Method, device, equipment and medium for identifying dangerous behaviors of cleaning vehicles

Abstract

The present invention discloses a method, device, equipment and medium for identifying dangerous behaviors of a cleaning vehicle. The method comprises: obtaining driving status information and vehicle posture information of the cleaning vehicle in real time; performing waypoint search on the road where the cleaning vehicle is located according to multi-frame vehicle posture information, and determining the waypoint direction vector corresponding to the cleaning vehicle in the current frame; determining the vehicle head steering angular acceleration and steering wheel rotation angular acceleration corresponding to the cleaning vehicle in the current frame based on the driving status information and waypoint direction vectors of two adjacent frames; determining the behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration, in combination with a preset speed reference value; comparing the behavior score with a preset score threshold to determine whether the cleaning vehicle exhibits dangerous behaviors in the current frame, thereby more accurately identifying the cleaning vehicle's own dangerous behaviors.

Description

Technical Field

[0001] The present invention relates to the field of behavior recognition technology, and in particular to a method, device, equipment and medium for identifying dangerous behaviors of a cleaning vehicle. Background Art

[0002] As the pace of urbanization accelerates, road cleaning is also facing tremendous pressure. Traditional solutions usually rely on sanitation workers to clean the roads or cleaning vehicles to clean the roads.

[0003] With the continuous development of unmanned driving technology, vehicles are becoming more and more information-based and intelligent. Cleaning vehicles are easier to drive unmanned due to their slow speed and fixed routes. However, since cleaning vehicles are usually large vehicles, it is more important to pay attention to whether dangerous behaviors occur during driving compared to ordinary vehicles.

[0004] However, in the existing technology, dangerous behaviors of vehicles are mainly identified and judged by estimating the collision time based on the distance from other vehicles or the distance from the road shoulder, which makes it difficult to accurately identify the dangerous behaviors of the cleaning vehicle itself. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for identifying dangerous behaviors of cleaning vehicles, which solves the technical problem in the prior art that dangerous behaviors of vehicles are mainly identified and judged by estimating the collision time based on the distance from other vehicles or the distance from the road shoulder, making it difficult to accurately identify the dangerous behaviors of the cleaning vehicle itself.

[0006] A first aspect of the present invention provides a method for identifying dangerous behaviors of a cleaning vehicle, comprising:

[0007] Obtain the driving status information and vehicle posture information of the cleaning vehicle in real time;

[0008] Performing a waypoint search on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames, and determining the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame;

[0009] Determine the head steering angular acceleration and steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame according to the driving state information and the waypoint direction vector of two adjacent frames;

[0010] Determining a behavior score of the cleaning vehicle corresponding to the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration and a preset speed reference value;

[0011] The behavior score is compared with a preset score threshold to determine whether the cleaning vehicle has exhibited a dangerous behavior in the current frame.

[0012] Optionally, the vehicle posture information includes a vehicle position and a vehicle orientation; the step of performing a waypoint search on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames and determining a direction vector of the waypoint corresponding to the cleaning vehicle in the current frame includes:

[0013] Performing a waypoint search on the road where the cleaning vehicle is located with the vehicle position in the current frame as the center, and determining a plurality of initial waypoints within a to-be-searched area with a preset radius;

[0014] Determining a target waypoint corresponding to the cleaning vehicle in the current frame based on a comparison result of the vehicle orientation, the initial waypoint, and a plurality of historical candidate waypoints obtained from a preset waypoint database;

[0015] The vector of the target waypoint in the world coordinate system is determined as the waypoint direction vector corresponding to the cleaning vehicle in the current frame.

[0016] Optionally, the step of determining a target waypoint corresponding to the cleaning vehicle in the current frame based on a comparison result of the vehicle orientation, the initial waypoint, and a plurality of historical candidate waypoints obtained from a preset waypoint database includes:

[0017] Calculating the difference between the vehicle orientation and each of the initial waypoints;

[0018] Selecting a first preset number of the initial waypoints as candidate waypoints according to the direction differences from smallest to largest;

[0019] Acquire a plurality of candidate frames before the current frame from a preset waypoint database according to a second preset number, and acquire a plurality of historical candidate waypoints corresponding to each of the candidate frames;

[0020] The candidate waypoints are compared with the historical candidate waypoints, and the candidate waypoint with the largest number of overlaps and closest to the vehicle position is selected as the target waypoint corresponding to the cleaning vehicle in the current frame.

[0021] Optionally, the driving state information includes a vehicle angle vector and a steering wheel rotation speed; the step of determining the vehicle head steering angular acceleration and the steering wheel rotation angular acceleration corresponding to the current frame of the cleaning vehicle based on the driving state information and the waypoint direction vector corresponding to two adjacent frames includes:

[0022] Calculating a time difference between a historical frame adjacent to the current frame and the current frame;

[0023] Determine the vehicle head steering angular acceleration corresponding to the current frame of the cleaning vehicle according to the vehicle angle vector and the waypoint direction vector of the historical frame and the current frame;

[0024] Calculating a rotation speed difference between the steering wheel rotation speed of the historical frame and the steering wheel rotation speed of the current frame;

[0025] The ratio of the rotation speed difference to the time difference is determined as the steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame.

[0026] Optionally, the step of determining the head steering angular acceleration of the cleaning vehicle corresponding to the current frame based on the vehicle angle vector and the waypoint direction vector of the historical frame and the current frame includes:

[0027] Calculating a vector dot product between the vehicle angle vector and the waypoint direction vector;

[0028] Calculating a vector length product between the vector length of the vehicle angle vector and the vector length of the waypoint direction vector;

[0029] Calculating a first ratio between the vector dot product and the vector length product, and using the arc cosine value of the first ratio as a first angle value between the cleaning vehicle and the road;

[0030] Obtaining a first angle value corresponding to the historical frame as a second angle value;

[0031] Calculating an angle difference between the second angle value and the first angle value;

[0032] The ratio between the angle difference and the time difference is determined as the vehicle head steering angular acceleration corresponding to the cleaning vehicle in the current frame.

[0033] Optionally, the speed reference value includes a steering angular acceleration reference value and a rotation angular acceleration reference value; and the step of determining the behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with a preset speed reference value includes:

[0034] determining a vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value;

[0035] determining a steering wheel rotation score based on the steering wheel rotation angular acceleration and the rotation angular acceleration reference value;

[0036] The multiplication value between the vehicle steering score and the steering wheel rotation score is calculated to obtain the behavior score of the cleaning vehicle corresponding to the current frame.

[0037] Optionally, the steering angular acceleration reference value includes a full-score steering angular acceleration value and a zero-score steering angular acceleration value; and the step of determining the vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value includes:

[0038] calculating a first difference between the steering angular acceleration and a full score value of the steering angular acceleration;

[0039] calculating a second difference between the zero-score value of the steering angular acceleration and the full-score value of the steering angular acceleration;

[0040] calculating a second ratio between the first difference and the second difference;

[0041] A difference between a preset constant and the second ratio is determined as a vehicle steering score.

