Vehicle obstacle avoidance method, device, equipment and medium based on scene recognition

Abstract

The present application provides a vehicle obstacle avoidance method, apparatus, device, and medium based on scene recognition. The method comprises: determining, based on a preset recognition device, the lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and the longitudinal obstacle and its longitudinal distance from the vehicle, and determining the vehicle's obstacle avoidance mode based on the lateral distance, longitudinal distance, vehicle width, and vehicle length; when the vehicle's obstacle avoidance mode is an oblique mode, determining a target offset distance based on the obstacle avoidance boundary of the longitudinal obstacle, and determining a vector angle based on the target offset distance and the vehicle's corresponding path reference line; determining the vehicle's corresponding retreat distance based on the target offset distance, the vector angle, and the vehicle's wheel steering angle threshold; determining the oblique travel starting point based on the retreat distance, and determining the vehicle's oblique travel trajectory based on the oblique travel mode and a quadratic programming algorithm; and avoiding the longitudinal obstacle based on the oblique travel starting point and the oblique travel trajectory. The flexibility and accuracy of the vehicle's obstacle avoidance are improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle obstacle avoidance technology, and in particular to a vehicle obstacle avoidance method, device, equipment and medium based on scene recognition. Background Art

[0002] The port environment is a highly complex and dynamic scene, which contains various types of vehicles and other complex obstacles. Vehicles need to be able to accurately judge complex road conditions and avoid obstacles in a timely manner.

[0003] Existing vehicle obstacle avoidance methods generally use front wheel steering to change lanes and avoid obstacles. They are mainly used for steering on conventional roads and rely on the front wheel steering angle to control the trajectory. Obstacle avoidance can also be achieved through a single diagonal mode.

[0004] However, existing vehicle obstacle avoidance methods have the problems of poor flexibility and low accuracy. Summary of the Invention

[0005] The present application provides a vehicle obstacle avoidance method, apparatus, device and medium based on scene recognition, which are used to solve the problems of poor flexibility and low accuracy of existing vehicle obstacle avoidance methods.

[0006] In a first aspect, the present application provides a vehicle obstacle avoidance method based on scene recognition, the method comprising:

[0007] According to the preset recognition equipment, the lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, as well as the longitudinal obstacle and its longitudinal distance from the vehicle are determined, and the vehicle's obstacle avoidance mode is determined based on the lateral distance, longitudinal distance, vehicle width distance and vehicle length distance;

[0008] When the vehicle's obstacle avoidance mode is oblique, the target offset distance is determined based on the obstacle avoidance boundary of the longitudinal obstacle, and the vector angle is determined based on the target offset distance and the path reference line corresponding to the vehicle;

[0009] Determine the corresponding retreat distance of the vehicle based on the target offset distance, the vector angle, and the vehicle's wheel steering angle threshold;

[0010] Determine the starting point of the oblique movement based on the retreat distance, and determine the oblique movement trajectory of the vehicle based on the oblique movement pattern and the quadratic programming algorithm;

[0011] According to the oblique starting point and oblique trajectory, longitudinal obstacles are circumvented.

[0012] In some embodiments of the present application, a preset recognition device is used to determine a lateral obstacle corresponding to a vehicle and its lateral distance from the vehicle, as well as a longitudinal obstacle and its longitudinal distance from the vehicle. The vehicle's obstacle avoidance mode is determined based on the lateral distance, longitudinal distance, vehicle width, and vehicle length, including:

[0013] Determine a target distance threshold according to the sum of the vehicle length and the preset distance;

[0014] Compare the lateral distance and vehicle width distance, as well as the longitudinal distance and target distance threshold to obtain a comparison result;

[0015] If the comparison result shows that the lateral distance is less than the vehicle width distance, and the longitudinal distance is less than the target distance threshold, the obstacle avoidance mode is determined to be the oblique driving mode;

[0016] If the comparison result is that the lateral distance is not less than the vehicle width distance, or the longitudinal distance is not less than the target distance threshold, the obstacle avoidance mode is determined to be the figure-eight obstacle avoidance mode.

[0017] In some embodiments of the present application, a target offset distance is determined based on the obstacle avoidance boundary of the longitudinal obstacle, and a vector angle is determined based on the target offset distance and the path reference line corresponding to the vehicle, including:

[0018] According to the frenet coordinate system, determine the path reference line corresponding to the vehicle;

[0019] Determine the vehicle's obstacle avoidance direction based on the path reference line and the longitudinal obstacle, and determine the boundary of each boundary of the longitudinal obstacle that corresponds to the obstacle avoidance direction as the obstacle avoidance boundary;

[0020] Determine the distance between the obstacle avoidance boundary and the path reference line as the offset boundary distance, and determine the preset obstacle avoidance safety distance;

[0021] Determine the target offset distance based on the sum of the offset boundary distance, the preset obstacle avoidance safety distance, and the vehicle width distance;

[0022] Determine the vector angle based on the target offset distance and the path reference line.

[0023] In some embodiments of the present application, determining the vector angle based on the target offset distance and the path reference line includes:

[0024] According to the Frenet coordinate system, determine the initial point of the path reference line and the lower boundary of the longitudinal obstacle, and determine the distance between the initial point and the lower boundary as the obstacle boundary distance;

[0025] Determine the coordinates of the target offset point based on the target offset distance and the obstacle boundary distance, and determine the target vector based on the coordinates of the target offset point and the coordinates corresponding to the center point of the vehicle;

[0026] Determine the projection point of the vehicle on the path reference line and determine the projection heading angle corresponding to the projection point;

[0027] Determine the vector angle based on the target vector and the projection orientation angle.

[0028] In some embodiments of the present application, determining the corresponding backward distance of the vehicle based on the vector angle, the vehicle's wheel steering angle threshold, and the target offset distance includes:

[0029] Determining a wheel steering angle threshold according to a maximum wheel steering angle of the vehicle and a preset steering coefficient, and comparing the vector angle with the wheel steering angle threshold to obtain a comparison result;

[0030] If the comparison result shows that the vector angle is not greater than the wheel steering angle threshold, the retreat distance is determined to be zero;

[0031] If the comparison result shows that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined based on the target offset distance.

