Typical road surface recognition methods, systems, computers, and readable storage media

Abstract

This invention provides a typical driving surface recognition method, system, computer, and readable storage medium. The method includes acquiring a vehicle's unsprung acceleration signal, vehicle speed signal, and a training driving surface set, wherein the training driving surface set includes several template driving surface types; calculating a first signal feature based on the unsprung acceleration signal, and calculating a second signal feature based on the unsprung acceleration signal and the vehicle speed signal; inputting the training driving surface set into a preset classification model, encoding each template driving surface type into a preset number of error correction codes, and training the preset classification model based on several error correction codes to obtain a training classification model; inputting the first signal feature and the second signal feature into the training classification model, and outputting the driving surface recognition result. This invention can achieve automated signal processing and driving surface classification recognition, and improves the accuracy and efficiency of driving surface classification recognition.

Description

Technical Field

[0001] This invention belongs to the technical field of road surface recognition, specifically relating to typical driving road surface recognition methods, systems, devices, and readable storage media. Background Technology

[0002] Vehicle durability, as one of the key performance indicators of automobiles, is directly related to their safety, reliability, and economy, and affects their core competitiveness, thus receiving increasing attention. In vehicle durability engineering, the vehicle structural load spectrum is an important basis for component structural design, life prediction, experimental verification, and structural optimization. Only when the vehicle structural load spectrum can accurately and reasonably reflect the load state of the vehicle in the actual use environment can the components and the entire vehicle designed based on it achieve good durability performance.

[0003] Automotive structural load spectrum is often obtained through the experience of durability engineers or through long-distance real-vehicle data collection. However, the former requires a high level of experience and judgment from engineers and is difficult to ensure that it reflects the load of the vehicle under real-world usage conditions. The latter often requires a large investment of manpower and resources, increasing the research and development time and financial costs of durability engineering. Therefore, to better understand the loads borne by the vehicle, it is necessary to identify and record the road surface types that the vehicle travels on during driving, providing important reference value for the compilation of automotive structural load spectrum. In the current technology, road surface type identification is usually done by taking pictures of the vehicle in motion, which requires a large investment of manpower and resources and cannot guarantee the accuracy of road surface identification. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a typical driving road surface recognition method, system, device, and readable storage medium to resolve the technical issues existing in the prior art.

[0005] In a first aspect, the invention provides the following technical solution: a typical road surface recognition method, the method comprising:

[0006] Acquire the vehicle's unsprung acceleration signal, vehicle speed signal, and training road surface set, wherein the training road surface set includes several template road surface types;

[0007] Calculate the first signal feature based on the unsprung acceleration signal, and calculate the second signal feature based on the unsprung acceleration signal and the vehicle speed signal;

[0008] The training road surface set is input into a preset classification model, and each template road surface type is written as a preset number of error correction codes. The preset classification model is trained based on several error correction codes to obtain a training classification model.

[0009] The first signal feature and the second signal feature are input into the training classification model to output the road surface recognition result.

[0010] Compared with existing technologies, the beneficial effects of this application are as follows: This application first acquires the unsprung acceleration signal, vehicle speed signal, and a training road surface set, wherein the training road surface set includes several template road surface types; then, based on the unsprung acceleration signal, a first signal feature is calculated, and based on the unsprung acceleration signal and the vehicle speed signal, a second signal feature is calculated; the training road surface set is input into a preset classification model, and each template road surface type is written as a preset number of error correction codes. The preset classification model is trained based on several error correction codes to obtain a training classification model; finally, the first signal feature and the second signal feature are input into the training classification model, and the road surface recognition result is output. This application acquires the temporal and spatial domain features of the unsprung acceleration signal and vehicle speed signal, and outputs the final road surface recognition result based on the extracted features and a decision tree. This enables the invention to achieve automated signal processing and road surface classification and recognition, and improves the accuracy and efficiency of road surface classification and recognition.

[0011] Preferably, the step of calculating the first signal feature based on the vehicle speed signal includes:

[0012] Calculate the root mean square value (rms) of the unsprung acceleration signal in the time domain to obtain the first signal characteristic:

[0013]

[0014] In the formula, f(t) is the unsprung acceleration signal, and N is the number of signal points in the unsprung acceleration signal.