[0042] Optionally, the angular acceleration reference value includes a full score value of the steering wheel angular acceleration and a zero score value of the steering wheel angular acceleration; and the step of determining the steering wheel rotation score based on the steering wheel angular acceleration and the angular acceleration reference value includes:

[0043] calculating a third difference between the steering wheel angular acceleration and a full score value of the steering wheel angular acceleration;

[0044] calculating a fourth difference between the zero-score value of the steering wheel rotation angular acceleration and the full-score value of the steering wheel rotation angular acceleration;

[0045] calculating a third ratio between the third difference and the fourth difference;

[0046] A difference between the preset constant and the third ratio is determined as a steering wheel rotation score.

[0047] Optionally, the step of comparing the behavior score with a preset score threshold to determine whether the cleaning vehicle has engaged in a dangerous behavior in the current frame includes:

[0048] comparing the behavior score with a preset score threshold;

[0049] If the behavior score is greater than or equal to the score threshold, it is determined that the cleaning vehicle has not exhibited any dangerous behavior in the current frame, and the process jumps to executing the step of obtaining the driving state information and vehicle posture information of the cleaning vehicle in real time;

[0050] If the behavior score is less than the score threshold, it is determined that the cleaning vehicle has engaged in a dangerous behavior in the current frame, and a danger warning message is output.

[0051] Optionally, the method further includes:

[0052] When a simulation test request is received, the simulation candidate waypoints and the simulation vehicle position corresponding to each frame of the cleaning vehicle within a preset time period are obtained;

[0053] Selecting a simulation current frame frame by frame, and selecting multiple frames adjacent to the simulation current frame as simulation history frames;

[0054] Selecting the simulation candidate waypoint with the largest number of overlaps between all the simulation history frames and the simulation current frame as the simulation target waypoint corresponding to the simulation current frame;

[0055] Selecting multiple frames after the current simulation frame as future simulation frames;

[0056] Calculating the Euclidean distance between the simulated vehicle position corresponding to the simulated future frame and the simulated vehicle position corresponding to the simulated current frame for each frame;

[0057] Selecting the first simulated vehicle position whose Euclidean distance is greater than a preset distance threshold as the target vehicle position;

[0058] The simulated vehicle position and the target vehicle position are connected to construct a simulation angle vector corresponding to the cleaning vehicle in the current simulation frame.

[0059] Optionally, the method further includes:

[0060] When a program version comparison request is received, the current running program and the program to be put online corresponding to the cleaning vehicle are obtained;

[0061] Executing the currently running program based on the simulation target waypoint and the simulation angle vector to obtain a corresponding first behavior score;

[0062] Executing the program to be put online based on the simulation target waypoint and the simulation angle vector to obtain a corresponding second behavior score;

[0063] Calculating an absolute value of a version score difference between the second behavior score and the first behavior score;

[0064] If the absolute value is greater than a preset version score threshold, it is determined that there is improvement in the program to be put online;

[0065] If the absolute value is less than or equal to a preset version score threshold, it is determined that there is no improvement in the program to be put online.

[0066] A second aspect provides a dangerous behavior identification device for a cleaning vehicle, comprising:

[0067] Vehicle information acquisition module, used to obtain the driving status information and vehicle posture information of the cleaning vehicle in real time;

[0068] A waypoint direction search module is used to search for waypoints on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames, and determine the waypoint direction vector corresponding to the cleaning vehicle in the current frame;

[0069] an angular acceleration determination module, configured to determine the head steering angular acceleration and the steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame based on the driving state information and the waypoint direction vector of two adjacent frames;

[0070] a behavior score calculation module, configured to determine a behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with a preset speed reference value;

[0071] The dangerous behavior judgment module is used to compare the behavior score with a preset score threshold to determine whether the cleaning vehicle has performed a dangerous behavior in the current frame.

[0072] The third aspect provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the dangerous behavior identification method for a cleaning vehicle as described in any one of the first aspects of the present invention.

[0073] A fourth aspect provides a computer-readable storage medium having a computer program stored thereon, which, when executed, implements the dangerous behavior identification method for a cleaning vehicle as described in any one of the first aspects of the present invention.

[0074] It can be seen from the above technical solutions that the present invention has the following advantages:

[0075] The present invention obtains driving state information and vehicle posture information of a cleaning vehicle, then performs a waypoint search on the road segment information of the road where the cleaning vehicle is located based on the vehicle posture information to determine the waypoint direction vector corresponding to the cleaning vehicle in the current frame. The present invention further calculates the angular difference between the cleaning vehicle in the current frame and the historical frame based on the waypoint direction vectors of two adjacent frames combined with the vehicle angle vector within the driving state information. The angular difference is used to calculate the vehicle's head steering angular acceleration corresponding to the current frame, and the steering wheel angular acceleration is also calculated based on the driving state information. After obtaining the vehicle's head steering angular acceleration and steering wheel angular acceleration, the present invention determines the vehicle's behavior score corresponding to the current frame in combination with a preset speed reference value. Finally, the behavior score is compared with a preset score threshold, and the comparison result is used to determine whether the cleaning vehicle has engaged in dangerous behavior in the current frame. Thus, by calculating the vehicle behavior score based on the cleaning vehicle's driving state and vehicle posture as data, the vehicle's own dangerous behavior can be more accurately identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0077] Figure 1 A flowchart of a method for identifying dangerous behaviors of a cleaning vehicle provided in Example 1 of the present invention;

[0078] Figure 2 A flowchart of a method for identifying dangerous behaviors of a cleaning vehicle provided in a second embodiment of the present invention;

[0079] Figure 3 This is a structural block diagram of a dangerous behavior identification device for a cleaning vehicle provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0080] The embodiments of the present invention provide a method, device, equipment and medium for identifying dangerous behaviors of cleaning vehicles, which are used to solve the technical problem that in the prior art, dangerous behaviors of vehicles are mainly identified and judged by estimating the collision time based on the distance from other vehicles or the distance from the road shoulder, making it difficult to accurately identify the dangerous behaviors of the cleaning vehicle itself.

[0081] In order to make the purpose, features, and advantages of the present invention more obvious and easy to understand, the technical solutions 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 embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0082] See also Figure 1 , Figure 1 This is a flowchart of the steps of a method for identifying dangerous behaviors of a cleaning vehicle provided in Example 1 of the present invention.

[0083] The present invention provides a method for identifying dangerous behaviors of a cleaning vehicle, comprising:

[0084] Step 101: Acquire driving status information and vehicle posture information of the cleaning vehicle;

[0085] Driving state information refers to the vehicle angle vector and steering wheel angular acceleration when the cleaning vehicle undergoes vehicle direction change operations such as changing lanes, turning, or making a U-turn during driving.

[0086] Vehicle posture information refers to the vehicle position and orientation of the cleaning vehicle at each moment during its driving process.

[0087] In specific implementations, compared to bypassing obstacles, cleaning vehicles usually adopt a strategy of stopping and waiting. At the same time, since cleaning vehicles have slow steering speeds and large sizes, it is necessary to identify dangerous behaviors of cleaning vehicles when their vehicle angles change. To this end, the driving status information and vehicle posture information of the cleaning vehicle can be obtained as the data basis for subsequent dangerous behavior identification.