[0032] In some embodiments of the present application, if the comparison result shows that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined based on the target offset distance, including:

[0033] Determine the vehicle boundary distance between the upper boundary of the vehicle and the initial reference point of the path reference line according to the Frenet coordinate system, and determine the obstacle boundary distance corresponding to the longitudinal obstacle;

[0034] Determine the target driving distance based on the difference between the vehicle boundary distance, the obstacle boundary distance and the preset driving safety distance;

[0035] The retreat distance is determined based on the target driving distance, the target offset distance and the wheel steering angle threshold.

[0036] In some embodiments of the present application, the starting point of the oblique movement is determined based on the backward distance, and the oblique movement trajectory of the vehicle is determined based on the oblique movement pattern and the quadratic programming algorithm, including:

[0037] According to the oblique pattern, determine the penalty coefficient and penalty weight corresponding to the quadratic programming algorithm;

[0038] The oblique trajectory is determined according to the target offset distance, penalty coefficient, penalty weight and quadratic programming algorithm.

[0039] In a second aspect, the present application provides a vehicle obstacle avoidance device based on scene recognition, the device comprising:

[0040] a mode determination module, configured to determine, based on a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and a longitudinal obstacle and its longitudinal distance from the vehicle, and determine an obstacle avoidance mode for the vehicle based on the lateral distance, longitudinal distance, vehicle width, and vehicle length;

[0041] A vector angle determination module is used to determine a target offset distance based on the obstacle avoidance boundary of the longitudinal obstacle when the vehicle's obstacle avoidance mode is the oblique mode, and to determine a vector angle based on the target offset distance and the path reference line corresponding to the vehicle;

[0042] A retreat distance determination module is used to determine the corresponding retreat distance of the vehicle based on the target offset distance, the vector angle and the vehicle's wheel steering angle threshold;

[0043] The trajectory determination module is used to determine the starting point of the oblique movement according to the retreat distance, and to determine the oblique movement trajectory of the vehicle according to the oblique movement pattern and the quadratic programming algorithm;

[0044] The obstacle avoidance module is used to avoid longitudinal obstacles based on the oblique starting point and oblique trajectory.

[0045] In a third aspect, the present application provides a computer device, comprising: a processor, and a memory communicatively connected to the processor;

[0046] Memory stores computer-executable instructions;

[0047] The processor executes the computer-executable instructions stored in the memory to implement the method of the present application.

[0048] In a fourth aspect, the present application provides a computer-readable storage medium, in which program code is stored. When the program code is executed by a processor, it is used to implement the method of the present application.

[0049] The vehicle obstacle avoidance method, apparatus, device, and medium based on scene recognition provided in the present application determine, by using a preset recognition device, the lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and the longitudinal obstacle and its longitudinal distance from the vehicle, and determine the vehicle's obstacle avoidance mode based on the lateral distance, longitudinal distance, vehicle width, and vehicle length. When the vehicle's obstacle avoidance mode is an oblique mode, a target offset distance is determined based on the obstacle avoidance boundary of the longitudinal obstacle, and a vector angle is determined based on the target offset distance and the path reference line corresponding to the vehicle. The corresponding retreat distance of the vehicle is determined based on the target offset distance, the vector angle, and the vehicle's wheel steering angle threshold. The oblique starting point is determined based on the retreat distance, and the oblique trajectory of the vehicle is determined based on the oblique mode and a quadratic programming algorithm. The longitudinal obstacle is avoided based on the oblique starting point and the oblique trajectory.

[0050] In this way, it is possible to autonomously identify different scenarios and actively switch between diagonal and figure-eight motion modes according to the characteristics of the current scenario. In diagonal mode, the four wheels of the vehicle can use the same steering angle to keep the vehicle moving in a parallel and straight line, thereby effectively circumventing obstacles and getting out of trouble. This obstacle avoidance method not only avoids the risk of scratches caused by vehicle rotation, but also allows for more flexible adjustment of the driving trajectory in narrow spaces or complex road conditions, improving the efficiency and safety of obstacle avoidance. In curve scenarios, scenarios with large heading angle errors, and other scenarios requiring rapid steering, the system automatically switches to figure-eight steering to achieve rapidly changing IGV heading angles, effectively improving the vehicle's steering efficiency and flexibility in special scenarios, ensuring that it can respond smoothly, safely, and quickly to various steering requirements, maintain a stable driving trajectory, and provide strong guarantees for the vehicle's automated and efficient operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0052] Figure 1 A schematic diagram of a process flow of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application;

[0053] Figure 2 A schematic diagram of an oblique obstacle avoidance method for a vehicle based on scene recognition provided in an embodiment of the present application;

[0054] Figure 3 A schematic diagram of a figure-eight obstacle avoidance method for a vehicle based on scene recognition provided in an embodiment of the present application;

[0055] Figure 4 A schematic diagram of the architecture of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application;

[0056] Figure 5 A schematic structural diagram of a vehicle obstacle avoidance device based on scene recognition provided in an embodiment of the present application;

[0057] Figure 6 This is a structural block diagram of a device for executing a vehicle obstacle avoidance method based on scene recognition according to an embodiment of the present application. DETAILED DESCRIPTION

[0058] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0059] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0060] Figure 1 This is a flow chart of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application. Figure 1 As shown, the vehicle obstacle avoidance method based on scene recognition may include the following steps:

[0061] S110. Determine, based on a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and a longitudinal obstacle and its longitudinal distance from the vehicle, and determine an obstacle avoidance mode for the vehicle based on the lateral distance, longitudinal distance, vehicle width, and vehicle length.

[0062] Among them, the preset recognition device is a pre-set recognition device used to identify obstacles within a preset range of the vehicle, such as a sensor or radar device; the preset range can be a pre-set recognition range, such as a range of 50 meters of the vehicle body.

[0063] A lateral obstacle is a horizontal obstacle that exists within the lateral range corresponding to the vehicle body, and a longitudinal obstacle is a vertical obstacle that exists within the longitudinal range corresponding to the vehicle width. It can be understood that lateral obstacles and longitudinal obstacles are obstacles within the preset range of the vehicle that may collide with the vehicle and therefore need to be circumvented to ensure that the vehicle continues to drive normally.

[0064] The lateral distance is the shortest distance between the lateral obstacle and the vehicle, the longitudinal distance is the shortest distance between the longitudinal obstacle and the vehicle, the vehicle width distance is the length corresponding to the vehicle width, and the vehicle length distance is the length corresponding to the vehicle body.