[0015] Preferably, the step of calculating the second signal characteristics based on the unsprung acceleration signal and the vehicle speed signal includes:

[0016] The unsprung acceleration signal is divided into several data segments, and the single-sided power spectrum G of each data segment is calculated based on the vehicle speed signal. xx (f):

[0017]

[0018]

[0019] In the formula, R xx (τ) is the autocorrelation function of the unsprung acceleration signal, f is the spatial domain sampling frequency, and Fs is the time domain sampling frequency. τ represents the average vehicle speed, τ represents the unsprung acceleration signal, and j represents the imaginary number.

[0020] The single-sided power spectrum G corresponding to all the data segments xx (f) Add them together and take the average value to obtain the spatial power density spectrum pxx(n);

[0021] Calculate the energy distribution ratio η of the unsprung acceleration signal in the low-frequency region to the high-frequency region based on the spatial power spectral density pxx(n):

[0022]

[0023]

[0024]

[0025] In the formula, n is the spatial frequency, S1 is the energy distribution of the unsprung acceleration signal in the low-frequency region, and S2 is the energy distribution of the unsprung acceleration signal in the high-frequency region.

[0026] Preferably, the steps of inputting the training road surface set into a preset classification model, encoding each template road surface type as a preset number of error correction codes, and training the preset classification model based on several of the error correction codes to obtain a trained classification model include:

[0027] The training road surface set is input into a preset classification model, and each template road surface type is written as a binary error correction code of a preset number of bits.

[0028] Extract the training features of each bit of the error correction code corresponding to all the template road surface types, and integrate the training features into a training dataset;

[0029] The preset classification model is trained using the training dataset to obtain the trained classification model.

[0030] Preferably, the step of training the preset classification model using the training dataset to obtain the trained classification model includes:

[0031] Define the impurity of the training dataset as Gini(D):

[0032]

[0033] In the formula, D is the training dataset, m is the number of tuples in the training dataset, and p i For any tuple to belong to C i The probability, C i To determine the types of subsets classified in the training dataset;

[0034] The training dataset D is divided into a first subset D1 and a second subset D2, and the Gini index after the partition is calculated. A (D):

[0035]

[0036] In the formula, Gini(D1) is the impurity of the first subset D1, and Gini(D2) is the impurity of the second subset D2;

[0037] Select the Gini index A (D) The first subset D1 and the second subset D2 corresponding to the minimum value, and take the first subset D1 and the second subset D2 as the classification subset;

[0038] Based on the classification subset, a decision tree corresponding to the number of bits of the error correction code is trained in the preset classification model to obtain the trained classification model.

[0039] Preferably, the step of inputting the first signal feature and the second signal feature into the training classification model and outputting the road surface recognition result includes:

[0040] The first signal feature and the second signal feature are input into the training classification model;

[0041] The decision tree in the training classification model outputs a string of decision characters consisting of 0 or 1 based on the first signal feature and the second signal feature in sequence.

[0042] The decision characters are integrated to obtain the output code, and the Hamming distance between the output code and each of the error correction codes is calculated.

[0043] The template road surface type corresponding to the error correction code with the smallest Hamming distance is selected as the road surface recognition result output.

[0044] Preferably, the unsprung acceleration signal is an acceleration signal acquired by an acceleration sensor installed at the wheel axle head and installed in a direction perpendicular to the horizontal plane.

[0045] Secondly, the invention provides the following technical solution: a typical road surface recognition system, the system comprising:

[0046] The acquisition module is used to acquire the unsprung acceleration signal, vehicle speed signal, and training driving road surface set of the vehicle, wherein the training driving road surface set includes several template road surface types;

[0047] The feature calculation module is used to calculate a first signal feature based on the unsprung acceleration signal and to calculate a second signal feature based on the unsprung acceleration signal and the vehicle speed signal.

[0048] The training module is used to input the training road surface set into a preset classification model, write each template road surface type as a preset number of error correction codes, and train the preset classification model based on several error correction codes to obtain a training classification model.