[0088] Step 102: searching for waypoints on the road where the cleaning vehicle is located according to the multi-frame vehicle posture information to determine the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame;

[0089] Road segment information refers to the road test data corresponding to the road on which the cleaning vehicle is traveling during driving, including but not limited to road waypoints, lane relationship information, surrounding obstacle information, and traffic light information.

[0090] In an embodiment of the present application, since the cleaning vehicle is in the process of driving, it is necessary to plan the driving route. At this time, according to the vehicle posture information corresponding to the current frame and the historical frame before the current frame, the road segment information of the road where the cleaning vehicle is located can be searched for waypoints based on the vehicle position and vehicle orientation to determine the target waypoint corresponding to the cleaning vehicle in the current frame, and further determine its vector in the world coordinate system as the waypoint direction vector corresponding to the cleaning vehicle in the current frame.

[0091] Step 103: determining the vehicle head turning angular acceleration and steering wheel rotation angular acceleration corresponding to the cleaning vehicle in the current frame based on the driving state information and the waypoint direction vector of two adjacent frames;

[0092] After obtaining the waypoint direction vector, we can further calculate the steering wheel angular acceleration between the historical frame and the current frame based on the driving state information corresponding to the current frame and the adjacent historical frame. At the same time, we use the waypoint direction vector between two adjacent frames to calculate the vehicle head steering angular acceleration between the historical frame and the current frame.

[0093] It should be noted that the steering wheel angular acceleration and the vehicle head turning angular acceleration can be used to measure whether the direction of the cleaning vehicle changes suddenly in the current frame, so as to further judge the safety in the current frame.

[0094] Step 104: Determine the behavior score of the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with a preset speed reference value;

[0095] The speed reference value refers to the full score value and zero score value corresponding to the steering angular acceleration and the steering wheel rotation angular acceleration, respectively, and is used as a reference value for scoring the behavior of the current frame when the cleaning vehicle is driving.

[0096] After obtaining the steering angular acceleration and the steering wheel rotation angular acceleration, they are further used as input data and combined with the preset speed reference value to calculate the behavior score of the cleaning vehicle in the current frame.

[0097] It should be noted that the vehicle steering score and the steering wheel rotation score are calculated using the steering angular acceleration and the steering wheel rotation angular acceleration respectively, and the two are further superimposed to obtain the behavior score.

[0098] Step 105 : Compare the behavior score with a preset score threshold to determine whether the cleaning vehicle has engaged in a dangerous behavior in the current frame.

[0099] In a specific implementation, the behavior score can be compared with a preset score threshold. If the behavior score is greater than or equal to the score threshold, it indicates that the cleaning vehicle is in a relatively safe state and can be determined to have not engaged in dangerous behavior in the current frame. If the behavior score is less than the score threshold, it indicates that the cleaning vehicle has engaged in dangerous behavior and can further prompt the user to apply the brakes or automatically control the cleaning vehicle to stop the current driving operation.

[0100] In an embodiment of the present application, by obtaining the driving state information and vehicle posture information of the cleaning vehicle, and then performing a waypoint search on the road segment information of the road where the cleaning vehicle is located according to the vehicle posture information, the waypoint direction vector corresponding to the cleaning vehicle in the current frame is determined. The angle difference between the cleaning vehicle in the current frame and the historical frame is further calculated based on the waypoint direction vectors of the two adjacent frames combined with the vehicle angle vector in the driving state information. The steering angular acceleration corresponding to the cleaning vehicle in the current frame is calculated based on the angle difference, and the steering wheel angular acceleration is calculated based on the driving state information. After obtaining the steering angular acceleration and the steering wheel angular acceleration, the behavior score corresponding to the cleaning vehicle in the current frame is determined in combination with a preset speed reference value. Finally, the behavior score is compared with a preset score threshold, and the comparison result is used to determine whether the cleaning vehicle has engaged in dangerous behavior in the current frame. Thus, by calculating the vehicle behavior score based on the driving state and vehicle posture of the cleaning vehicle, the cleaning vehicle's own dangerous behavior can be more accurately identified.

[0101] See also Figure 2 , Figure 2 This is a flowchart of the steps of a method for identifying dangerous behaviors of a cleaning vehicle provided in Example 2 of the present invention.

[0102] The present invention provides a method for identifying dangerous behaviors of a cleaning vehicle, comprising:

[0103] Step 201: Acquire driving state information and vehicle posture information of the cleaning vehicle; the driving state information includes the vehicle angle vector and steering wheel rotation speed;

[0104] In the embodiment of the present application, the specific implementation process of step 201 is similar to step 101 in embodiment 1 and will not be repeated here.

[0105] Step 202: searching for waypoints on the road where the cleaning vehicle is located according to the multi-frame vehicle posture information to determine the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame;

[0106] Optionally, the vehicle posture information includes the vehicle position and vehicle orientation; step 202 may include the following sub-steps S11-S13:

[0107] S11, performing a waypoint search on the road where the cleaning vehicle is located with the vehicle position in the current frame as the center, and determining multiple initial waypoints within a to-be-searched area with a preset radius;

[0108] Road waypoints refer to multiple road coordinates on a preset semantic map.

[0109] In an embodiment of the present application, a waypoint search may be performed on multiple road waypoints on the road with the cleaning vehicle as the center, so as to determine multiple initial waypoints within a to-be-searched area with a preset length radius.

[0110] It should be noted that the waypoint search can be implemented using a path planning algorithm, such as the A-Star algorithm or a dynamic programming algorithm.

[0111] S12, determining the target waypoint corresponding to the cleaning vehicle in the current frame based on the vehicle orientation, the initial waypoint, and a comparison result of multiple historical candidate waypoints obtained from a preset waypoint database;

[0112] Furthermore, step S12 may include the following sub-steps:

[0113] Calculate the difference between the vehicle's orientation and each initial waypoint's orientation;

[0114] Selecting a first preset number of initial waypoints as candidate waypoints according to the direction difference from smallest to largest;

[0115] Acquire a plurality of candidate frames before the current frame from a preset waypoint database according to a second preset number, and acquire a plurality of historical candidate waypoints corresponding to each candidate frame;

[0116] Compare the alternative waypoints with the historical alternative waypoints, and select the alternative waypoint with the largest number of overlaps and closest to the vehicle position as the target waypoint corresponding to the cleaning vehicle in the current frame.

[0117] The vehicle orientation refers to the orientation of the front of the cleaning vehicle relative to the world coordinate system at the current moment, which can be expressed in the form of a vector.

[0118] In this embodiment of the present application, the vehicle orientation of the cleaning vehicle corresponding to the current frame is obtained, and the orientation difference between the vehicle orientation and each initial waypoint is calculated, that is, the angular deviation between the vehicle orientation and the initial waypoint. A first preset number of initial waypoints are selected as candidate waypoints corresponding to the current frame in ascending order of orientation difference.

