A method, device, vehicle and storage medium for classifying and identifying road roughness

Abstract

The present invention discloses a method, device, vehicle and storage medium for classifying and identifying road surface roughness. The method comprises obtaining at least one set of training data sequences, training a preset neural network model for at least one round according to the at least one set of training data sequences, and obtaining a road surface roughness prediction model; after a front vehicle occludes the vehicle on the predicted future driving trajectory, collecting the real-time posture data of the front vehicle; inputting the real-time posture data into the road surface roughness prediction model, and determining the roughness classification and recognition result that matches the road surface on which the front vehicle is traveling. The technical solution of the present invention increases the accuracy and robustness of the road surface roughness prediction model, and improves the accuracy of identifying the road surface roughness ahead.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a method, device, vehicle and storage medium for classifying and recognizing road roughness. Background Art

[0002] Under urban road conditions, the vehicle's forward vision sensor field of view is frequently occupied by other vehicles on the road, making it unable to perceive the road conditions ahead, which in turn causes the failure of the existing road surface roughness recognition function. This function affects other functions such as vehicle mode switching, vehicle dynamics state prediction, and active suspension. If it fails, it will lead to a decrease in vehicle safety, comfort and other performance. Existing technologies usually use visual sensors to identify the elevation and roughness of the road ahead.

[0003] In the process of realizing the present invention, the inventors found that the prior art had the following defects: the existing method for detecting the unevenness of the road ahead did not take into account the problem of obstructed vision in urban conditions that is widely present in actual applications, resulting in the inability to perceive the road ahead conditions. Summary of the invention

[0004] The present invention provides a method, device, vehicle and storage medium for classifying and identifying road surface roughness, so as to increase the accuracy and robustness of a road surface roughness prediction model and improve the accuracy of identifying road surface roughness ahead.

[0005] According to one aspect of the present invention, a method for classifying and identifying road surface roughness is provided, the method comprising:

[0006] Acquire at least one set of training data sequences, and train a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road roughness prediction model;

[0007] After a leading vehicle blocks the vehicle's future predicted driving trajectory, real-time position and posture data of the leading vehicle is collected;

[0008] The real-time position data is input into the road surface roughness prediction model to determine an roughness classification recognition result that matches the road surface on which the preceding vehicle is traveling.

[0009] According to another aspect of the present invention, a device for classifying and identifying road surface roughness is provided, the device comprising:

[0010] A road surface roughness prediction model acquisition module is used to acquire at least one set of training data sequences, and to train a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model;

[0011] A posture data acquisition module is used to collect real-time posture data of the preceding vehicle after the preceding vehicle blocks the vehicle's future travel prediction trajectory;

[0012] The classification recognition result determination module is used to input the real-time posture data into the road surface roughness prediction model to determine the roughness classification recognition result that matches the road surface on which the preceding vehicle is traveling.

[0013] According to another aspect of the present invention, a vehicle is provided, the vehicle comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for classifying and identifying road surface roughness described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for classifying and identifying road surface roughness described in any embodiment of the present invention when executed.

[0018] The technical solution of the embodiment of the present invention obtains at least one set of training data sequences, and trains a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model; after a leading vehicle appears to be blocked on the predicted future driving trajectory of the vehicle, the real-time posture data of the leading vehicle is collected; the real-time posture data is input into the road surface roughness prediction model to determine the roughness classification and recognition result matching the road surface on which the leading vehicle is traveling, that is, the road surface roughness prediction model is obtained by training the neural network model with a large amount of training data, and the roughness classification and recognition result of the road surface on which the leading vehicle is traveling is determined according to the posture signal of the leading vehicle through the road surface roughness prediction model, which solves the problem in the prior art that the vehicle's own forward visual sensor field of view is blocked, resulting in the inability to perceive the road surface condition ahead, increases the accuracy and robustness of the road surface roughness prediction model, and improves the accuracy of identifying the road surface roughness ahead.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1a A flow chart of a method for classifying and identifying road surface roughness provided in the first embodiment of the present invention;

[0022] Figure 1b This is an application scenario diagram of a road surface roughness classification and recognition method provided in the first embodiment of the present invention;

[0023] Figure 2 A flowchart of another method for classifying and identifying road surface roughness provided in the second embodiment of the present invention;

[0024] Figure 3 A schematic diagram of the structure of a device for classifying and identifying road surface roughness provided in Embodiment 3 of the present invention;