[0065] Obstacle avoidance mode is a driving mode in which the vehicle avoids obstacles. For example, it can be a figure-eight avoidance or a diagonal avoidance. Figure-eight avoidance means that the front and rear wheels of the vehicle turn in opposite directions, forming a steering pattern similar to the shape of an "eight". This steering method can significantly reduce the vehicle's turning radius and improve steering flexibility. Diagonal mode refers to a movement mode in which all wheels of the vehicle turn at the same angle, allowing the vehicle to move diagonally. In this mode, the vehicle can move laterally while maintaining the direction of the body unchanged. Four-wheel steering vehicles achieve this by independently controlling the steering angle of each wheel.

[0066] Based on this, during the normal driving of the vehicle, the lateral obstacles and longitudinal obstacles within the preset range of the vehicle, as well as the distance between each obstacle and the vehicle, are determined through the identification device, so that the distance between the obstacle and the vehicle can be evaluated based on the lateral distance, longitudinal distance, vehicle width distance and vehicle length, so as to determine that the current vehicle needs to switch the driving mode to the oblique driving mode to complete the circumvention of the obstacle.

[0067] S120: When the obstacle avoidance mode of the vehicle is the oblique mode, determine a target offset distance according to the obstacle avoidance boundary of the longitudinal obstacle, and determine a vector angle according to the target offset distance and a path reference line corresponding to the vehicle.

[0068] The obstacle avoidance boundary can be understood as the boundary closest to the longitudinal obstacle and the vehicle's obstacle avoidance route. For example, the longitudinal obstacle can be a square box. If the vehicle avoids the obstacle from the left side of the box according to the diagonal mode, the left boundary of the box is the obstacle avoidance boundary; if the vehicle avoids the obstacle from the right side of the box according to the diagonal mode, the right boundary of the box is the obstacle avoidance boundary.

[0069] The target offset distance is the target offset, which refers to the final lateral position that the IGV needs to reach when bypassing an obstacle. It is specifically defined as the lateral distance between the end point of the obstacle bypass path and the lane reference line in the Frenet coordinate system, that is, the safe position that the vehicle should reach after bypassing the obstacle. IGV (Intelligent Guided Vehicle) refers to an intelligent guided transport vehicle, also known as an unmanned container truck (an autonomous driving truck without a cab). It can drive autonomously through environmental perception and path planning, with higher path flexibility and flexibility, and is suitable for dynamic scheduling scenarios such as ports, warehouses, and factories.

[0070] A path reference line is the ideal path for a vehicle or robot in trajectory planning (such as a lane centerline or a globally planned path). A path reference line is a smooth path consisting of a series of reference points. It contains information such as the orientation, coordinates, and curvature of the path points and can be simply understood as the lane centerline.

[0071] The vector angle can be understood as the angle between the target point or target path direction after the vehicle circumvents the obstacle and the current direction. It is used to determine whether the vehicle can successfully circumvent the obstacle by directly adopting the diagonal mode at the current distance. That is, a comprehensive check is carried out on the preliminary safety conditions for diagonal obstacle avoidance. If the angles with the obstacle meet the requirements for safe diagonal obstacle avoidance, the path planning is directly carried out according to the final state of the obstacle and the quadratic planning optimization method. If the current conditions cannot meet the requirements for safe diagonal obstacle avoidance, the reversing distance required to achieve safe diagonal obstacle avoidance is calculated, and a reversing trajectory is first planned. Then, the IGV is guided to reverse to the appropriate position for diagonal obstacle avoidance. After the vehicle is adjusted to meet the diagonal obstacle conditions, it is switched to the diagonal mode for obstacle avoidance.

[0072] Based on this, by determining the boundary of the longitudinal obstacle relative to the obstacle avoidance route, the target offset distance between the vehicle and the boundary after the obstacle is determined. Then, based on the target offset distance and the lane reference line in the vehicle trajectory planning, the angle between the target path direction and the current direction is further determined, so as to subsequently determine whether the current vehicle can directly avoid the obstacle diagonally.

[0073] S130 : Determine a corresponding backward distance of the vehicle according to the target offset distance, the vector angle, and a wheel steering angle threshold of the vehicle.

[0074] The wheel steering angle threshold is a threshold corresponding to the steering angle of the wheel, and is used to determine whether the vehicle can directly diagonally avoid obstacles at the current angle and distance. For example, the wheel steering angle threshold may be 0.9 times the maximum steering angle of the wheel.

[0075] The back-off distance is the distance the vehicle needs to back off to avoid obstacles. Figure 2 , Figure 2 A schematic diagram of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application; Figure 2 As shown in the figure, OBS represents an obstacle, and IGV can be understood as an unmanned container truck. Based on the vector angle and the wheel steering angle threshold, it is determined that a collision will occur when the current vehicle directly switches to the oblique driving mode to avoid the obstacle. That is, the vehicle cannot directly avoid the obstacle at the current distance. At this time, it is necessary to determine the reversing distance required to achieve safe oblique driving. First, a reversing trajectory is planned, and then the IGV is guided to reverse to the appropriate position. After the vehicle is adjusted to meet the oblique driving conditions, it is switched to the oblique driving mode to avoid the obstacle.

[0076] Based on this, by determining the vector angle and wheel steering angle threshold, a collision will occur when the current vehicle directly switches to the oblique mode to avoid obstacles, and thus determining the retreat distance according to the target offset distance, so that the vehicle first retreats according to the retreat distance, and then plans the oblique trajectory again to avoid obstacles.

[0077] S140 : Determine the starting point of the oblique movement according to the backward distance, and determine the oblique movement trajectory of the vehicle according to the oblique movement pattern and the quadratic programming algorithm.

[0078] The starting point of the oblique movement is the point where the vehicle moves backward according to the backward distance and then moves obliquely around obstacles again.

[0079] The quadratic programming algorithm, also known as the QP (Quadratic Programming) algorithm, aims to solve the minimum (or maximum) value of a quadratic objective function under linear constraints. It is a typical example of a convex optimization problem. It can be used to replan the vehicle's diagonal trajectory when it circumvents obstacles at the distance and angle corresponding to the starting point of the diagonal movement.