[0119] At the same time, multiple candidate frames before the current frame are obtained from a preset waypoint database according to a second preset number, and multiple historical candidate waypoints corresponding to each candidate frame are obtained as a data basis for subsequent comparison.

[0120] After obtaining the historical alternative waypoints, the alternative waypoints can be compared with the historical alternative waypoints to select the alternative waypoint with the largest number of overlaps from the alternative frame to the current frame and the closest distance to the vehicle position as the target waypoint corresponding to the cleaning vehicle in the current frame.

[0121] It should be noted that multiple candidate waypoints are selected for each frame and stored in the waypoint database, and are associated with the corresponding candidate frames. When selecting the target waypoint, the candidate with the most overlaps is selected first. If there is only one, it is used as the target waypoint. If there are multiple candidate waypoints with the same number of overlaps, the candidate closest to the vehicle's position is selected as the target waypoint.

[0122] The first preset number may be 2 or 3, and the second preset number may be 5 frames. This embodiment does not limit the specific values.

[0123] S13. Determine the vector of the target waypoint in the world coordinate system as the waypoint direction vector corresponding to the cleaning vehicle in the current frame.

[0124] After obtaining the target waypoint, it can be mapped to the world coordinate system to obtain the vector between the target waypoint and the origin in the world coordinate system, which is used as the waypoint direction vector corresponding to the cleaning vehicle in the current frame.

[0125] In another example of the present application, the initial waypoint may be updated further in combination with lane relationship information of the road where the cleaning vehicle is located, where the lane relationship information includes multiple traffic markings and marking types.

[0126] The search area can be divided into multiple intermediate areas using multiple traffic markings. These intermediate areas include not only the vehicle's location but also other areas. The intermediate areas are then filtered based on the marking type of each traffic marking and the corresponding driving rules to determine the lane-changing area for the cleaning vehicle. Finally, initial waypoints within the vehicle's location and the lane-changing area are selected and used as the new initial waypoints for the cleaning vehicle in the current frame.

[0127] Step 203, calculating the time difference between the current frame and the historical frame adjacent to the current frame;

[0128] Since the frame rate is different, the number of frames obtained will also be different. Therefore, in order to improve the accuracy of subsequent calculations, the time difference between the historical frame and the current frame can be calculated.

[0129] Step 204: determining the vehicle head steering angular acceleration corresponding to the current frame of the cleaning vehicle based on the vehicle angle vector and the waypoint direction vector of the historical frame and the current frame;

[0130] Optionally, step 204 may include the following sub-steps:

[0131] Calculate the vector dot product between the vehicle angle vector and the waypoint direction vector;

[0132] Calculate the vector length product between the vector length of the vehicle angle vector and the vector length of the waypoint direction vector;

[0133] Calculating a first ratio between the vector dot product and the vector length product, and using the arc cosine value of the first ratio as a first angle value between the cleaning vehicle and the road on which the cleaning vehicle is located;

[0134] Obtaining a first angle value corresponding to the historical frame as a second angle value;

[0135] Calculating an angle difference between the second angle value and the first angle value;

[0136] The ratio between the angle difference and the time difference is determined as the steering angular acceleration of the cleaning vehicle corresponding to the current frame.

[0137] In one example of the present application, by calculating the vector dot product between the vehicle angle vector and the waypoint direction vector, and calculating the vector length product between the vector lengths of the two, a first ratio between the vector dot product and the vector length product is further calculated:

[0138]

[0139] Among them, θ2 is the first angle value, p1 is the vehicle angle vector, and p2 is the waypoint direction vector.

[0140] The arc cosine value of the first ratio is used as a first angle value θ2 between the cleaning vehicle and the road.

[0141] At the same time, since the first angle value is calculated for each current frame, the first angle value corresponding to the historical frame adjacent to the current frame can be obtained from the cleaning vehicle database as the second angle value θ1. The angle difference between the second angle value and the first angle value is further calculated, and the ratio between the angle difference and the time difference is calculated to obtain the cleaning vehicle's head turning speed a1 corresponding to the current frame:

[0142]

[0143] Wherein, θ1 is the second angle value, t1 is the time corresponding to the historical frame, and t2 is the time corresponding to the current frame.

[0144] Step 205 , calculating the steering wheel rotation speed difference between the historical frame and the steering wheel rotation speed of the current frame;

[0145] Step 206 : Determine the ratio between the rotation speed difference and the time difference as the steering wheel rotation angular acceleration corresponding to the cleaning vehicle in the current frame.

[0146] At the same time, the steering wheel speed difference between the historical frame and the current frame is calculated, and the ratio between the speed difference and the time difference is further calculated to determine it as the steering wheel angular acceleration a2 corresponding to the cleaning vehicle in the current frame:

[0147]

[0148] Among them, v1 is the steering wheel rotation speed of the historical frame, and v2 is the steering wheel rotation speed of the current frame.

[0149] Step 207 , determining a behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration, combined with a preset speed reference value;

[0150] Optionally, the speed reference value includes a steering angular acceleration reference value and a rotation angular acceleration reference value; step 207 may include the following sub-steps S21-S23:

[0151] S21. Determine a vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value;

[0152] Furthermore, the steering angular acceleration reference value includes a steering angular acceleration full score value and a steering angular acceleration zero score value; S21 may include the following sub-steps:

[0153] calculating a first difference between the steering angular acceleration and a full-score value of the steering angular acceleration;

[0154] calculating a second difference between the zero-score value of the steering angular acceleration and the full-score value of the steering angular acceleration;

[0155] calculating a second ratio between the first difference and the second difference;

[0156] A difference between the preset constant and the second ratio is determined as a vehicle steering score.

[0157] The full score of steering angular acceleration refers to the upper limit of the vehicle's steering angular acceleration during driving operation. In this case, the cleaning vehicle has the highest behavioral safety in terms of steering angular acceleration and can be set to 0.05rad / s 2 .

[0158] The zero-point value for steering angular acceleration refers to the lower limit of the vehicle's steering angular acceleration during driving. In this case, the vehicle's behavior safety in terms of steering angular acceleration is the lowest, and can be set to 0.15 rad / s 2 .

[0159] In the embodiment of the present application, the steering angular acceleration reference value includes a full-score steering angular acceleration value and a zero-score steering angular acceleration value. The vehicle steering score is obtained by calculating a first difference between the steering angular acceleration and the full-score steering angular acceleration value, and simultaneously calculating a second difference between the zero-score steering angular acceleration value and the full-score steering angular acceleration value. After obtaining a second ratio between the first difference and the second difference, the difference between a preset constant and the second ratio is calculated. This can be specifically implemented using the following program formula:

[0160]

[0161] Among them, angle_score is the vehicle steering score, a 1max is the full score of steering angular acceleration, a 1min is the zero score of the steering angular acceleration, and 1 is a preset constant.