[0025] Figure 4 It is a schematic diagram of the structure of a vehicle for implementing the method for classifying and identifying road surface roughness according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] Embodiment 1

[0029] Figure 1a This is a flow chart of a method for classifying and identifying road surface roughness provided in the first embodiment of the present invention. This embodiment is applicable to the situation where a vehicle ahead blocks the road surface on the predicted future driving trajectory of the vehicle. The method can be executed by a device for classifying and identifying road surface roughness. The device for classifying and identifying road surface roughness can be implemented in the form of hardware and / or software. The device for classifying and identifying road surface roughness can be configured in the main controller of the vehicle. Figure 1a As shown, the method includes:

[0030] S110, obtaining at least one set of training data sequences, and training a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road roughness prediction model.

[0031] Under urban road conditions, the vehicle's forward visual sensor field of view is frequently occupied by other vehicles on the road, making it unable to perceive the road conditions ahead, which in turn causes the existing road roughness recognition function to fail. Based on this, when the vehicle's forward visual sensor field of view is occupied, the road roughness can be judged through the posture information of the leading vehicle.

[0032] Among them, at least one set of training data sequences can be obtained from data collected from a large number of user vehicles during actual driving, and each set of training data sequences can include the posture data of the front vehicle and the matching road surface roughness label data, for example, the posture data of the front vehicle at a certain time point and the roughness label of the road surface on which the front vehicle is driving at the same time. The preset neural network model can be an RNN (Recurrent Neural Network) suitable for the calculation of time series signals, such as LSTM (Long Short-Term Memory); the preset neural network model can also be CNN (Convolutional Neural Networks). The road surface roughness prediction model can be a model obtained by training a preset neural network model using at least one set of training data sequences.

[0033] In this embodiment, at least one set of training data sequences can be obtained in advance, and at least one round of training is performed on the preset neural network model based on multiple sets of training data sequences to obtain a trained model with enhanced accuracy and robustness as a road roughness prediction model.

[0034] S120: After a leading vehicle blocks the vehicle's future predicted driving trajectory, real-time position and posture data of the leading vehicle is collected.

[0035] In this embodiment, reference Figure 1bWhen the vehicle is driving on the road, the future driving trajectory of the vehicle can be predicted based on the steering wheel angle at this time. If the viewing angle of the vehicle's forward-looking sensor is blocked by the vehicle in front on the predicted future driving trajectory, it means that the forward-looking sensor cannot identify the unevenness of the road surface of the vehicle at the future moment, and the vehicle will pass through the road surface under the wheels of the vehicle in front at this moment in the future. After that, the real-time posture data of the vehicle in front can be collected, and the unevenness of the road surface under the front wheels can be identified by analyzing the posture data of the vehicle in front.

[0036] S130, inputting the real-time posture data into the road surface roughness prediction model to determine an roughness classification recognition result that matches the road surface on which the preceding vehicle is traveling.

[0037] The roughness classification recognition result may include the roughness level classification and the confidence of each level. The roughness level classification may include, for example, slight, general, severe and other categories.

[0038] In this embodiment, the collected posture data of the preceding vehicle may be input into a pre-trained road surface roughness prediction model to obtain roughness grade classification results matching the road surface on which the preceding vehicle is traveling and the confidence level of each grade.

[0039] Optionally, after determining the roughness classification result of the road surface on which the preceding vehicle is traveling, input signals can be provided to other control systems of the vehicle according to the vehicle's own vehicle condition to improve the comfort and safety of vehicle driving.

[0040] The technical solution of the embodiment of the present invention obtains at least one set of training data sequences, and trains a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model; after a leading vehicle appears to be blocked on the predicted future driving trajectory of the vehicle, the real-time posture data of the leading vehicle is collected; the real-time posture data is input into the road surface roughness prediction model to determine the roughness classification and recognition result matching the road surface on which the leading vehicle is traveling, that is, the road surface roughness prediction model is obtained by training the neural network model with a large amount of training data, and the roughness classification and recognition result of the road surface on which the leading vehicle is traveling is determined according to the posture signal of the leading vehicle through the road surface roughness prediction model, which solves the problem in the prior art that the vehicle's own forward visual sensor field of view is blocked, resulting in the inability to perceive the road surface condition ahead, increases the accuracy and robustness of the road surface roughness prediction model, and improves the accuracy of identifying the road surface roughness ahead.