[0080] Based on this, by determining the corresponding backward distance of the vehicle, the starting point of the oblique movement after the backward movement is determined, so that under the driving conditions corresponding to the starting point of the oblique movement, the corresponding trajectory of the vehicle's oblique movement mode is planned again according to the oblique movement mode and the quadratic programming algorithm to obtain the oblique movement trajectory under the starting point of the oblique movement.

[0081] S150: circumventing longitudinal obstacles based on the oblique travel starting point and the oblique travel trajectory.

[0082] Based on this, when the vehicle is unable to avoid obstacles diagonally under the current driving conditions, the starting point of the diagonal movement can be determined by determining the backward distance, and the diagonal trajectory under the starting point of the diagonal movement can be determined, so that the vehicle can avoid obstacles diagonally again, thereby achieving precise obstacle avoidance in different scenarios and improving the flexibility and accuracy of obstacle avoidance.

[0083] Based on the feasible implementation of the above S110, the present application further provides a method for determining, according to a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and a longitudinal obstacle and its longitudinal distance from the vehicle, and determining an obstacle avoidance mode for the vehicle based on the lateral distance, longitudinal distance, vehicle width distance, and vehicle length distance, including the following steps:

[0084] Determine a target distance threshold according to the sum of the vehicle length and the preset distance;

[0085] Compare the lateral distance and vehicle width distance, as well as the longitudinal distance and target distance threshold to obtain a comparison result;

[0086] If the comparison result shows that the lateral distance is less than the vehicle width distance, and the longitudinal distance is less than the target distance threshold, the obstacle avoidance mode is determined to be the oblique driving mode;

[0087] If the comparison result is that the lateral distance is not less than the vehicle width distance, or the longitudinal distance is not less than the target distance threshold, the obstacle avoidance mode is determined to be the figure-eight obstacle avoidance mode.

[0088] The preset distance can be understood as a pre-set safe distance for obstacle avoidance, which is used to improve the safety of obstacle avoidance. The target distance threshold is the distance threshold used for obstacle avoidance judgment. Please refer to Figure 3 , Figure 3 A figure eight obstacle avoidance diagram of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application; Figure 3 As shown, there are obstacles in both the left and right lanes, and they do not overlap longitudinally. However, the lateral distance is less than the width of the IGV, and the longitudinal distance is less than the length of the IGV + 5m. That is, the longitudinal space in the obstacle avoidance scenario does not meet the requirements for the IGV to travel diagonally. In this case, a figure-eight pattern is required to avoid the obstacle. Furthermore, in the curved scenario, the figure-eight pattern can also be used to quickly change the IGV's heading angle to achieve rapid steering and cornering.

[0089] Based on this, when it is determined that there are corresponding obstacles within the lateral range and longitudinal range of the vehicle and they do not overlap, it is necessary to determine the obstacle avoidance mode required for the current scenario based on the distance between the vehicle and the obstacle. Therefore, by comparing the lateral distance and vehicle width distance, as well as the longitudinal distance and target distance threshold, it is possible to determine whether to avoid the obstacle in a figure eight or oblique manner in the current scenario based on the comparison results, so as to perform corresponding obstacle avoidance according to the real-time scenario, thereby improving the flexibility of obstacle avoidance.

[0090] Based on the feasible implementation of the above S120, the present application further provides a method for determining a target offset distance based on the obstacle avoidance boundary of the longitudinal obstacle, and determining a vector angle based on the target offset distance and the path reference line corresponding to the vehicle, including the following steps:

[0091] According to the frenet coordinate system, determine the path reference line corresponding to the vehicle;

[0092] Determine the vehicle's obstacle avoidance direction based on the path reference line and the longitudinal obstacle, and determine the boundary of each boundary of the longitudinal obstacle that corresponds to the obstacle avoidance direction as the obstacle avoidance boundary;

[0093] Determine the distance between the obstacle avoidance boundary and the path reference line as the offset boundary distance, and determine the preset obstacle avoidance safety distance;

[0094] Determine the target offset distance based on the sum of the offset boundary distance, the preset obstacle avoidance safety distance, and the vehicle width distance;

[0095] Determine the vector angle based on the target offset distance and the path reference line.

[0096] Among them, the Frenet coordinate system is a local coordinate system related to the path. Its coordinate axes will change dynamically with the shape and position of the reference path. The Frenet coordinate system is based on the reference path (such as the road centerline, planned trajectory) and consists of three orthogonal coordinate axes, including the tangent direction, that is, the tangent direction along the reference path; the normal direction, perpendicular to the tangent direction, pointing to the normal direction of the path; the vertical direction, perpendicular to the first two directions (that is, the height direction in three-dimensional space), which is often ignored in two-dimensional problems.

[0097] The preset obstacle avoidance safety distance can be understood as a pre-set safety distance threshold, which is used to improve the safety of obstacle avoidance.

[0098] Based on this, by determining the obstacle avoidance boundary corresponding to the longitudinal obstacle and the obstacle avoidance direction of the vehicle's oblique obstacle avoidance, the distance between the obstacle avoidance boundary and the vehicle path reference line is determined as the offset boundary distance, that is, the lateral distance between the obstacle boundary and the lane reference line is determined, so as to determine the distance and value between the offset boundary distance, the obstacle avoidance safety distance and the vehicle width distance, and obtain the target offset distance. That is, through the vehicle width, the distance between the obstacle and the current reference line, and the safety distance reserved in advance for obstacle avoidance, the distance between the vehicle and the reference line after successful obstacle avoidance is determined, so that the vector angle can be further determined based on the target offset distance.

[0099] Based on the feasible implementation of the above S120, the present application further provides a method for determining a vector angle based on a target offset distance and a path reference line, including the following steps:

[0100] According to the Frenet coordinate system, determine the initial point of the path reference line and the lower boundary of the longitudinal obstacle, and determine the distance between the initial point and the lower boundary as the obstacle boundary distance;

[0101] Determine the coordinates of the target offset point based on the target offset distance and the obstacle boundary distance, and determine the target vector based on the coordinates of the target offset point and the coordinates corresponding to the center point of the vehicle;

[0102] Determine the projection point of the vehicle on the path reference line and determine the projection heading angle corresponding to the projection point;

[0103] Determine the vector angle based on the target vector and the projection orientation angle.