[0049] The road surface recognition module is used to input the first signal feature and the second signal feature into the training classification model and output the road surface recognition result.

[0050] Thirdly, the invention provides the following technical solution: a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described typical driving road surface recognition method.

[0051] Fourthly, the invention provides the following technical solution: a readable storage medium storing a computer program, which, when executed by a processor, implements the above-described typical road surface recognition method. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart of a typical road surface recognition method provided in the first embodiment of the present invention;

[0054] Figure 2 A detailed flowchart of step S22 in the typical road surface recognition method provided in the first embodiment of the present invention;

[0055] Figure 3 This is a schematic diagram of the spatial frequency domain characteristics of various template pavement types provided in the first embodiment of the present invention;

[0056] Figure 4 A detailed flowchart of step S3 in the typical road surface recognition method provided in the first embodiment of the present invention;

[0057] Figure 5 A detailed flowchart of step S33 in the typical road surface recognition method provided in the first embodiment of the present invention;

[0058] Figure 6 A detailed flowchart of step S4 in the typical road surface recognition method provided in the first embodiment of the present invention;

[0059] Figure 7 This is a structural block diagram of a typical road surface recognition system provided in the second embodiment of the present invention;

[0060] Figure 8 This is a hardware structure block diagram of a computer provided for another embodiment of the present invention.

[0061] The embodiments of the present invention will be further described below with reference to the accompanying drawings. Detailed Implementation

[0062] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain embodiments of the present invention, and should not be construed as limiting the present invention.

[0063] Example 1

[0064] like Figure 1 As shown, in the first embodiment of the present invention, the invention provides the following technical solution: a typical road surface recognition method, the method comprising:

[0065] S1. Acquire the unsprung acceleration signal, vehicle speed signal, and training road surface set of the vehicle, wherein the training road surface set includes several template road surface types;

[0066] The unsprung acceleration signal and vehicle speed signal can both be obtained by sensors installed on the vehicle. The unsprung acceleration signal is the acceleration signal obtained by an acceleration sensor installed at the wheel axle head with the installation direction perpendicular to the horizontal plane. The vehicle speed signal can be obtained by wheel speed sensor or by reading the CAN bus. The unsprung acceleration signal and vehicle speed signal need to be preprocessed, including but not limited to drift removal and low-pass filtering.

[0067] S2. Calculate the first signal characteristic based on the unsprung acceleration signal, and calculate the second signal characteristic based on the unsprung acceleration signal and the vehicle speed signal;

[0068] Specifically, step S2 includes: S21, calculating a first signal feature based on the unsprung acceleration signal; and S22, calculating a second signal feature based on the unsprung acceleration signal and the vehicle speed signal.

[0069] Step S21 includes:

[0070] Calculate the root mean square value (rms) of the unsprung acceleration signal in the time domain to obtain the first signal characteristic:

[0071]

[0072] In the formula, f(t) is the unsprung acceleration signal, and N is the number of signal points in the unsprung acceleration signal;

[0073] The unsprung acceleration signal is a signal containing N points. Therefore, when calculating the unsprung acceleration, it is necessary to first calculate the energy E of this unsprung acceleration signal.

[0074]

[0075] Then the average power P at each point is:

[0076]

[0077] Therefore, the root mean square value (rms) of the unsprung acceleration signal in the time domain is the square root of the average power at each point:

[0078]

[0079] As can be seen from the above formula, the root mean square value of a complete unsprung acceleration signal is equal to the average power of the average power of each point in that signal segment. Therefore, if the signal points contained in each segment are vertically the same, the root mean square value of each segment can be used to represent the amount of energy contained in that segment. Moreover, the energy transferred to the vehicle is different for different road surface types. In other words, the root mean square value of the unsprung acceleration signal per unit time will be different depending on the road surface type. Therefore, the root mean square value (rms) of the unsprung acceleration signal in the time domain can be used as the first signal feature.