[0041] Embodiment 2

[0042] Figure 2 This is a flow chart of another method for classifying and identifying road surface roughness provided in the second embodiment of the present invention. Based on the above embodiments, this embodiment refines the method by obtaining at least one set of training data sequences. Figure 2 As shown, the method includes:

[0043] S210. When the target preceding vehicle blocks the target vehicle's predicted future driving trajectory, a first time point and a target position of the target preceding vehicle are recorded, and real-time position and posture data of the target preceding vehicle are collected after the first time point.

[0044] The first time point may refer to the time point at which the forward-looking sensor of the target vehicle detects the target vehicle ahead on the future driving prediction trajectory. The target position may refer to the actual position of the target vehicle ahead at the first time point.

[0045] In this embodiment, in order to train the preset neural network model, at least one set of training data sequences may be obtained in advance. Taking obtaining a set of training data sequences as an example: when the target vehicle ahead is blocked on the predicted future driving trajectory of the target vehicle, the first time point and the target position of the target vehicle ahead are recorded, and the real-time position data of the target vehicle ahead are collected after the first time point.

[0046] S220: When it is preliminarily determined that the target road surface where the target vehicle is located is severely uneven based on the real-time position data of the target vehicle in front, it is determined whether the target vehicle will travel to the target position according to the future predicted travel trajectory after the first time point.

[0047] The real-time position and posture data of the target vehicle in front may include the real-time vehicle speed and the real-time roll angle. The real-time position and posture data of the target vehicle in front may be collected by sensors configured on the target vehicle.

[0048] Due to differences in body size, suspension system performance, etc. between vehicles, different vehicles have different body posture responses on the same roughness level of the road. Therefore, it is impossible to directly judge the road roughness based on the posture data of the preceding vehicle. The vehicle must pass the road ahead and use its own dynamic response signal to judge the specific road roughness level.

[0049] In this embodiment, after preliminarily determining that the target road surface on which the target vehicle is located is severely uneven based on the real-time vehicle speed and real-time roll angle of the target vehicle in front, it is also necessary to further determine whether the target vehicle travels to the target position according to the future predicted driving trajectory after the first time point.

[0050] In an optional embodiment, preliminarily determining that the target road surface where the target front vehicle is located is severely uneven based on the real-time posture data may include: obtaining the current vehicle speed and the current roll angle of the target front vehicle; calculating the current roll variance value of the target front vehicle based on the current vehicle speed and the current roll angle; and determining that the target road surface where the target front vehicle is located at the current moment is severely uneven based on the current roll variance value and a preset variance threshold.

[0051] Optionally, the current roll variance value of the target leading vehicle is calculated based on the current vehicle speed and the current roll angle, and the required quantity of historical target leading vehicle posture data at the current moment can be calculated based on the current vehicle speed, a preset acquisition period and a preset unevenness judgment interval length; based on the required quantity, a corresponding quantity of historical target leading vehicle roll angles are obtained; based on the required quantity and each of the historical target leading vehicle roll angles, the current roll variance value of the target leading vehicle is calculated.

[0052] Among them, the preset collection period may refer to the collection period of the target vehicle posture data. The preset roughness judgment interval may refer to the length of the road roughness judgment. For example, the roughness judgment interval may be predetermined based on the body length, shockproof performance and other information of the target vehicle. For example, if the target vehicle body is 5 meters long, the roughness judgment interval may be preset to 5 meters. The number of historical target vehicle posture data required at the current moment may refer to the number of target vehicle posture data required collected before the current moment.

[0053] Specifically, the current roll variance value can be obtained by Calculate, where S 2 Indicates the fluctuation degree of the body roll of the target vehicle in front during a period of time before the current moment. represents the roll angle of the target vehicle at different times, n represents the number of sampling points, represents the length of time intercepted before the current time, and represents the number of required historical target vehicle posture data. In order to spatially correspond the posture data of the vehicle in front to the actual road roughness, the number of sampling points should change with the vehicle speed, which can be specifically obtained by OK, where L c is the preset roughness judgment interval length, which is a calibration quantity; t s is the preset sampling period of the target vehicle’s posture data, u f is the speed of the target vehicle ahead.