[0162] S22. Determine a steering wheel rotation score based on the steering wheel rotation angular acceleration and the rotation angular acceleration reference value;

[0163] Furthermore, the angular acceleration reference value includes a full score value of the steering wheel angular acceleration and a zero score value of the steering wheel angular acceleration; S22 may include the following sub-steps:

[0164] calculating a third difference between the steering wheel angular acceleration and a full score value of the steering wheel angular acceleration;

[0165] calculating a fourth difference between a zero-score value of the steering wheel angular acceleration and a full-score value of the steering wheel angular acceleration;

[0166] calculating a third ratio between the third difference and the fourth difference;

[0167] A difference between the preset constant and the third ratio is determined as a steering wheel rotation score.

[0168] The maximum value of the steering wheel angular acceleration refers to the upper limit of the vehicle's steering wheel angular acceleration during driving. In this case, the cleaning vehicle has the highest behavior safety under the steering wheel angular acceleration, which can be set to 1.0rad / s 2 .

[0169] The zero score for steering wheel angular acceleration refers to the lower limit of the vehicle's steering wheel angular acceleration during driving. In this case, the vehicle's behavior safety under the steering wheel angular acceleration is the lowest, and it can be set to 2.5 rad / s 2 .

[0170] In the embodiment of the present application, the steering wheel angular acceleration reference value includes a full-score steering wheel angular acceleration value and a zero-score steering wheel angular acceleration value. The steering wheel rotation score is obtained by calculating a third difference between the steering wheel angular acceleration and the full-score steering wheel angular acceleration value, and simultaneously calculating a fourth difference between the zero-score steering wheel angular acceleration value and the full-score steering wheel angular acceleration value. After obtaining a third ratio between the third difference and the fourth difference, the difference between a preset constant and the third ratio is calculated. This can be specifically implemented using the following program formula:

[0171]

[0172] Among them, wheel_score is the steering wheel rotation score, a 2max is the full score of steering angular acceleration, a 2min is the zero score of the steering angular acceleration, and 1 is a preset constant.

[0173] S23. Calculate the product of the vehicle steering score and the steering wheel rotation score to obtain the behavior score corresponding to the cleaning vehicle in the current frame.

[0174] After calculating the vehicle steering score and the steering wheel rotation score, the multiplication value between the two is further calculated to obtain the behavior score corresponding to the cleaning vehicle in the current frame.

[0175] Step 208 : Compare the behavior score with a preset score threshold to determine whether the cleaning vehicle has engaged in a dangerous behavior in the current frame.

[0176] In one example of the present application, step 208 may include the following sub-steps:

[0177] Comparing the behavior score with a preset score threshold;

[0178] If the behavior score is greater than or equal to the score threshold, it is determined that the cleaning vehicle has not performed any dangerous behavior in the current frame, and the process jumps to the step of obtaining the driving state information and vehicle posture information of the cleaning vehicle;

[0179] If the behavior score is less than the score threshold, it is determined that the cleaning vehicle has engaged in dangerous behavior in the current frame and a danger warning message is output.

[0180] In this embodiment, after calculating the behavior score, the behavior score can be further compared with a preset score threshold to determine whether the cleaning vehicle has engaged in dangerous behavior in the current frame. If the behavior score is greater than or equal to the score threshold, it indicates that the cleaning vehicle is in a relatively safe state in both vehicle angular acceleration and steering wheel angular acceleration in the current frame. Therefore, it can be determined that the cleaning vehicle has not engaged in dangerous behavior in the current frame, and the process can directly skip to step 201 to obtain the driving state information and vehicle posture information for the next frame.

[0181] If the behavior score is less than the score threshold, it indicates that the cleaning vehicle may have performed dangerous operations in the vehicle angular acceleration or the steering wheel angular acceleration. It can be determined that the cleaning vehicle has performed dangerous behavior in the current frame, and a danger warning message is output to stop the vehicle or require the user to perform subsequent operations.

[0182] In a specific implementation, the dangerous behavior identification method used by the cleaning vehicle, that is, the technical solution composed of the various technical means of steps 201 to 208, can usually be integrated into a program or storage medium and loaded into each cleaning vehicle for use. In order to ensure the safety of the method used by the cleaning vehicle as much as possible and to be able to repair program bugs as much as possible, it is necessary to continuously update the underlying program or code corresponding to the above solution. Therefore, before the program goes online, it is necessary to conduct a road test on the above solution to obtain road test data. When a simulation test request is received, the road test data of the cleaning vehicle within a preset time period is obtained, such as simulated alternative waypoints and simulated vehicle positions to execute the following steps S31-S37:

[0183] S31. When a simulation test request is received, obtaining the simulation candidate waypoints and the simulation vehicle position corresponding to each frame of the cleaning vehicle within a preset time period;

[0184] S32, selecting a simulation current frame frame by frame, and selecting multiple frames adjacent to the simulation current frame as simulation history frames;

[0185] S33, selecting the simulation candidate waypoint with the largest number of overlaps between all simulation history frames and the simulation current frame as the simulation target waypoint corresponding to the simulation current frame;

[0186] In an embodiment of the present application, when a simulation test request is received, the road test data of the cleaning vehicle can be obtained from a preset database. Specifically, the simulation alternative waypoints and simulation vehicle positions corresponding to each frame of the cleaning vehicle within a preset time period can be obtained; then the obtained simulation alternative waypoints and simulation vehicle positions are processed frame by frame, and the first frame is first selected as the simulation current frame, and multiple frames adjacent to the simulation current frame are selected as simulation history frames.

[0187] Since each frame has a corresponding simulation candidate waypoint and simulation vehicle position, the simulation candidate waypoint with the most overlaps in all simulation history frames and the simulation current frame can be further selected as the simulation target waypoint corresponding to the simulation current frame at the current moment.

[0188] The simulation history frames include not only the multiple frames after the current simulation frame, but also the multiple frames before the current simulation frame. For example, when the fifth frame is selected as the current simulation frame, the simulation history frames include the sixth to tenth frames adjacent to the fifth frame, and the fourth to first frames. The number of selected simulation history frames can be equal or unequal.

[0189] S34, selecting multiple frames after the current simulation frame as future simulation frames;

[0190] S35, respectively calculating the Euclidean distance between the simulated vehicle position corresponding to the simulated future frame and the simulated vehicle position corresponding to the simulated current frame for each frame;

[0191] S36, selecting the first simulated vehicle position whose Euclidean distance is greater than a preset distance threshold as the target vehicle position;

[0192] S37. Connect the simulated vehicle position and the target vehicle position to construct a simulation angle vector corresponding to the cleaning vehicle in the current simulation frame.

[0193] At the same time, multiple frames after the current simulation frame are selected as simulated future frames, and the Euclidean distances between the simulated vehicle position corresponding to each simulated future frame and the simulated vehicle position corresponding to the current simulation frame are calculated respectively. The first simulated vehicle position whose Euclidean distance is greater than the preset distance threshold is selected as the target vehicle position, and the simulated vehicle position and the target vehicle position are connected to construct the simulation angle vector corresponding to the current simulation frame of the cleaning vehicle.

[0194] It should be noted that after the simulation target waypoint and simulation angle vector corresponding to the first frame are determined, the subsequent second frames are selected in sequence as the simulation current frames until all frames are selected.