[0104] The initial point of the path reference line refers to the starting position of the path reference line. It is the reference point for constructing the path reference line and is usually used to define the starting point, direction, and subsequent geometric characteristics of the path. It can be determined through path continuation and trajectory tracking. The end point of the existing path is taken as the initial point of the new path, inheriting its tangent direction. For example, in a lane keeping scenario, the initial point of the reference line is the lane centerline point corresponding to the vehicle's current position.

[0105] The lower boundary of the longitudinal obstacle may be understood as the boundary of the longitudinal obstacle that is closest to the vehicle in the current driving path of the vehicle.

[0106] The target offset point is the point obtained by adjusting the position of the points on the driving path based on the reference line and taking into account the obstacle avoidance requirements. It is used to more accurately describe the vehicle's expected driving position or target position. It can be understood as the corresponding driving point after the vehicle circumvents the obstacle.

[0107] The center point of the vehicle is the center point of the front of the vehicle.

[0108] The target vector is a vector determined based on the coordinates of the target offset point and the vehicle's center point. It is used to subsequently determine whether the vehicle can directly maneuver around obstacles at an angle at the current distance and angle. The projected heading angle refers to the angular representation of the vehicle's heading direction on the reference line. It is used to describe the vehicle's movement direction or posture in a two-dimensional plane. The vector angle can then be determined based on the target vector and the projected heading angle, thereby further determining whether the vehicle needs to retreat a certain distance, thereby establishing driving conditions that allow it to maneuver around obstacles at an angle.

[0109] Based on this, by determining the coordinates of the target offset point corresponding to the vehicle after avoiding the obstacle, the target vector is determined based on the coordinates of the target offset point and the vehicle center point. That is, the vector between the current vehicle and the ideal vehicle after avoiding the obstacle is determined. Then, the vector angle is determined based on the projection of the target vector and the current vehicle on the reference line.

[0110] Based on the feasible implementation of the above S130, the present application further provides a method for determining the corresponding backward distance of the vehicle based on the vector angle, the vehicle's wheel steering angle threshold, and the target offset distance, including the following steps:

[0111] Determining a wheel steering angle threshold according to a maximum wheel steering angle of the vehicle and a preset steering coefficient, and comparing the vector angle with the wheel steering angle threshold to obtain a comparison result;

[0112] If the comparison result shows that the vector angle is not greater than the wheel steering angle threshold, the retreat distance is determined to be zero;

[0113] If the comparison result shows that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined based on the target offset distance.

[0114] The maximum steering angle of the wheel is the maximum angle that the wheel can turn. The preset steering coefficient is a pre-set coefficient, for example, it can be 0.9. The maximum steering angle of the wheel can be expressed as follows: Indicates that the wheel steering angle threshold is 0.9 .

[0115] Based on this, by determining the wheel steering angle threshold, the vector angle and the wheel steering angle threshold are compared. If the vector angle is not greater than the wheel steering angle threshold, the vehicle can directly avoid obstacles diagonally under the current driving conditions, that is, the retreat distance is zero; if the vector angle is greater than the wheel steering angle threshold, the vehicle cannot directly avoid obstacles diagonally under the current driving conditions, and the retreat distance needs to be determined based on the target offset distance.

[0116] Based on the feasible implementation of the above S130, the present application further provides that if the comparison result shows that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined according to the target offset distance, including the following steps:

[0117] Determine the vehicle boundary distance between the upper boundary of the vehicle and the initial reference point of the path reference line according to the Frenet coordinate system, and determine the obstacle boundary distance corresponding to the longitudinal obstacle;

[0118] Determine the target driving distance based on the difference between the vehicle boundary distance, the obstacle boundary distance and the preset driving safety distance;

[0119] The retreat distance is determined based on the target driving distance, the target offset distance and the wheel steering angle threshold.

[0120] The upper boundary of the vehicle is the boundary of the front part; the target driving distance can be understood as the longitudinal distance between the front of the IGV and the rear of the obstacle, which can be expressed as If it is expressed as follows:

[0121] obs_start_s-buffer_lon-adc_end_s;

[0122] obs_start_s is the vehicle boundary distance, which can be understood as the longitudinal distance between the lower boundary of the obstacle and the starting point of the reference line in the Frenet coordinate system.

[0123] adc_end_s is the obstacle boundary distance, which can be understood as the longitudinal distance between the upper boundary of the IGV and the starting point of the reference line in the Frenet coordinate system.

[0124] buffer_lon is a preset driving safety distance, that is, a pre-set distance used to improve driving safety, which can be understood as a parameter-adjustable longitudinal safety distance from obstacles.

[0125] The target driving distance can be Indicates that

[0126]

[0127] is the target offset distance, is the wheel steering angle threshold, and tan is the tangent function in the trigonometric function.

[0128] Based on this, if the vector angle exceeds 0.9 times the corresponding maximum wheel steering angle, it means that the diagonal driving mode cannot be used in the current scenario. It is necessary to first construct the conditions for diagonal obstacle avoidance. That is, based on the current target driving distance, target offset distance and wheel steering angle threshold, the vehicle's reverse distance is calculated, and the vehicle's diagonal trajectory is planned.

[0129] Based on the feasible implementation of the above S140, the present application further provides a method for determining the starting point of the oblique movement according to the retreat distance, and determining the oblique movement trajectory of the vehicle according to the oblique movement pattern and the quadratic programming algorithm, including the following steps:

[0130] According to the oblique pattern, determine the penalty coefficient and penalty weight corresponding to the quadratic programming algorithm;

[0131] The oblique trajectory is determined according to the target offset distance, penalty coefficient, penalty weight and quadratic programming algorithm.

[0132] Among them, in the quadratic programming algorithm, the penalty coefficient (Penalty Coefficient) and penalty weight (Penalty Weight) usually appear in constrained optimization problems, especially when the penalty function method (Penalty Function Method) is used to convert the constrained problem into an unconstrained problem; the core idea of ​​the penalty function method is to convert the original constrained optimization problem into an unconstrained optimization problem by adding a penalty term to the objective function. If the iteration point violates the constraint condition, the penalty term will "penalize" (increase or decrease) the objective function value, forcing the optimization process to tend to satisfy the constraint.