[0054] Optionally, based on the current roll variance value and the preset variance threshold, it is determined that the target road surface where the target front vehicle is located at the current moment is seriously uneven, and it can be judged whether the current roll variance value is greater than the preset variance threshold; when the current roll variance value is greater than the preset variance threshold, it is determined that the target road surface where the target front vehicle is located at the current moment is seriously uneven.

[0055] Due to the limitation of the sensor installation height of the target vehicle, the roll angle data in the posture data of the target front parking space has more obvious characteristics and higher accuracy. Therefore, the variance of the roll angle data can be used as the basis for judging the road surface roughness under the target front wheel. If the calculated roll variance value S of the target front vehicle is 2 If S is greater than the preset variance threshold, it can be considered that the target vehicle is passing through a road with a large roughness. 2If the variance is not greater than the preset threshold, the position data of the target vehicle in front can be discarded, that is, it can no longer be used as a sampling object.

[0056] In another optional implementation, determining whether the target vehicle travels to the target location according to the future predicted travel trajectory after the first time point may include:

[0057] When the target preceding vehicle blocks the target vehicle's predicted future driving trajectory, obtain the predicted length of the target vehicle and the target preceding vehicle on the predicted future driving trajectory, and the current speed of the target vehicle; obtain the predicted time for the target vehicle to travel to the target position according to the predicted future driving trajectory based on the predicted length and the current speed, and obtain the second time point at which the target vehicle is predicted to travel to the target position; collect the first yaw angular velocity of the target vehicle at the first time point, the second yaw angular velocity of the target vehicle actually traveling to the second time point, and collect the yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the first time point and the second yaw angular velocity of the target vehicle at the a real-time vehicle speed between the first time point and the second time point; calculating the actual driving length of the target vehicle between the first time point and the second time point according to the real-time vehicle speed, the first time point and the second time point; calculating the yaw angular velocity integral value of the target vehicle between the first time point and the second time point according to the first yaw acceleration, the second yaw angular velocity, the first time point and the second time point; judging whether the target vehicle travels to the target position according to the future driving predicted trajectory after the first time point according to the actual driving length, the predicted length, the yaw angular velocity integral value and a preset integral threshold value.

[0058] refer to Figure 1b , the predicted length may refer to S, which may be obtained, for example, by a radar sensor or a laser sensor mounted on the target vehicle. The predicted duration may refer to the duration predicted for the target vehicle to travel S at its current speed. The second time point may refer to the predicted time point when the target vehicle travels to the target position of the target preceding vehicle at S.

[0059] Specifically, the actual travel distance of the target vehicle between the first time point and the second time point can be obtained by Among them, t1 represents the first time point, t2 represents the second time point, and u0 represents the actual speed of the target vehicle. This formula can be understood as accumulating the actual speed of the target vehicle between the first time point and the second time point.

[0060] The yaw rate integral value of the target vehicle between the first time point and the second time point can be obtained by Among them, w yaw represents the second yaw rate of the target vehicle, Indicates the first yaw rate of the target vehicle.

[0061] Based on the above implementation, when the first difference between the actual driving length and the predicted length does not exceed the preset difference threshold, and the yaw angular velocity integral value is less than the preset integral threshold, it is determined that the target vehicle travels to the target position according to the future driving predicted trajectory after the first time point.

[0062] Specifically, a first difference between the actual travel length and the predicted length of the target vehicle may be calculated, and when the first difference does not exceed a preset difference threshold, and the yaw angular velocity integral value is less than a preset integral threshold, it is determined that the target vehicle travels to the target position according to the future travel prediction trajectory after the first time point. Correspondingly, if the first difference exceeds the preset difference threshold, and / or the yaw angular velocity integral value is not less than the preset integral threshold, it may be determined that the target vehicle does not travel to the target position according to the future travel prediction trajectory after the first time point, and the target vehicle may not continue to be a sampling object.

[0063] S230: If yes, after the target vehicle travels to the target position according to the future travel prediction trajectory, a real-time dynamic response signal of the target vehicle is collected.

[0064] Among them, the dynamic response signal may include a suspension vertical acceleration signal, a sprung mass vertical acceleration signal, a vehicle body pitch angular velocity signal, a vehicle body roll angle signal, etc., which may be collected by the sensors configured by the vehicle itself.

[0065] S240, determining again the roughness classification and recognition result of the target road surface according to the real-time dynamic response signal of the target vehicle, and using the roughness classification and recognition result as a label data sequence matching the target road surface.