[0195] Furthermore, if a program version comparison request is received, the method further includes the following steps S41-S46:

[0196] S41. When a program version comparison request is received, the currently running program and the program to be put online corresponding to the cleaning vehicle are obtained;

[0197] S42, executing the current running program based on the simulation target waypoint and the simulation angle vector, and obtaining a corresponding first behavior score;

[0198] S43, executing the program to be put online based on the simulation target waypoint and the simulation angle vector, and obtaining a corresponding second behavior score;

[0199] S44. Calculate the absolute value of the version score difference between the second behavior score and the first behavior score;

[0200] S45. If the absolute value is greater than a preset version score threshold, it is determined that the program to be put online has been improved;

[0201] S46. If the absolute value is less than or equal to the preset version score threshold, it is determined that there is no improvement in the program to be put online.

[0202] The currently running program refers to various technical means for executing steps 201 to 208 , and is a set of instructions composed of and compiled by program codes that can be recognized and executed by a computer.

[0203] The program to be put online refers to the various technical means for executing steps 201 to 208, which is composed of and compiled from program codes and is a set of instructions that can be recognized and executed by a computer. The program code of the instructions is different from the currently running program.

[0204] In this embodiment of the present application, after determining the simulation target waypoint and simulation angle vector, if a program version comparison request is received, the current running program and the program to be launched corresponding to the cleaning vehicle are obtained as comparison objects. The simulation target waypoint is used as the new target waypoint, and the simulation angle vector is used as the new vehicle angle vector. Steps 201-208 of the current running program are executed to obtain the corresponding first behavior score. At the same time, steps 201-208 of the program to be launched are executed to obtain the corresponding second behavior score. By calculating the absolute value of the version score difference between the second behavior score and the first behavior score, it is determined whether the program to be launched has improved compared to the current running program.

[0205] If the absolute value is greater than the preset version score threshold, it is determined that the program to be launched has been improved. If the absolute value is less than or equal to the preset version score threshold, it is determined that the program to be launched has not been improved.

[0206] It should be noted that when it is determined that there are improvements in the program to be put online, since the absolute value of the above-mentioned version score difference cannot indicate the quality of the improvement, the behavioral score of the program to be put online can be calculated in response to the user's verification request to further determine the specific improvement of the program to be put online.

[0207] In an embodiment of the present application, by obtaining the driving state information and vehicle posture information of the cleaning vehicle, and then performing a waypoint search on the road segment information of the road where the cleaning vehicle is located according to the vehicle posture information, the waypoint direction vector corresponding to the cleaning vehicle in the current frame is determined. The angle difference between the cleaning vehicle in the current frame and the historical frame is further calculated based on the waypoint direction vectors of the two adjacent frames combined with the vehicle angle vector in the driving state information. The steering angular acceleration corresponding to the cleaning vehicle in the current frame is calculated based on the angle difference, and the steering wheel angular acceleration is calculated based on the driving state information. After obtaining the steering angular acceleration and the steering wheel angular acceleration, the behavior score corresponding to the cleaning vehicle in the current frame is determined in combination with a preset speed reference value. Finally, the behavior score is compared with a preset score threshold, and the comparison result is used to determine whether the cleaning vehicle has engaged in dangerous behavior in the current frame. Thus, by calculating the vehicle behavior score based on the driving state and vehicle posture of the cleaning vehicle, the cleaning vehicle's own dangerous behavior can be more accurately identified.

[0208] See also Figure 3 , Figure 3 This is a structural block diagram of a dangerous behavior identification device for a cleaning vehicle provided in Example 3 of the present invention.

[0209] An embodiment of the present invention provides a dangerous behavior identification device for a cleaning vehicle, comprising:

[0210] The vehicle information acquisition module 301 is used to obtain the driving status information and vehicle posture information of the cleaning vehicle in real time;

[0211] A waypoint direction search module 302 is used to search for waypoints on the road where the cleaning vehicle is located according to the multi-frame vehicle posture information, and determine the waypoint direction vector corresponding to the cleaning vehicle in the current frame;

[0212] The angular acceleration determination module 303 is used to determine the vehicle head turning angular acceleration and steering wheel rotation angular acceleration corresponding to the cleaning vehicle in the current frame based on the driving state information and the waypoint direction vector of two adjacent frames;

[0213] The behavior score calculation module 304 is used to determine the behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with a preset speed reference value;

[0214] The dangerous behavior judgment module 305 is used to compare the behavior score with a preset score threshold to determine whether the cleaning vehicle has performed a dangerous behavior in the current frame.

[0215] Optionally, the vehicle posture information includes the vehicle position and vehicle orientation; the waypoint direction search module 302 includes:

[0216] The initial waypoint search submodule is used to search for waypoints on the road where the cleaning vehicle is located with the vehicle position in the current frame as the center, and determine multiple initial waypoints within the search area with a preset radius;

[0217] The target waypoint determination submodule is used to determine the target waypoint corresponding to the cleaning vehicle in the current frame based on the vehicle orientation, the initial waypoint, and the comparison results of multiple historical candidate waypoints obtained from a preset waypoint database;

[0218] The direction vector conversion submodule is used to determine the vector of the target waypoint in the world coordinate system as the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame.

[0219] Optionally, the target waypoint determination submodule is specifically configured to:

[0220] Calculate the difference between the vehicle's orientation and each initial waypoint's orientation;

[0221] Selecting a first preset number of initial waypoints as candidate waypoints according to the direction difference from smallest to largest;

[0222] Acquire a plurality of candidate frames before the current frame from a preset waypoint database according to a second preset number, and acquire a plurality of historical candidate waypoints corresponding to each candidate frame;

[0223] Compare the alternative waypoints with the historical alternative waypoints, and select the alternative waypoint with the largest number of overlaps and closest to the vehicle position as the target waypoint corresponding to the cleaning vehicle in the current frame.

[0224] Optionally, the driving state information includes a vehicle angle vector and a steering wheel rotation speed; the angular acceleration determination module 303 includes:

[0225] The time difference calculation submodule is used to calculate the time difference between the historical frames adjacent to the current frame and the current frame;

[0226] The vehicle head steering angular acceleration determination submodule is used to determine the vehicle head steering angular acceleration corresponding to the current frame based on the vehicle angle vector and the waypoint direction vector of the historical frame and the current frame;

[0227] A rotation speed difference calculation submodule, configured to calculate a rotation speed difference between a steering wheel rotation speed in a historical frame and a steering wheel rotation speed in a current frame;

[0228] The steering wheel rotation angular acceleration determination submodule is used to determine the ratio between the rotation speed difference and the time difference as the steering wheel rotation angular acceleration corresponding to the cleaning vehicle in the current frame.

[0229] Optionally, the vehicle head steering angular acceleration determination submodule is specifically used to:

[0230] Calculate the vector dot product between the vehicle angle vector and the waypoint direction vector;

[0231] Calculate the vector length product between the vector length of the vehicle angle vector and the vector length of the waypoint direction vector;

[0232] Calculating a first ratio between the vector dot product and the vector length product, and using the arc cosine value of the first ratio as a first angle value between the cleaning vehicle and the road on which the cleaning vehicle is located;

[0233] Obtaining a first angle value corresponding to the historical frame as a second angle value;

[0234] Calculating an angle difference between the second angle value and the first angle value;

[0235] The ratio between the angle difference and the time difference is determined as the steering angular acceleration of the cleaning vehicle corresponding to the current frame.