[0133] For example:

[0134]

[0135] In the above formula, It can be understood as the objective function. , They represent lateral displacement s, lateral velocity v, and lateral acceleration a respectively, represents the acceleration, Represent the weights of lateral displacement, velocity, acceleration, and jerk, respectively. Represents the weight of the horizontal distance difference from the target reference line, is the lateral displacement of the reference line (here the reference line of the target final state). Additional penalty coefficient , that is, the penalty weight of the final state, Additional penalty coefficient ,right Additional penalty coefficient .

[0136] Constraints:

[0137] Boundary Constraints:

[0138]

[0139]

[0140]

[0141]

[0142] Where, 、 They represent the lower limit of lateral displacement, the lower limit of lateral velocity, the lower limit of lateral acceleration, and the lower limit of lateral jerk respectively; 、 、 、 They represent the upper limit of lateral displacement, the upper limit of lateral velocity, the upper limit of lateral acceleration, and the upper limit of lateral jerk respectively; Represents a very small unit lateral displacement, Indicates the change in acceleration per unit lateral displacement.

[0143] Continuity constraints:

[0144]

[0145] =0

[0146] Where, Expressed as The lateral velocity of the next point; Represents a very small unit lateral displacement, Represents the square of a very small unit lateral displacement.

[0147] Initial condition constraints:

[0148]

[0149]

[0150]

[0151] Where, 、 、 Represents the lateral displacement, velocity and acceleration of the first point; 、 、 They represent the initial lateral displacement, initial lateral velocity and initial lateral acceleration of the planning starting point respectively.

[0152] Based on this, the final state offset strategy is used to solve the path with the help of QP (quadratic programming) algorithm to generate an accurate and feasible planning path.

[0153] Please refer to Figure 4 , Figure 4 A schematic diagram of the architecture of a vehicle obstacle avoidance method based on scene recognition provided in an embodiment of the present application; Figure 4 As shown in the figure, an intelligent scene recognition system has been designed to automatically determine whether to activate the figure-eight steering mode by accurately analyzing the traversable space at the boundary of the drivable path. In special scenarios such as curves, when the vehicle's heading angle exceeds a threshold relative to the reference line's heading angle, and when obstacles exist in both left and right lanes and do not overlap longitudinally, when the lateral distance is less than the IGV width, and when the longitudinal distance is less than the IGV length + 5m, the system accurately identifies and switches to the figure-eight steering mode. Furthermore, in scenarios with longitudinal intersections and lateral misalignments, the system uses the oblique steering mode for obstacle avoidance. Intelligently switching between obstacle avoidance modes in different complex scenarios improves the flexibility and adaptability of the IGV's obstacle avoidance.

[0154] In some embodiments of the present application, a preset recognition device is used to identify the lateral and longitudinal obstacles corresponding to the vehicle, so that the corresponding obstacle avoidance method is determined according to the distance between the current vehicle and the obstacle. That is, in special scenarios where there are obstacles in the left and right lanes and they do not overlap longitudinally, the lateral distance is less than the IGV width, and the longitudinal distance is less than the IGV length + 5m, accurate identification is performed and the vehicle is switched to the figure-eight steering mode. For scenarios of longitudinal intersection and lateral misalignment, the oblique mode is used to avoid the obstacle, and further, based on the angle and distance information between the current vehicle and the longitudinal obstacle, it is determined whether the oblique mode can be directly used under the current driving conditions. If not, it is necessary to reconstruct the conditions for oblique obstacle avoidance based on the target driving distance, target offset distance and wheel steering angle threshold, calculate the vehicle's reverse distance, and perform secondary trajectory planning for oblique obstacle avoidance based on the secondary planning algorithm.

[0155] This intelligently switches between obstacle avoidance modes in different complex scenarios, improving the IGV's flexibility and adaptability. A preliminary safety condition check mechanism is incorporated into the oblique obstacle avoidance process. Based on the obstacle's boundaries and traversable space, the system obtains a target offset, projects the target point, and calculates the heading angle difference between the corresponding target point and the reference line projection point. If this exceeds a certain threshold (0.9 times the maximum wheel steering angle), the system establishes a condition for oblique obstacle avoidance. The system then calculates the vehicle's reverse distance, replans the reverse trajectory, and then proceeds to oblique mode planning. Furthermore, the system employs a final state offset strategy and a QP-based path planning solution. Different collision detection methods are used for different obstacle avoidance modes (figure-of-eight and oblique). Figure-of-eight uses the path, curvature, heading angle, structure, and envelope, while oblique uses the reference line projection point, heading angle, structure, and envelope). Finally, a safe and reliable trajectory is output, effectively improving the safety of oblique obstacle avoidance and the rationality of path planning.

[0156] Figure 5 Schematic diagram of the structure of a vehicle obstacle avoidance device 500 based on scene recognition provided in an embodiment of the present application. Figure 5 As shown, the scene recognition-based vehicle obstacle avoidance device 500 includes: a mode determination module 510, a vector angle determination module 520, a retreat distance determination module 530, a trajectory determination module 540, and an obstacle avoidance module 550; wherein:

[0157] Mode determination module 510 is configured to determine, based on a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and a longitudinal obstacle and its longitudinal distance from the vehicle, and determine an obstacle avoidance mode for the vehicle based on the lateral distance, longitudinal distance, vehicle width, and vehicle length.

[0158] A vector angle determination module 520 is configured to determine a target offset distance based on the obstacle avoidance boundary of the longitudinal obstacle when the vehicle's obstacle avoidance mode is the oblique travel mode, and to determine a vector angle based on the target offset distance and the path reference line corresponding to the vehicle;

[0159] A retreat distance determination module 530 is configured to determine a corresponding retreat distance of the vehicle based on the target offset distance, the vector angle, and the vehicle's wheel steering angle threshold;

[0160] The trajectory determination module 540 is used to determine the starting point of the oblique movement according to the backward distance, and determine the oblique movement trajectory of the vehicle according to the oblique movement pattern and the quadratic programming algorithm;

[0161] The obstacle avoidance module 550 is used to avoid longitudinal obstacles according to the oblique starting point and the oblique trajectory.