[0066] In this embodiment, the roughness classification and recognition result of the target road surface on which the target preceding vehicle previously traveled can be determined again according to the fluctuation of the real-time dynamic response signal of the target vehicle, so that the roughness classification and recognition result is used as a label data sequence matching the target road surface.

[0067] S250, aligning the label data sequence with the real-time position data of the target front vehicle in time series to obtain an alignment result, and using the alignment result as a set of training data sequences.

[0068] In this embodiment, since there is a distance difference between the target vehicle and the target vehicle in front, there is a difference in timing between the label data sequence obtained according to the dynamic response signal of the target vehicle and the real-time posture data of the target vehicle in front. Therefore, the label data sequence can be aligned with the real-time posture data of the target vehicle in front in timing, and the alignment result can be used as a set of training data sequences.

[0069] The benefit of such setting of S210-S250 operation is that it can improve the authenticity of the training data set, thereby increasing the robustness of the road roughness prediction model algorithm.

[0070] S260, determining whether the number of training data sequences meets a preset condition, if so, obtaining at least one set of training data sequences to execute S270, otherwise, returning to execute S210.

[0071] The preset condition may refer to a preset quantity requirement of the training data sequence, for example, several hundred or several thousand groups.

[0072] In this embodiment, the number of training data sequences is multiple groups, that is, at least one group of training data sequences needs to be obtained according to multiple groups of target vehicles and target preceding vehicles. After obtaining a group of training data sequences, it can be determined whether the number of training data sequences meets the preset condition. If so, at least one group of training data sequences is obtained. If not, the operation of S210-S250 is returned and re-executed until the training data sequence meets the preset condition.

[0073] S270. According to the at least one set of training data sequences, a preset neural network model is trained for at least one round to obtain a road roughness prediction model.

[0074] S280: After a leading vehicle blocks the vehicle's future predicted driving trajectory, real-time position and posture data of the leading vehicle is collected.

[0075] S290: Input the real-time posture data into the road surface roughness prediction model to determine an roughness classification recognition result that matches the road surface on which the preceding vehicle is traveling.

[0076] The technical solution of the embodiment of the present invention is to initially judge the roughness of the target road surface by collecting the posture data of the target front vehicle in actual driving, and when the target vehicle travels to the target road surface on which the target front vehicle travels according to the future driving prediction trajectory, the roughness of the target road surface is judged again by collecting the dynamic response signal of the target vehicle, and at least one set of training data sequences is obtained according to the posture data of the target front vehicle and the roughness label sequence of the target road surface finally obtained, and a preset neural network model is trained for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model; after the front vehicle is blocked on the vehicle's future driving prediction trajectory, Collect the real-time posture data of the leading vehicle; input the real-time posture data into the road surface roughness prediction model to determine the roughness classification and recognition result that matches the road surface on which the leading vehicle is traveling, that is, train the neural network model with a large amount of training data to obtain the road surface roughness prediction model, and determine the roughness classification and recognition result of the road surface on which the leading vehicle is traveling according to the posture signal of the leading vehicle through the road surface roughness prediction model, thereby solving the problem in the prior art that the vehicle's own forward visual sensor field of view is blocked, resulting in the inability to perceive the road surface condition ahead, increasing the accuracy and robustness of the road surface roughness prediction model, and improving the accuracy of identifying the roughness of the road ahead.

[0077] Embodiment 3

[0078] Figure 3 This is a schematic diagram of the structure of a road surface roughness classification and identification device provided in the third embodiment of the present invention. Figure 3 As shown, the device includes: a road roughness prediction model acquisition module 310, a posture data acquisition module 320 and a classification recognition result determination module 330. Among them:

[0079] A road roughness prediction model acquisition module 310 is used to acquire at least one set of training data sequences, and to train a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road roughness prediction model;

[0080] The posture data acquisition module 320 is used to collect the real-time posture data of the preceding vehicle after the preceding vehicle blocks the vehicle's future travel trajectory;

[0081] The classification recognition result determination module 330 is used to input the real-time posture data into the road surface roughness prediction model to determine the roughness classification recognition result that matches the road surface on which the preceding vehicle is traveling.