[0236] Optionally, the speed reference value includes a steering angular acceleration reference value and a rotation angular acceleration reference value; the behavior score calculation module 304 includes:

[0237] A vehicle steering score calculation submodule, configured to determine a vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value;

[0238] a steering wheel rotation score calculation submodule, configured to determine a steering wheel rotation score based on the steering wheel rotation angular acceleration and the rotation angular acceleration reference value;

[0239] The behavior score determination submodule is used to calculate the product between the vehicle steering score and the steering wheel rotation score to obtain the behavior score corresponding to the cleaning vehicle in the current frame.

[0240] Optionally, the steering angular acceleration reference value includes a steering angular acceleration full score value and a steering angular acceleration zero score value; the vehicle steering score calculation submodule is specifically used to:

[0241] calculating a first difference between the steering angular acceleration and a full-score value of the steering angular acceleration;

[0242] calculating a second difference between the zero-score value of the steering angular acceleration and the full-score value of the steering angular acceleration;

[0243] calculating a second ratio between the first difference and the second difference;

[0244] A difference between the preset constant and the second ratio is determined as a vehicle steering score.

[0245] Optionally, the angular acceleration reference value includes a full score value of the steering wheel angular acceleration and a zero score value of the steering wheel angular acceleration; the steering wheel rotation score calculation submodule is specifically used to:

[0246] calculating a third difference between the steering wheel angular acceleration and a full score value of the steering wheel angular acceleration;

[0247] calculating a fourth difference between a zero-score value of the steering wheel angular acceleration and a full-score value of the steering wheel angular acceleration;

[0248] calculating a third ratio between the third difference and the fourth difference;

[0249] A difference between the preset constant and the third ratio is determined as a steering wheel rotation score.

[0250] Optionally, the dangerous behavior judgment module 305 includes:

[0251] A behavior score comparison submodule, used to compare the behavior score with a preset score threshold;

[0252] A first determination submodule is configured to determine that the cleaning vehicle has not exhibited any dangerous behavior in the current frame if the behavior score is greater than or equal to a score threshold, and to jump to the step of acquiring driving state information and vehicle posture information of the cleaning vehicle in real time;

[0253] The second determination submodule is configured to determine that the cleaning vehicle has engaged in a dangerous behavior in the current frame and output a danger warning message if the behavior score is less than a score threshold.

[0254] Optionally, the device further comprises:

[0255] A simulation data acquisition module is used to obtain the simulation alternative waypoints and simulation vehicle positions corresponding to each frame of the cleaning vehicle within a preset time period when a simulation test request is received;

[0256] The frame selection module is used to select the current simulation frame frame by frame and select multiple frames adjacent to the current simulation frame as simulation history frames;

[0257] A simulation target waypoint selection module is used to select the simulation candidate waypoint with the largest number of overlaps between all simulation history frames and the simulation current frame as the simulation target waypoint corresponding to the simulation current frame;

[0258] A simulation future frame selection module is used to select multiple frames after the simulation current frame as simulation future frames;

[0259] A Euclidean distance calculation module is used to calculate the Euclidean distance between the simulated vehicle position corresponding to the simulated future frame and the simulated vehicle position corresponding to the simulated current frame for each frame;

[0260] A target vehicle position selection module is used to select the first simulated vehicle position whose Euclidean distance is greater than a preset distance threshold as the target vehicle position;

[0261] The simulation angle vector construction module is used to connect the simulated vehicle position and the target vehicle position to construct the simulation angle vector corresponding to the cleaning vehicle in the current simulation frame.

[0262] Optionally, the device further comprises:

[0263] A program acquisition module is used to obtain the current running program and the program to be put online corresponding to the cleaning vehicle when receiving a program version comparison request;

[0264] The first behavior score calculation module is used to execute the current running program based on the simulation target waypoint and the simulation angle vector to obtain the corresponding first behavior score;

[0265] The second behavior score calculation module is used to execute the program to be run online based on the simulation target waypoint and the simulation angle vector to obtain the corresponding second behavior score;

[0266] A version score difference calculation module, configured to calculate an absolute value of a version score difference between a second behavior score and a first behavior score;

[0267] A first improvement determination module is configured to determine that there is improvement in the program to be put online if the absolute value is greater than a preset version score threshold;

[0268] The second improvement determination module is configured to determine that there is no improvement in the program to be put online if the absolute value is less than or equal to a preset version score threshold.

[0269] An embodiment of the present invention also provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory. When the computer program is executed by the processor, the processor executes the steps of the dangerous behavior identification method for a cleaning vehicle as described in any embodiment of the present invention.

[0270] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed, the method for identifying dangerous behaviors of a cleaning vehicle as described in any embodiment of the present invention is implemented.

[0271] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, modules and sub-modules can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0272] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0273] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0274] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0275] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0276] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying dangerous behaviors of a cleaning vehicle, characterized in that: include: Obtain the driving status information and vehicle posture information of the cleaning vehicle in real time; Performing a waypoint search on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames, and determining the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame; Determine the head steering angular acceleration and steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame according to the driving state information and the waypoint direction vector of two adjacent frames; Determining a behavior score of the cleaning vehicle corresponding to the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration and a preset speed reference value; Comparing the behavior score with a preset score threshold to determine whether the cleaning vehicle exhibits a dangerous behavior in the current frame; The vehicle posture information includes the vehicle position and vehicle orientation; the step of searching for a waypoint on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames and determining the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame includes: Performing a waypoint search on the road where the cleaning vehicle is located with the vehicle position in the current frame as the center, and determining a plurality of initial waypoints within a to-be-searched area with a preset radius; Determining a target waypoint corresponding to the cleaning vehicle in the current frame based on a comparison result of the vehicle orientation, the initial waypoint, and a plurality of historical candidate waypoints obtained from a preset waypoint database; Determine the vector of the target waypoint in the world coordinate system as the waypoint direction vector corresponding to the cleaning vehicle in the current frame; The step of determining the target waypoint corresponding to the cleaning vehicle in the current frame based on the comparison results of the vehicle orientation, the initial waypoint, and multiple historical candidate waypoints obtained from a preset waypoint database includes: Calculating the difference between the vehicle orientation and each of the initial waypoints; Selecting a first preset number of the initial waypoints as candidate waypoints according to the direction differences from smallest to largest; Acquire a plurality of candidate frames before the current frame from a preset waypoint database according to a second preset number, and acquire a plurality of historical candidate waypoints corresponding to each of the candidate frames; The candidate waypoints are compared with the historical candidate waypoints, and the candidate waypoint with the largest number of overlaps and closest to the vehicle position is selected as the target waypoint corresponding to the cleaning vehicle in the current frame.