[0162] In the embodiment of the present application, the mode determination module 510 may also be specifically configured to:

[0163] Determine a target distance threshold according to the sum of the vehicle length and the preset distance;

[0164] Compare the lateral distance and vehicle width distance, as well as the longitudinal distance and target distance threshold to obtain a comparison result;

[0165] If the comparison result shows that the lateral distance is less than the vehicle width distance, and the longitudinal distance is less than the target distance threshold, the obstacle avoidance mode is determined to be the oblique driving mode;

[0166] If the comparison result is that the lateral distance is not less than the vehicle width distance, or the longitudinal distance is not less than the target distance threshold, the obstacle avoidance mode is determined to be the figure-eight obstacle avoidance mode.

[0167] In the embodiment of the present application, the vector angle determination module 520 may also be specifically configured to:

[0168] According to the frenet coordinate system, determine the path reference line corresponding to the vehicle;

[0169] Determine the vehicle's obstacle avoidance direction based on the path reference line and the longitudinal obstacle, and determine the boundary of each boundary of the longitudinal obstacle that corresponds to the obstacle avoidance direction as the obstacle avoidance boundary;

[0170] Determine the distance between the obstacle avoidance boundary and the path reference line as the offset boundary distance, and determine the preset obstacle avoidance safety distance;

[0171] Determine the target offset distance based on the sum of the offset boundary distance, the preset obstacle avoidance safety distance, and the vehicle width distance;

[0172] Determine the vector angle based on the target offset distance and the path reference line.

[0173] In the embodiment of the present application, the vector angle determination module 520 may also be specifically configured to:

[0174] According to the Frenet coordinate system, determine the initial point of the path reference line and the lower boundary of the longitudinal obstacle, and determine the distance between the initial point and the lower boundary as the obstacle boundary distance;

[0175] Determine the coordinates of the target offset point based on the target offset distance and the obstacle boundary distance, and determine the target vector based on the coordinates of the target offset point and the coordinates corresponding to the center point of the vehicle;

[0176] Determine the projection point of the vehicle on the path reference line and determine the projection heading angle corresponding to the projection point;

[0177] Determine the vector angle based on the target vector and the projection orientation angle.

[0178] In the embodiment of the present application, the retreat distance determination module 530 may also be specifically configured to:

[0179] Determining a wheel steering angle threshold according to a maximum wheel steering angle of the vehicle and a preset steering coefficient, and comparing the vector angle with the wheel steering angle threshold to obtain a comparison result;

[0180] If the comparison result shows that the vector angle is not greater than the wheel steering angle threshold, the retreat distance is determined to be zero;

[0181] If the comparison result shows that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined based on the target offset distance.

[0182] In the embodiment of the present application, the retreat distance determination module 530 may also be specifically configured to:

[0183] Determine the vehicle boundary distance between the upper boundary of the vehicle and the initial reference point of the path reference line according to the Frenet coordinate system, and determine the obstacle boundary distance corresponding to the longitudinal obstacle;

[0184] Determine the target driving distance based on the difference between the vehicle boundary distance, the obstacle boundary distance and the preset driving safety distance;

[0185] The retreat distance is determined based on the target driving distance, the target offset distance and the wheel steering angle threshold.

[0186] In the embodiment of the present application, the trajectory determination module 540 may also be specifically configured to:

[0187] According to the oblique pattern, determine the penalty coefficient and penalty weight corresponding to the quadratic programming algorithm;

[0188] The oblique trajectory is determined according to the target offset distance, penalty coefficient, penalty weight and quadratic programming algorithm.

[0189] Figure 6 A schematic diagram of the structure of a device for executing a vehicle obstacle avoidance method based on scene recognition according to an embodiment of the present application is provided in the embodiment of the present application. Figure 6 As shown, the device 600 includes:

[0190] The device 600 may include one or more processors 601 , one or more computer-readable storage media memories 602 , a communication component 603 , and other components. The processor 601 , the memory 602 , and the communication component 603 are connected via a bus 604 .

[0191] During the specific implementation process, at least one processor 601 executes the computer-executable instructions stored in the memory 602, so that the at least one processor 601 executes the above-mentioned vehicle obstacle avoidance method based on scene recognition.

[0192] The specific implementation process of the processor 601 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0193] Furthermore, the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), or application-specific integrated circuits (ASICs). A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in this application may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0194] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0195] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0196] In some embodiments, a computer program product is further proposed, comprising a computer program or instructions, which, when executed by a processor, implements the steps of any of the above-mentioned scene recognition-based vehicle obstacle avoidance methods.

[0197] The specific implementation of the above operations can be found in the previous embodiments and will not be repeated here.

[0198] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be accomplished by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0199] To this end, an embodiment of the present application provides a computer-readable storage medium, in which multiple program codes are stored. The program codes can be loaded by a processor to execute the steps of any scene recognition-based vehicle obstacle avoidance method provided in the embodiment of the present application.

[0200] The storage medium may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0201] According to one aspect of the present application, a computer program product or a computer program is provided. The computer program product or the computer program comprises computer instructions stored in a computer-readable storage medium.

[0202] Since the instructions stored in the storage medium can execute the steps of any of the scene recognition-based vehicle obstacle avoidance methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the scene recognition-based vehicle obstacle avoidance methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0203] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered merely as exemplary, and the true scope and spirit of the present application are indicated by the appended claims.