[0082] The technical solution of the embodiment of the present invention obtains at least one set of training data sequences, and trains a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model; after a leading vehicle appears to be blocked on the predicted future driving trajectory of the vehicle, the real-time posture data of the leading vehicle is collected; the real-time posture data is input into the road surface roughness prediction model to determine the roughness classification and recognition result matching the road surface on which the leading vehicle is traveling, that is, the road surface roughness prediction model is obtained by training the neural network model with a large amount of training data, and the roughness classification and recognition result of the road surface on which the leading vehicle is traveling is determined according to the posture signal of the leading vehicle through the road surface roughness prediction model, which solves the problem in the prior art that the vehicle's own forward visual sensor field of view is blocked, resulting in the inability to perceive the road surface condition ahead, increases the accuracy and robustness of the road surface roughness prediction model, and improves the accuracy of identifying the road surface roughness ahead.

[0083] Optionally, the road roughness prediction model acquisition module 310 includes:

[0084] Perform the following operations for a set of target vehicles and target preceding vehicles:

[0085] A real-time posture data acquisition unit, configured to record a first time point and a target position of the target vehicle in front when the target vehicle in front is blocked on the predicted future travel trajectory of the target vehicle, and to acquire real-time posture data of the target vehicle in front after the first time point;

[0086] A target vehicle driving position determination unit, configured to determine whether the target vehicle drives to the target position according to the future driving prediction trajectory after the first time point after preliminarily determining that the target road surface where the target vehicle is located is seriously uneven according to the real-time position data of the target vehicle in front;

[0087] A real-time dynamic response signal acquisition unit, configured to acquire a real-time dynamic response signal of the target vehicle after the target vehicle travels to the target position according to the future travel prediction trajectory;

[0088] a label data sequence acquisition unit, used to re-determine the roughness classification recognition result of the target road surface according to the real-time dynamic response signal of the target vehicle, and use the roughness classification recognition result as the label data sequence matching the target road surface;

[0089] The training data sequence acquisition unit is used to align the label data sequence with the real-time posture data of the target front vehicle in time sequence to obtain an alignment result, and use the alignment result as a set of training data sequences.

[0090] Optionally, the real-time position and posture data of the target front vehicle includes real-time vehicle speed and real-time roll angle;

[0091] Correspondingly, the target vehicle driving position determination unit includes:

[0092] A vehicle speed and roll angle acquisition subunit, used to acquire the current vehicle speed and current roll angle of the target front vehicle;

[0093] A roll variance value calculation subunit, configured to calculate a current roll variance value of the target front vehicle according to the current vehicle speed and the current roll angle;

[0094] The target road surface serious unevenness determination subunit is used to determine the target road surface serious unevenness where the target front vehicle is located at the current moment according to the current roll variance value and a preset variance threshold.

[0095] Optionally, the roll variance value calculation subunit may be used to:

[0096] Calculate the required amount of historical target vehicle posture data at the current moment according to the current vehicle speed, the preset collection cycle and the preset roughness judgment interval length;

[0097] According to the required quantity, obtaining a corresponding number of historical target leading vehicle roll angles;

[0098] The current roll variance value of the target preceding vehicle is calculated according to the required quantity and each of the historical target preceding vehicle roll angles.

[0099] Optionally, the target road surface serious unevenness determination subunit may be used to:

[0100] Determining whether the current roll variance value is greater than the preset variance threshold;

[0101] When the current roll variance value is greater than the preset variance threshold, it is determined that the target road surface where the target front vehicle is located at the current moment is seriously uneven.

[0102] Optionally, the target vehicle driving position determination unit further includes:

[0103] The current speed acquisition subunit of the target vehicle is used to acquire the predicted length of the target vehicle and the target vehicle in front on the predicted future travel trajectory, and the current speed of the target vehicle when the target vehicle in front blocks the predicted future travel trajectory of the target vehicle;

[0104] A second time point acquisition subunit is used to acquire the predicted time length for the target vehicle to travel to the target position according to the future travel prediction trajectory according to the predicted length and the current vehicle speed, and acquire the second time point when the target vehicle is predicted to travel to the target position;

[0105] A real-time speed collection subunit of the target vehicle, used to collect a first yaw angular velocity of the target vehicle at the first time point, a second yaw angular velocity of the target vehicle when the target vehicle actually travels to a second time point, and collect a real-time speed of the target vehicle between the first time point and the second time point;

[0106] an actual travel length calculation subunit, configured to calculate an actual travel length of the target vehicle between the first time point and the second time point according to the real-time vehicle speed, the first time point and the second time point;

[0107] a yaw rate integral value calculation subunit, configured to calculate a yaw rate integral value of the target vehicle between the first time point and the second time point according to the first yaw acceleration, the second yaw rate, the first time point and the second time point;

[0108] The target vehicle driving position judgment subunit is used to judge whether the target vehicle travels to the target position according to the future driving prediction trajectory after the first time point based on the actual driving length, the predicted length, the yaw angular velocity integral value and the preset integral threshold.

