibration signal at the center of mass, a prediction result regarding the type of road surface on which the target vehicle is traveling is generated through the trained neural network model.

2. The method according to claim 1, wherein acquiring an acceleration signal about an unsprung position of the target vehicle comprises: Acquire a speed signal of an unsprung position of the target vehicle based on a predetermined sampling period and sampling frequency, wherein the unsprung position includes at least an unsprung position of a left front wheel and an unsprung position of a right front wheel of the target vehicle; as well as The acquired acceleration signal of the unsprung position is integrated based on a predetermined time step and then a trend term is removed to acquire a velocity signal of the unsprung position of the target vehicle.

3. The method according to claim 2, wherein acquiring a velocity signal about an unsprung position of the target vehicle and an angular velocity signal at a center of mass comprises: Controlling an inertial measurement unit installed at the center of mass of the target vehicle to collect a pitch angular velocity signal and a roll angular velocity signal of the target vehicle; Controlling a first acceleration sensor installed at an unsprung position of a left front wheel of a target vehicle to collect an acceleration signal of the unsprung left front wheel; Controlling a second acceleration sensor installed at a right front wheel undersprung position of the target vehicle to collect an acceleration signal under the right front wheel undersprung; as well as The first acceleration sensor and the second acceleration sensor are both zero-frequency sensors, and the predetermined sampling period and sampling frequency of the first acceleration sensor, the second acceleration sensor, and the inertial measurement unit are the same.

4. The method according to any one of claims 1 to 3, wherein determining a vibration signal about an unsprung position of the target vehicle and an angular vibration signal at a center of mass comprises: A tap delay device extracts a predetermined number of sample data from the acquired velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position based on a predetermined delay time to generate a sample sequence corresponding to the velocity signal, pitch angular velocity signal, and roll angular velocity signal of the unsprung position, respectively; Performing fast Fourier transform on the generated sampling sequences respectively to obtain corresponding spectrum information; as well as Based on the acquired frequency spectrum information, a vibration signal about an unsprung position and an angular vibration signal at a center of mass of the target vehicle are determined.

5. The method according to claim 4, wherein generating a prediction result about the type of road surface on which the target vehicle is traveling by means of the trained neural network model comprises: Based on the obtained plurality of groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass, amplitude-frequency data about each group of vibration signals at the unsprung position and the angular vibration signals at the center of mass are obtained, so as to determine a plurality of groups of maximum amplitudes and frequencies corresponding to the maximum amplitudes respectively corresponding to the plurality of groups of vibration signals at the unsprung position and the angular vibration signals at the center of mass; generating input feature data based on the maximum amplitude and the frequency corresponding to the maximum amplitude, a velocity signal about an unsprung position of the target vehicle, and an angular velocity signal at the center of mass; as well as The generated input feature data is input into the trained neural network model to predict the type of road surface on which the target vehicle is currently traveling.

6. The method according to claim 1, wherein the trained neural network model is obtained by training through the following steps: Acquire multiple sets of training vibration data about the target vehicle, vehicle speed data corresponding to the training vibration data, and a road surface type corresponding to the training vibration data, wherein the multiple sets of training vibration data are generated based on a left front unsprung vibration signal, a right front unsprung vibration signal, a pitch angle vibration signal at the center of mass, and a roll angle vibration signal at the center of mass; Determining the road surface type corresponding to the training vibration data as verification data, and determining the training vibration data and the vehicle speed data corresponding to the training vibration data as training data of the neural network model to obtain the road surface type predicted by the neural network model; and Based on the training vibration data, the vehicle speed data corresponding to the training vibration data, the road surface type corresponding to the training vibration data and the predicted road surface type, the neural network model is trained with a predetermined loss function until the model converges to obtain a trained neural network model.

7. The method according to claim 1, wherein the prediction result about the type of road surface on which the target vehicle is traveling includes at least one or more of ordinary roads, wavy roads, impact roads, cross-axle roads and bumpy roads.

8. The method according to claim 7, wherein the road surface type is related to the flatness, slope and / or unevenness type of the road on which the vehicle is traveling, and the road surface type corresponding to the training vibration data is associated with at least one of the amplitude characteristics and frequency characteristics of the left front unsprung vibration signal, the right front unsprung vibration signal, the pitch angle vibration signal at the center of mass, and the roll angle vibration signal at the center of mass of the target vehicle.

9. The method according to any one of claims 1-3 or 5-8, further comprising: Based on the determined road type on which the target vehicle is currently traveling, a vibration reduction strategy is matched for the target vehicle, so as to control a vibration reduction system carried by the target vehicle based on the matched vibration reduction strategy.

10. The method according to claim 9, wherein the vibration reduction strategy comprises at least two of the following: a standard vibration reduction control strategy, a vibration isolation control strategy, an impact vibration reduction control strategy, and a resonance control strategy; and The vibration reduction system carried by the control target vehicle at least includes: Controls the current of the magnetorheological damper mounted on the target vehicle.

11. A computing device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 10.

12. A computer-readable storage medium, characterized in that: A computer program is stored on a computer-readable storage medium, and when the computer program is executed by a machine, the method according to any one of claims 1 to 10 is implemented.

Images