[0204] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A vehicle obstacle avoidance method based on scene recognition, characterized in that: The method comprises: According to a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, as well as a longitudinal obstacle and its longitudinal distance from the vehicle are determined. Based on the lateral distance, the longitudinal distance, the vehicle width, and the vehicle length, an obstacle avoidance mode for the vehicle is determined, specifically: Determining a target distance threshold according to a sum of the vehicle length and a preset distance; Comparing the lateral distance with the vehicle width distance, and the longitudinal distance with the target distance threshold, to obtain a comparison result; If the comparison result shows that the lateral distance is less than the vehicle width distance, and the longitudinal distance is less than the target distance threshold, determining that the obstacle avoidance mode is the oblique driving mode; If the comparison result is that the lateral distance is not less than the vehicle width distance, or the longitudinal distance is not less than the target distance threshold, determining that the obstacle avoidance mode is a figure-eight obstacle avoidance mode; When the obstacle avoidance mode of the vehicle is the oblique mode, a target offset distance is determined according to the obstacle avoidance boundary of the longitudinal obstacle, and a vector angle is determined according to the target offset distance and the path reference line corresponding to the vehicle, specifically: Determining a path reference line corresponding to the vehicle according to the Frenet coordinate system; determining an obstacle avoidance direction of the vehicle according to the path reference line and the longitudinal obstacle, and determining a boundary corresponding to the obstacle avoidance direction among the boundaries of the longitudinal obstacle as the obstacle avoidance boundary; Determining the distance between the obstacle avoidance boundary and the path reference line as the offset boundary distance, and determining a preset obstacle avoidance safety distance; Determining the target offset distance according to the sum of the offset boundary distance, the preset obstacle avoidance safety distance, and the vehicle width distance; Determining, according to the Frenet coordinate system, an initial point of the path reference line and a lower boundary of the longitudinal obstacle, and determining a distance between the initial point and the lower boundary as an obstacle boundary distance; Determining the coordinates of a target offset point based on the target offset distance and the obstacle boundary distance, and determining a target vector based on the coordinates of the target offset point and the coordinates corresponding to the center point of the vehicle; Determining a projection point of the vehicle on the path reference line, and determining a projection heading angle corresponding to the projection point; Determining the vector angle according to the target vector and the projection orientation angle; determining a corresponding backward distance of the vehicle according to the target offset distance, the vector angle, and a wheel steering angle threshold of the vehicle; Determining the starting point of the oblique movement according to the retreat distance, and determining the oblique movement trajectory of the vehicle according to the oblique movement pattern and a quadratic programming algorithm; The longitudinal obstacle is circumvented according to the oblique starting point and the oblique trajectory.

2. The method according to claim 1, characterized in that The determining, based on the vector angle, the wheel steering angle threshold of the vehicle, and the target offset distance, of a corresponding backward distance of the vehicle includes: determining the wheel steering angle threshold according to the maximum wheel steering angle of the vehicle and a preset steering coefficient, and comparing the vector angle with the wheel steering angle threshold to obtain a comparison result; If the comparison result shows that the vector angle is not greater than the wheel steering angle threshold, determining that the retreat distance is zero; If the comparison result is that the vector angle is greater than the wheel steering angle threshold, the retreat distance is determined according to the target offset distance.

3. The method according to claim 2, characterized in that If the comparison result shows that the vector angle is greater than the wheel steering angle threshold, determining the retreat distance according to the target offset distance includes: determining a vehicle boundary distance between an upper boundary of the vehicle and an initial reference point of the path reference line according to a Frenet coordinate system, and determining an obstacle boundary distance corresponding to the longitudinal obstacle; determining a target driving distance according to a difference between the vehicle boundary distance, the obstacle boundary distance, and a preset driving safety distance; The retreat distance is determined according to the target driving distance, the target offset distance, and the wheel steering angle threshold.

4. The method according to claim 1, wherein The step of determining the starting point of the oblique movement according to the backward distance, and determining the oblique movement trajectory of the vehicle according to the oblique movement pattern and a quadratic programming algorithm, includes: Determining a penalty coefficient and a penalty weight corresponding to the quadratic programming algorithm according to the oblique movement pattern; The oblique trajectory is determined according to the target offset distance, the penalty coefficient and the penalty weight and the quadratic programming algorithm.

5. A vehicle obstacle avoidance device based on scene recognition, characterized in that: The device comprises: The mode determination module is configured to determine, based on a preset recognition device, a lateral obstacle corresponding to the vehicle and its lateral distance from the vehicle, and a longitudinal obstacle and its longitudinal distance from the vehicle, and determine an obstacle avoidance mode for the vehicle based on the lateral distance, the longitudinal distance, the vehicle width, and the vehicle length, specifically: Determining a target distance threshold according to a sum of the vehicle length and a preset distance; Comparing the lateral distance with the vehicle width distance, and the longitudinal distance with the target distance threshold, to obtain a comparison result; If the comparison result shows that the lateral distance is less than the vehicle width distance, and the longitudinal distance is less than the target distance threshold, determining that the obstacle avoidance mode is the oblique driving mode; If the comparison result is that the lateral distance is not less than the vehicle width distance, or the longitudinal distance is not less than the target distance threshold, determining that the obstacle avoidance mode is a figure-eight obstacle avoidance mode; The vector angle determination module is configured to determine a target offset distance based on the obstacle avoidance boundary of the longitudinal obstacle when the obstacle avoidance mode of the vehicle is the oblique mode, and to determine a vector angle based on the target offset distance and the path reference line corresponding to the vehicle, specifically: Determining a path reference line corresponding to the vehicle according to the Frenet coordinate system; determining an obstacle avoidance direction of the vehicle according to the path reference line and the longitudinal obstacle, and determining a boundary corresponding to the obstacle avoidance direction among the boundaries of the longitudinal obstacle as the obstacle avoidance boundary; Determining the distance between the obstacle avoidance boundary and the path reference line as the offset boundary distance, and determining a preset obstacle avoidance safety distance; Determining the target offset distance according to the sum of the offset boundary distance, the preset obstacle avoidance safety distance, and the vehicle width distance; Determining, according to the Frenet coordinate system, an initial point of the path reference line and a lower boundary of the longitudinal obstacle, and determining a distance between the initial point and the lower boundary as an obstacle boundary distance; Determining the coordinates of a target offset point based on the target offset distance and the obstacle boundary distance, and determining a target vector based on the coordinates of the target offset point and the coordinates corresponding to the center point of the vehicle; Determining a projection point of the vehicle on the path reference line, and determining a projection heading angle corresponding to the projection point; Determining the vector angle according to the target vector and the projection orientation angle; a retreat distance determination module, configured to determine a retreat distance corresponding to the vehicle based on the target offset distance, the vector angle, and a wheel steering angle threshold of the vehicle; a trajectory determination module, configured to determine a starting point of the oblique movement according to the backward distance, and to determine an oblique movement trajectory of the vehicle according to the oblique movement pattern and a quadratic programming algorithm; The obstacle avoidance module is used to avoid the longitudinal obstacle according to the oblique starting point and the oblique trajectory.

6. A computer device, characterized in that: include: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in a memory and configured to be executed by one or more processors, the one or more programs being configured to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores program codes, which can be called by a processor to execute the method according to any one of claims 1 to 4.

Images