[0080] like Figure 2 As shown, step S22 includes:

[0081] S221. Divide the unsprung acceleration signal into several data segments, and calculate the single-sided power spectrum G of each data segment based on the vehicle speed signal. xx (f):

[0082]

[0083]

[0084] In the formula, R xx (τ) is the autocorrelation function of the unsprung acceleration signal, f is the spatial domain sampling frequency, and Fs is the time domain sampling frequency. τ represents the average vehicle speed, τ represents the unsprung acceleration signal, and j represents the imaginary number.

[0085] The time-domain signal of the unsprung acceleration signal can be converted to spatial frequency using a specific algorithm, therefore the corresponding spatial domain sampling frequency f is:

[0086]

[0087] In the formula, L is the average speed of the vehicle in t seconds. The distance traveled, corresponding to N t The number of points collected.

[0088] S222, All the single-sided power spectra G corresponding to the data segments xx (f) Add them together and take the average value to obtain the spatial power density spectrum pxx(n);

[0089] Specifically, after calculating the spatial domain sampling frequency, the Welch method is used to calculate the spatial power density spectrum of the unsprung acceleration signal. The Welch method involves dividing a time-domain signal into multiple segments, each of which overlaps with the adjacent data segments. The power spectral density of each segment is calculated based on the above formula, and the power spectra are summed and averaged to obtain the spatial power density spectrum pxx(n). Compared with the method of directly calculating the power spectrum of an entire signal, the spatial power density spectrum pxx(n) obtained by the above method has better spectral variance characteristics and a smoother curve.

[0090] S223. Calculate the energy distribution ratio η of the unsprung acceleration signal in the low-frequency region and the high-frequency region based on the spatial power spectral density pxx(n):

[0091]

[0092]

[0093]

[0094] In the formula, n is the spatial frequency, S1 is the energy distribution of the unsprung acceleration signal in the low-frequency region, and S2 is the energy distribution of the unsprung acceleration signal in the high-frequency region;

[0095] like Figure 3 As shown, different road surface types contain different spatial frequency components, including undulating road, railway crossing, bumps, and saphalt roadhills. The energy of undulating roads is mainly concentrated in the [1, 3] m range. -1Within this range, there is also some energy distribution at higher frequencies, but it is much less than in the concentrated range. The energy at the railway crossing is clearly concentrated at low frequencies, especially when the spatial frequency is greater than 2m. -1 At that time, there is almost no energy distribution; the energy of the convex path is mainly concentrated in [2, 3]m. -1 In this interval, compared to undulating paths, the energy is more concentrated on convex paths, [2, 3]m -1 There is less energy distributed outside this interval; the energy of the hill is mainly concentrated in the [0, 2]m range. -1 Within this range, its energy distribution is similar to that of a railway crossing, concentrating towards the low-frequency region. However, the energy is more dispersed, as can be seen when the frequency is greater than 2m. -1 Even after that, a certain amount of energy is still distributed. In contrast, railway crossings have a spatial frequency greater than 2m. -1 At that time, there is almost no energy distribution. Therefore, in the subsequent road surface classification process, a decision tree is introduced to obtain the final classification result based on the different energy distributions of different road surface types at different spatial frequencies.

[0096] Therefore, this invention divides the spatial frequency into two intervals to obtain the low-frequency region [0, 2]m. -1 With high frequency region [2, 4]m -1 The spatial power spectral density pxx(n) is calculated, and the energy distribution ratio η of the unsprung acceleration signal in the low-frequency region and the high-frequency region is calculated. This value can characterize the distribution ratio of the energy acting on the vehicle by the road surface type in the low-frequency region and the high-frequency region. This value only describes the relative magnitude, while the absolute magnitude of the energy is reflected by the root mean square value in the first signal feature.

[0097] S3. Input the training road surface set into the preset classification model, write each template road surface type as a preset number of error correction codes, and train the preset classification model based on several error correction codes to obtain the training classification model.

[0098] Specifically, the training road surface set includes several template road surface types. In this embodiment, the training road surface set includes, but is not limited to, undulating roads, railway crossings, bumpy roads, and road hills.