2. The method according to claim 1, characterized in that The driving state information includes a vehicle angle vector and a steering wheel rotation speed; the step of determining the vehicle head steering angular acceleration and the steering wheel rotation angular acceleration corresponding to the current frame of the cleaning vehicle based on the driving state information and the waypoint direction vector of two adjacent frames includes: Calculating a time difference between a historical frame adjacent to the current frame and the current frame; Determine the vehicle head steering angular acceleration corresponding to the current frame of the cleaning vehicle according to the vehicle angle vector and the waypoint direction vector of the historical frame and the current frame; Calculating a rotation speed difference between the steering wheel rotation speed of the historical frame and the steering wheel rotation speed of the current frame; The ratio of the rotation speed difference to the time difference is determined as the steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame.

3. The method according to claim 1, characterized in that The speed reference value includes a steering angular acceleration reference value and a rotation angular acceleration reference value; the step of determining the behavior score of the cleaning vehicle corresponding to the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with the preset speed reference value includes: determining a vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value; determining a steering wheel rotation score based on the steering wheel rotation angular acceleration and the rotation angular acceleration reference value; The multiplication value between the vehicle steering score and the steering wheel rotation score is calculated to obtain the behavior score of the cleaning vehicle corresponding to the current frame.

4. The method according to claim 3, characterized in that The steering angular acceleration reference value includes a steering angular acceleration full score value and a steering angular acceleration zero score value; The step of determining the vehicle steering score based on the steering angular acceleration and the steering angular acceleration reference value comprises: calculating a first difference between the steering angular acceleration and a full score value of the steering angular acceleration; calculating a second difference between the zero-score value of the steering angular acceleration and the full-score value of the steering angular acceleration; calculating a second ratio between the first difference and the second difference; A difference between a preset constant and the second ratio is determined as a vehicle steering score.

5. The method according to claim 3, characterized in that The rotation angular acceleration reference value includes a full score value of the steering wheel rotation angular acceleration and a zero score value of the steering wheel rotation angular acceleration; and the step of determining the steering wheel rotation score based on the steering wheel rotation angular acceleration and the rotation angular acceleration reference value includes: calculating a third difference between the steering wheel angular acceleration and a full score value of the steering wheel angular acceleration; calculating a fourth difference between the zero-score value of the steering wheel rotation angular acceleration and the full-score value of the steering wheel rotation angular acceleration; calculating a third ratio between the third difference and the fourth difference; A difference between the preset constant and the third ratio is determined as a steering wheel rotation score.

6. The method according to claim 1, characterized in that The step of comparing the behavior score with a preset score threshold to determine whether the cleaning vehicle has engaged in a dangerous behavior in the current frame includes: comparing the behavior score with a preset score threshold; If the behavior score is greater than or equal to the score threshold, it is determined that the cleaning vehicle has not exhibited any dangerous behavior in the current frame, and the process jumps to executing the step of obtaining the driving state information and vehicle posture information of the cleaning vehicle in real time; If the behavior score is less than the score threshold, it is determined that the cleaning vehicle has engaged in a dangerous behavior in the current frame, and a danger warning message is output.

7. The method according to any one of claims 1 to 6, characterized in that The method further comprises: When a simulation test request is received, the simulation candidate waypoints and the simulation vehicle position corresponding to each frame of the cleaning vehicle within a preset time period are obtained; Selecting a simulation current frame frame by frame, and selecting multiple frames adjacent to the simulation current frame as simulation history frames; Selecting the simulation candidate waypoint with the largest number of overlaps between all the simulation history frames and the simulation current frame as the simulation target waypoint corresponding to the simulation current frame; Selecting multiple frames after the current simulation frame as future simulation frames; Calculating the Euclidean distance between the simulated vehicle position corresponding to the simulated future frame and the simulated vehicle position corresponding to the simulated current frame for each frame; Selecting the first simulated vehicle position whose Euclidean distance is greater than a preset distance threshold as the target vehicle position; The simulated vehicle position and the target vehicle position are connected to construct a simulation angle vector corresponding to the cleaning vehicle in the current simulation frame.

8. The method according to claim 7, characterized in that The method further comprises: When a program version comparison request is received, the current running program and the program to be put online corresponding to the cleaning vehicle are obtained; Executing the currently running program based on the simulation target waypoint and the simulation angle vector to obtain a corresponding first behavior score; Executing the program to be put online based on the simulation target waypoint and the simulation angle vector to obtain a corresponding second behavior score; Calculating an absolute value of a version score difference between the second behavior score and the first behavior score; If the absolute value is greater than a preset version score threshold, it is determined that there is improvement in the program to be put online; If the absolute value is less than or equal to a preset version score threshold, it is determined that there is no improvement in the program to be put online.

9. A dangerous behavior identification device for a cleaning vehicle, characterized in that: include: Vehicle information acquisition module, used to obtain the driving status information and vehicle posture information of the cleaning vehicle in real time; A waypoint direction search module is used to search for waypoints on the road where the cleaning vehicle is located according to the vehicle posture information of multiple frames, and determine the waypoint direction vector corresponding to the cleaning vehicle in the current frame; an angular acceleration determination module, configured to determine the head steering angular acceleration and the steering wheel rotation angular acceleration of the cleaning vehicle corresponding to the current frame based on the driving state information and the waypoint direction vector of two adjacent frames; a behavior score calculation module, configured to determine a behavior score corresponding to the cleaning vehicle in the current frame based on the steering angular acceleration and the steering wheel rotation angular acceleration in combination with a preset speed reference value; a dangerous behavior judgment module, configured to compare the behavior score with a preset score threshold to determine whether the cleaning vehicle has engaged in a dangerous behavior in the current frame; The vehicle posture information includes vehicle position and vehicle orientation; The waypoint direction search module includes: An initial waypoint search submodule is used to perform a waypoint search on the road where the cleaning vehicle is located with the vehicle position in the current frame as the center, and determine a plurality of initial waypoints within a to-be-searched area with a preset radius; A target waypoint determination submodule is configured to determine a target waypoint corresponding to the cleaning vehicle in the current frame based on a comparison result of the vehicle orientation, the initial waypoint, and a plurality of historical candidate waypoints obtained from a preset waypoint database; A direction vector conversion submodule, configured to determine the vector of the target waypoint in the world coordinate system as the direction vector of the waypoint corresponding to the cleaning vehicle in the current frame; The target waypoint determination submodule is specifically used to: Calculating the difference between the vehicle orientation and each of the initial waypoints; Selecting a first preset number of the initial waypoints as candidate waypoints according to the direction differences from smallest to largest; Acquire a plurality of candidate frames before the current frame from a preset waypoint database according to a second preset number, and acquire a plurality of historical candidate waypoints corresponding to each of the candidate frames; The candidate waypoints are compared with the historical candidate waypoints, and the candidate waypoint with the largest number of overlaps and closest to the vehicle position is selected as the target waypoint corresponding to the cleaning vehicle in the current frame.

10. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the dangerous behavior identification method of the cleaning vehicle according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the dangerous behavior identification method of the cleaning vehicle according to any one of claims 1 to 8 is implemented.

Images