[0109] Optionally, the target vehicle driving position determination subunit may be used to:

[0110] When a first difference between the actual driving length and the predicted length does not exceed a preset difference threshold, and the yaw angular velocity integral value is less than the preset integral threshold, it is determined that the target vehicle travels to the target position according to the future driving predicted trajectory after the first time point.

[0111] The device for classifying and identifying road surface roughness provided in the embodiment of the present invention can execute the method for classifying and identifying road surface roughness provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0112] Embodiment 4

[0113] Figure 4 A schematic diagram of a vehicle 400 that can be used to implement an embodiment of the present invention is shown. The vehicle is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The vehicle can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0114] like Figure 4 As shown, the vehicle 400 includes at least one processor 401, and a memory connected to the at least one processor 401 in communication, such as a read-only memory (ROM) 402, a random access memory (RAM) 403, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 401 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 402 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the vehicle 400 can also be stored. The processor 401, the ROM 402, and the RAM 403 are connected to each other through the bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0115] A number of components in the vehicle 400 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the vehicle 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0116] The processor 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 401 executes the various methods and processes described above, such as a classification and recognition method for road roughness.

[0117] In some embodiments, the method for classifying and identifying road surface roughness may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the vehicle 400 via the ROM 402 and / or the communication unit 19. When the computer program is loaded into the RAM 403 and executed by the processor 401, one or more steps of the method for classifying and identifying road surface roughness described above may be performed. Alternatively, in other embodiments, the processor 401 may be configured to execute the method for classifying and identifying road surface roughness in any other appropriate manner (e.g., by means of firmware).

[0118] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0119] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0120] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0121] To provide interaction with a user, the systems and techniques described herein can be implemented on a vehicle having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the vehicle. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0122] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0123] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0124] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0125] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for classifying and identifying road roughness, characterized in that: include: Acquire at least one set of training data sequences, and train a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road roughness prediction model; After a leading vehicle blocks the vehicle's future predicted driving trajectory, real-time position and posture data of the leading vehicle is collected; Inputting the real-time posture data into the road surface roughness prediction model to determine the roughness classification recognition result matching the road surface on which the preceding vehicle is traveling; Wherein, obtaining at least one set of training data sequences includes: Perform the following operations for a set of target vehicles and target preceding vehicles: When the target preceding vehicle blocks the target vehicle's predicted future driving trajectory, a first time point and a target position of the target preceding vehicle are recorded, and real-time position and posture data of the target preceding vehicle are collected after the first time point; Preliminarily determining that the target road surface where the target vehicle is located is severely uneven based on the real-time position data of the target vehicle ahead, and judging whether the target vehicle travels to the target position according to the future travel prediction trajectory after the first time point; If so, after the target vehicle travels to the target position according to the future travel prediction trajectory, collecting a real-time dynamic response signal of the target vehicle; Determining again the roughness classification and recognition result of the target road surface according to the real-time dynamic response signal of the target vehicle, and using the roughness classification and recognition result as a label data sequence matching the target road surface; The label data sequence is aligned with the real-time position data of the target front vehicle in time sequence to obtain an alignment result, and the alignment result is used as a set of training data sequences.

2. The method according to claim 1, characterized in that The real-time position data of the target front vehicle includes real-time vehicle speed and real-time roll angle; Accordingly, preliminarily determining that the target road surface on which the target front vehicle is located is seriously uneven according to the real-time position and posture data includes: Obtaining the current speed and current roll angle of the target front vehicle; Calculating a current roll variance value of the target front vehicle according to the current vehicle speed and the current roll angle; According to the current roll variance value and a preset variance threshold, it is determined that the target road surface where the target front vehicle is located at the current moment is seriously uneven.