[0099] like Figure 4 As shown, specifically, step S3 includes:

[0100] S31. Input the training road surface set into the preset classification model, and write each template road surface type as a binary error correction code of preset number of bits;

[0101] Specifically, in this embodiment, the preset number of bits is 7, so the corresponding template road surface type can be written as: undulating road 1111111, bump road 0000111, railway crossing 0011001, road mound 0101010.

[0102] S32. Extract the training features of each bit of the error correction code corresponding to all template road surface types, and integrate the training features into a training dataset;

[0103] Specifically, in this step, the training features are the same as the first signal features and the second signal features in the above steps. The energy ratio reflecting the distribution ratio of energy on the vehicle by the road surface type in the low-frequency region and the high-frequency region, as well as the root mean square value reflecting the absolute magnitude of the energy, are extracted. The above features are integrated into a training dataset, and training features can be obtained for each bit of the corresponding error correction code for each template road surface type. The whole dataset is then used as a complete training dataset.

[0104] S33. Train the preset classification model using the training dataset to obtain a trained classification model;

[0105] like Figure 5 As shown, specifically, step S33 includes:

[0106] S331. Define the impurity of the training dataset, Gini(D):

[0107]

[0108] In the formula, D is the training dataset, m is the number of tuples in the training dataset, and p i For any tuple to belong to C i The probability, C i The types of subsets classified in the training dataset.

[0109] S332. Divide the training dataset D into a first subset D1 and a second subset D2, and calculate the Gini index after the partition. A (D):

[0110]

[0111] In the formula, Gini(D1) is the impurity of the first subset D1, and Gini(D2) is the impurity of the second subset D2.

[0112] S333, Select the Gini index. A (D) The first subset D1 and the second subset D2 corresponding to the minimum value, and take the first subset D1 and the second subset D2 as the classification subset;

[0113] Specifically, since this step considers a binary partition, all possible subsets S that an attribute A with v distinct values ​​can form are considered. A There are 2 v There are (2) v -2) / 2 are possible ways to partition the dataset D into binary subsets. Therefore, for a given tuple, if the value of its attribute A belongs to the subset S... A Then the tuple is partitioned into the partition corresponding to the subset, so that after the binary partition D, there are D1 and D2. By selecting the smallest Gini index... A (D) The first subset D1 and the second subset D2 corresponding to it are used as the classification subsets.

[0114] S334. Based on the classification subset, train a decision tree corresponding to the number of bits of the error correction code in the preset classification model to obtain a trained classification model;

[0115] Specifically, since the decision tree is a tree structure containing decision conditions, when an element in the dataset satisfies the decision conditions, the first branch is output; otherwise, the second branch is output. In this embodiment, the first branch is 1 and the second branch is 0. By obtaining the classification subset, the decision tree and the decision conditions in the decision tree can be trained. In this embodiment, the decision conditions are the energy level of the road surface type reflected on the vehicle and the proportion of energy distribution in the low-frequency and high-frequency regions. Since the template road surface type is written as a seven-bit binary code consisting of 0 or 1, seven decision trees need to be trained in the model to output a seven-bit binary output code.

[0116] S4. Input the first signal feature and the second signal feature into the training classification model, and output the road surface recognition result;

[0117] like Figure 6 As shown, step S4 includes:

[0118] S41. Input the first signal feature and the second signal feature into the training classification model.

[0119] S42. The decision tree in the training classification model outputs a string of decision characters consisting of 0 or 1 according to the first signal feature and the second signal feature in sequence.

[0120] Specifically, since the seven decision trees have already been trained in step S3, by inputting the first signal feature and the second signal feature into the training classification model, the seven decision trees will output 0 or 1 in sequence according to their trained decision conditions and the first and second signal features, thus obtaining a string of decision characters.

[0121] S43. Integrate the decision characters to obtain the output code, and calculate the Hamming distance between the output code and each of the error correction codes;

[0122] Specifically, after obtaining the decision character, it is integrated into a seven-bit binary output code consisting of 0s or 1s, and the Hamming distance between the output code and each preset error correction code is calculated. Here, the Hamming distance refers to the number of bits that are different between the output code and the error correction code.

[0123] S44. Select the template road surface type corresponding to the error correction code with the smallest Hamming distance as the road surface recognition result output;

[0124] By selecting the error correction code corresponding to the smallest Hamming distance as the road surface recognition result, the smallest Hamming distance means that the similarity between the error correction code and the output code corresponding to the template road surface type is the best. Therefore, the template road surface type is output as the road surface recognition result.

[0125] The advantages of this embodiment are as follows: This application first acquires the unsprung acceleration signal, vehicle speed signal, and a training road surface set, wherein the training road surface set includes several template road surface types; then, based on the unsprung acceleration signal, a first signal feature is calculated, and based on the unsprung acceleration signal and the vehicle speed signal, a second signal feature is calculated; the training road surface set is input into a preset classification model, and each template road surface type is written as a preset number of error correction codes. The preset classification model is trained based on several error correction codes to obtain a training classification model; finally, the first signal feature and the second signal feature are input into the training classification model, and the road surface recognition result is output. This application acquires the temporal and spatial domain features of the unsprung acceleration signal and vehicle speed signal, and outputs the final road surface recognition result based on the extracted features and a decision tree. This enables the invention to achieve automated signal processing and road surface classification and recognition, and improves the accuracy and efficiency of road surface classification and recognition.

[0126] Example 2

[0127] like Figure 7 As shown, a typical road surface recognition system is provided in a second embodiment of the present invention, the system comprising:

[0128] The acquisition module 1 is used to acquire the unsprung acceleration signal, vehicle speed signal, and training driving road surface set of the vehicle, wherein the training driving road surface set includes several template road surface types;

[0129] Feature calculation module 2 is used to calculate a first signal feature based on the unsprung acceleration signal and to calculate a second signal feature based on the unsprung acceleration signal and the vehicle speed signal;

[0130] Training module 3 is used to input the training road surface set into a preset classification model, write each template road surface type into a preset number of error correction codes, and train the preset classification model based on several error correction codes to obtain a training classification model.

[0131] The road surface recognition module 4 is used to input the first signal feature and the second signal feature into the training classification model and output the road surface recognition result.

[0132] Specifically, the feature calculation module 2 is used for:

[0133] Calculate the root mean square value (rms) of the unsprung acceleration signal in the time domain to obtain the first signal characteristic:

[0134]

[0135] In the formula, f(t) is the unsprung acceleration signal, and N is the number of signal points in the unsprung acceleration signal.

[0136] The feature calculation module 2 includes:

[0137] The segmentation submodule is used to divide the unsprung acceleration signal into several data segments, and calculate the single-sided power spectrum G of each data segment based on the vehicle speed signal. xx (f):

[0138]

[0139]

[0140] In the formula, R xx (τ) is the autocorrelation function of the unsprung acceleration signal, f is the spatial domain sampling frequency, and Fs is the time domain sampling frequency. τ represents the average vehicle speed, τ represents the unsprung acceleration signal, and j represents the imaginary number.

[0141] The density spectrum calculation submodule is used to calculate the single-sided power spectrum G corresponding to all the data segments. xx (f) Add them together and take the average value to obtain the spatial power density spectrum pxx(n);

[0142] The energy ratio calculation submodule is used to calculate the energy distribution ratio η of the unsprung acceleration signal in the low-frequency region and the high-frequency region based on the spatial power spectral density pxx(n).

[0143]

[0144]

[0145]

[0146] In the formula, n is the spatial frequency, S1 is the energy distribution of the unsprung acceleration signal in the low-frequency region, and S2 is the energy distribution of the unsprung acceleration signal in the high-frequency region.

[0147] The training module 3 includes:

[0148] A submodule is written to input the training road surface set into a preset classification model and to write each template road surface type as a binary error correction code of a preset number of bits.

[0149] An extraction submodule is used to extract the training features of each bit of the error correction code corresponding to all the template road surface types, and integrate the training features into a training dataset.

[0150] The training submodule is used to train the preset classification model using the training dataset to obtain the training classification model.