3. The method according to claim 2, characterized in that Calculating the current roll variance value of the target front vehicle according to the current vehicle speed and the current roll angle includes: Calculate the required amount of historical target vehicle posture data at the current moment according to the current vehicle speed, the preset collection cycle and the preset roughness judgment interval length; According to the required quantity, obtaining a corresponding number of historical target leading vehicle roll angles; The current roll variance value of the target preceding vehicle is calculated according to the required quantity and each of the historical target preceding vehicle roll angles.

4. The method according to claim 2, characterized in that: Determining, according to the current roll variance value and a preset variance threshold, that the target road surface where the target front vehicle is located at the current moment is seriously uneven includes: Determining whether the current roll variance value is greater than the preset variance threshold; When the current roll variance value is greater than the preset variance threshold, it is determined that the target road surface where the target front vehicle is located at the current moment is seriously uneven.

5. The method according to claim 1, characterized in that Determining whether the target vehicle travels to the target position according to the future travel prediction trajectory after the first time point includes: When the target preceding vehicle blocks the target vehicle's future travel prediction trajectory, obtaining the predicted length of the target vehicle and the target preceding vehicle on the future travel prediction trajectory, and the current speed of the target vehicle; According to the predicted length and the current vehicle speed, the predicted time length for the target vehicle to travel to the target location according to the predicted future travel trajectory is obtained, and a second time point when the target vehicle is predicted to travel to the target location is obtained; Collecting a first yaw angular velocity of the target vehicle at the first time point, a second yaw angular velocity of the target vehicle when the target vehicle actually travels to a second time point, and collecting a real-time vehicle speed of the target vehicle between the first time point and the second time point; Calculate the actual travel distance of the target vehicle between the first time point and the second time point according to the real-time vehicle speed, the first time point and the second time point; Calculating an integral value of a yaw rate of the target vehicle between the first time point and the second time point according to the first yaw acceleration, the second yaw rate, the first time point, and the second time point; According to the actual driving length, the predicted length, the yaw angular velocity integral value and a preset integral threshold, it is determined whether the target vehicle travels to the target position according to the future driving predicted trajectory after the first time point.

6. The method according to claim 5, characterized in that According to the actual travel length, the predicted length, the yaw angular velocity integral value and the preset integral threshold, determining whether the target vehicle travels to the target position according to the future travel prediction trajectory after the first time point includes: When a first difference between the actual driving length and the predicted length does not exceed a preset difference threshold, and the yaw angular velocity integral value is less than the preset integral threshold, it is determined that the target vehicle travels to the target position according to the future driving predicted trajectory after the first time point.

7. A device for classifying and identifying road surface roughness, characterized in that: include: A road surface roughness prediction model acquisition module is used to acquire at least one set of training data sequences, and to train a preset neural network model for at least one round according to the at least one set of training data sequences to obtain a road surface roughness prediction model; A posture data acquisition module is used to collect real-time posture data of the preceding vehicle after the preceding vehicle blocks the vehicle's future travel prediction trajectory; A classification recognition result determination module, used for inputting the real-time posture data into the road surface roughness prediction model to determine the roughness classification recognition result matching the road surface on which the preceding vehicle is traveling; Among them, the road roughness prediction model acquisition module includes: Perform the following operations for a set of target vehicles and target preceding vehicles: A real-time posture data acquisition unit, configured to record a first time point and a target position of the target vehicle in front when the target vehicle in front is blocked on the predicted future travel trajectory of the target vehicle, and to acquire real-time posture data of the target vehicle in front after the first time point; A target vehicle driving position determination unit, configured to determine whether the target vehicle drives to the target position according to the future driving prediction trajectory after the first time point after preliminarily determining that the target road surface where the target vehicle is located is seriously uneven according to the real-time position data of the target vehicle in front; A real-time dynamic response signal acquisition unit, configured to acquire a real-time dynamic response signal of the target vehicle after the target vehicle travels to the target position according to the future travel prediction trajectory; a label data sequence acquisition unit, used to re-determine the roughness classification recognition result of the target road surface according to the real-time dynamic response signal of the target vehicle, and use the roughness classification recognition result as the label data sequence matching the target road surface; The training data sequence acquisition unit is used to align the label data sequence with the real-time posture data of the target front vehicle in time sequence to obtain an alignment result, and use the alignment result as a set of training data sequences.

8. A vehicle, characterized in that: The vehicle comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method for classifying and identifying road surface roughness according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for classifying and identifying road surface roughness according to any one of claims 1 to 6 when executed.

Images