[0151] The training submodule includes:

[0152] Define a unit to define the impurity Gini(D) of the training dataset:

[0153]

[0154] In the formula, D is the training dataset, m is the number of tuples in the training dataset, and p i For any tuple to belong to C i The probability, C i To determine the types of subsets classified in the training dataset;

[0155] The partitioning unit is used to divide the training dataset D into a first subset D1 and a second subset D2, and to calculate the Gini index after the partition. A (D):

[0156]

[0157] In the formula, Gini(D1) is the impurity of the first subset D1, and Gini(D2) is the impurity of the second subset D2;

[0158] Subset determination unit, used to select the Gini index. A (D) The first subset D1 and the second subset D2 corresponding to the minimum value, and take the first subset D1 and the second subset D2 as the classification subset;

[0159] The training unit is used to train a decision tree corresponding to the number of bits of the error correction code in the preset classification model based on the classification subset, so as to obtain a training classification model.

[0160] The road surface recognition module 4 includes:

[0161] An input submodule is used to input the first signal feature and the second signal feature into the training classification model;

[0162] The character output submodule is used to output a string of decision characters consisting of 0 or 1 sequentially based on the first signal feature and the second signal feature in the training classification model.

[0163] The Hamming distance calculation submodule is used to integrate the decision characters to obtain the output code, and calculate the Hamming distance between the output code and each of the error correction codes.

[0164] The output submodule is used to select the template road surface type corresponding to the error correction code with the smallest Hamming distance as the road surface recognition result output.

[0165] In other embodiments of the present invention, the present invention provides the following technical solution: a computer, including a memory 102, a processor 101, and a computer program stored in the memory 102 and executable on the processor 101, wherein the processor 101 executes the computer program to implement the typical driving road surface recognition method described above.

[0166] Specifically, the processor 101 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0167] The memory 102 may include a mass storage device for data or instructions. For example, and not limitingly, the memory 102 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 102 may include removable or non-removable (or fixed) media. Where appropriate, the memory 102 may be internal or external to a data processing device. In a particular embodiment, the memory 102 is non-volatile memory. In a particular embodiment, the memory 102 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random Access Memory (FPMDRAM), Extended Data Out Dynamic Random Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0168] The memory 102 can be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 101.

[0169] The processor 101 implements the above-described typical road surface recognition method by reading and executing computer program instructions stored in the memory 102.

[0170] In some embodiments, the computer may further include a communication interface 103 and a bus 100. For example, Figure 8 As shown, the processor 101, memory 102, and communication interface 103 are connected through bus 100 and complete communication with each other.

[0171] The communication interface 103 is used to enable communication between the various modules, devices, units, and / or equipment in the embodiments of this application. The communication interface 103 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0172] Bus 100 includes hardware, software, or both, that couples computer components together. Bus 100 includes, but is not limited to, at least one of the following: data bus, address bus, control bus, expansion bus, and local bus. For example, and not as a limitation, bus 100 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 100 may include one or more buses. Although specific buses are described and illustrated in the embodiments of this application, this application considers any suitable bus or interconnection.

[0173] The computer can use the typical driving road surface recognition system to execute the typical driving road surface recognition method of this application, thereby realizing the recognition of typical driving road surfaces.

[0174] In some further embodiments of the present invention, in conjunction with the above-described typical road surface recognition method, the present invention provides the following technical solution: a readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-described typical road surface recognition method.

[0175] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0176] More specific examples of readable media (a non-exhaustive list) include: electrical connections (electronic devices) with one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0177] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0178] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0179] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A typical road surface recognition method, characterized in that, The method includes: Acquire the vehicle's unsprung acceleration signal, vehicle speed signal, and training road surface set, wherein the training road surface set includes several template road surface types; Calculate the first signal feature based on the unsprung acceleration signal, and calculate the second signal feature based on the unsprung acceleration signal and the vehicle speed signal; The step of calculating the first signal feature based on the vehicle speed signal includes: Calculate the root mean square value of the unsprung acceleration signal in the time domain. rms To obtain the first signal characteristics: ； In the formula, This is the unsprung acceleration signal. This represents the number of signal points in the unsprung acceleration signal. The step of calculating the second signal characteristic based on the unsprung acceleration signal and the vehicle speed signal includes: The unsprung acceleration signal is divided into several data segments, and the single-sided power spectrum of each data segment is calculated based on the vehicle speed signal. : ； ； In the formula, Let be the autocorrelation function of the unsprung acceleration signal. For spatial domain sampling frequency, The sampling frequency in the time domain. Average vehicle speed This is the unsprung acceleration signal. It is the symbol for imaginary numbers; One-sided power spectra corresponding to all the data segments The sums are then averaged to obtain the spatial power density spectrum. ; According to the spatial power spectral density Calculate the energy distribution ratio of the unsprung acceleration signal in the low-frequency region to that in the high-frequency region. : ； ； ； In the formula, For spatial frequency, The energy distribution of the unsprung acceleration signal in the low-frequency region. The energy distribution of the unsprung acceleration signal in the high-frequency region; The training road surface set is input into a preset classification model, and each template road surface type is written as a preset number of error correction codes. The preset classification model is trained based on several error correction codes to obtain a training classification model. The first signal feature and the second signal feature are input into the training classification model to output the road surface recognition result.

2. The typical road surface recognition method according to claim 1, characterized in that, The steps of inputting the training road surface set into a preset classification model, writing each template road surface type as a preset number of error correction codes, and training the preset classification model based on several error correction codes to obtain a trained classification model include: The training road surface set is input into a preset classification model, and each template road surface type is written as a binary error correction code of a preset number of bits. Extract the training features of each bit of the error correction code corresponding to all the template road surface types, and integrate the training features into a training dataset; The preset classification model is trained using the training dataset to obtain the trained classification model.

3. The typical road surface recognition method according to claim 2, characterized in that, The step of training the preset classification model using the training dataset to obtain the trained classification model includes: Define the impurity of the training dataset : ； In the formula, For the training dataset, The number of tuples in the training dataset, For any tuple to belong to C i The probability, C i To determine the types of subsets classified in the training dataset; training dataset Divided into the first subset With the second subset And calculate the Gini index after division. : ； In the formula, For the first subset impurity, For the second subset Impurity; Select the Gini index The first subset corresponding to the minimum With the second subset and the first subset With the second subset As a subset of categories; Based on the classification subset, a decision tree corresponding to the number of bits of the error correction code is trained in the preset classification model to obtain the trained classification model.

4. The typical road surface recognition method according to claim 1, characterized in that, The step of inputting the first signal feature and the second signal feature into the training classification model and outputting the road surface recognition result includes: The first signal feature and the second signal feature are input into the training classification model; The decision tree in the training classification model outputs a string of decision characters consisting of 0 or 1 based on the first signal feature and the second signal feature in sequence. The decision characters are integrated to obtain the output code, and the Hamming distance between the output code and each of the error correction codes is calculated. The template road surface type corresponding to the error correction code with the smallest Hamming distance is selected as the road surface recognition result output.

5. The typical road surface recognition method according to claim 1, characterized in that, The unsprung acceleration signal is the acceleration signal acquired by an acceleration sensor installed at the wheel axle head with its installation direction perpendicular to the horizontal plane.

6. A typical driving road surface recognition system, wherein the system employs the typical driving road surface recognition method as described in claim 1, characterized in that, The system includes: The acquisition module is used to acquire the unsprung acceleration signal, vehicle speed signal, and training driving road surface set of the vehicle, wherein the training driving road surface set includes several template road surface types; The feature calculation module is used to calculate a first signal feature based on the unsprung acceleration signal and to calculate a second signal feature based on the unsprung acceleration signal and the vehicle speed signal. The training module is used to input the training road surface set into a preset classification model, write each template road surface type as a preset number of error correction codes, and train the preset classification model based on several error correction codes to obtain a training classification model. The road surface recognition module is used to input the first signal feature and the second signal feature into the training classification model and output the road surface recognition result.

7. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the typical driving road surface recognition method as described in any one of claims 1 to 5.

8. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements the typical driving road surface recognition method as described in any one of claims 1 to 5.

Images