Local path planning method and device, equipment and storage medium

Abstract

The invention discloses a local path planning method, device and equipment and a storage medium, and the method comprises the steps: generating a plurality of reference trajectories based on a target trajectory of vehicle driving, sequentially carrying out the obstacle collision detection and loss function value comparison of the reference trajectories, screening out an optimal trajectory, introducing a lane changing mechanism, and when the optimal trajectory has a lane changing condition, carrying out the lane changing of the vehicle. According to the method, a plurality of lane changing tracks are generated, the driving behavior of the vehicle on a structured road is better fitted, and finally, the lane changing track which does not have obstacle collision and has the shortest path is selected from the plurality of lane changing tracks as the local path of the vehicle, so that the requirements of vehicle constraint conditions and vehicle driving scenes can be met; and the optimal local path can be provided according to the position of the vehicle, and the vehicle passing efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a local path planning method, device, equipment and storage medium. Background Art

[0002] Local path planning is one of the key technologies in the field of autonomous driving. Local path planning refers to avoiding obstacles or other restrictions by selecting appropriate paths in the case of a known environment map, so that the target object can find the best path from the starting point to the end point within the specified range. At present, most local path planning algorithms are mainly based on robot path planning, which can be widely used in indoor, outdoor, structured and unstructured environments, but are not applicable to traffic environments based on structured roads, such as urban roads and highways. In addition, the operating carrier of the traffic environment of structured roads is generally a vehicle, which has specific dynamic characteristics and comfort requirements. Local path planning needs to consider these constraints. The vehicle needs to adapt to vehicle driving scenarios such as lane keeping, lane changing, and overtaking. The vehicle driving scenario is different from the simple obstacle avoidance of the robot. Since the traditional robot local path planning method does not require vehicle constraints such as dynamic characteristics and vehicle comfort, and does not consider the vehicle driving scenario, it is not applicable to the vehicle's local path planning process. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide a local path planning method, device, equipment and storage medium that can meet the requirements of vehicle constraints and vehicle driving scenarios, and can provide an optimal local path according to the location of the vehicle to improve vehicle travel efficiency.

[0004] To achieve the above object, an embodiment of the present invention provides a local path planning method, including:

[0005] Generate a number of reference trajectories based on the target trajectory of the vehicle;

[0006] Performing obstacle collision detection on each of the reference trajectories to select a candidate trajectory without obstacle collision from a plurality of reference trajectories;

[0007] When the number of the candidate trajectories is greater than 1, a loss function value of each candidate trajectory is calculated, and an optimal trajectory is selected from the candidate trajectories according to the loss function value; wherein the loss function value is used to evaluate the trajectory performance of the candidate trajectories;

[0008] When there is an adjacent lane at the location of the vehicle, and the adjacent lane of the optimal trajectory allows lane change, generating a plurality of lane change trajectories according to lane parameters of the adjacent lane;

[0009] A lane changing trajectory without obstacle collision and with the shortest path is selected from the plurality of lane changing trajectories as a local path of the vehicle.

[0010] As an improvement of the above solution, the loss function value is calculated by a loss function, and the input parameters of the loss function include: the product of the trajectory evaluation value and its corresponding weight value, and the ratio of the minimum distance value between the candidate trajectory and the obstacle and its weight value.

[0011] As an improvement of the above solution, the trajectory evaluation value includes a lateral trajectory evaluation value, a longitudinal trajectory evaluation value and a trajectory consistency evaluation value; wherein,

[0012] The lateral trajectory evaluation value is calculated according to the path planning duration and distance of the candidate trajectory, and the weight value of the lateral trajectory evaluation value is a preset first parameter value;

[0013] The longitudinal trajectory evaluation value is calculated according to the path planning duration and the vehicle speed difference parameter of the candidate trajectory, and the weight value of the longitudinal trajectory evaluation value is calculated according to the target speed and the current speed of the vehicle;

[0014] The trajectory consistency evaluation value is calculated according to the candidate trajectory and the average curvature and the curve offset value of the local path calculated last time, and the weight value of the trajectory consistency evaluation value is a preset second parameter value;

[0015] The weight value of the minimum distance value is calculated based on the number of drivable lanes of the candidate path and a fixed value.

[0016] As an improvement of the above solution, the method of generating a plurality of reference trajectories based on the target trajectory of the vehicle includes:

[0017] A reference coordinate system is constructed with the target trajectory as a reference line, and a plurality of reference points are obtained by projecting the target trajectory onto the reference line;

[0018] The reference points are sampled based on a speed parameter of the vehicle, an expected road width, and a trajectory planning duration parameter to generate a plurality of reference trajectories.

[0019] As an improvement of the above solution, the generating of a plurality of lane change trajectories according to the lane parameters of the adjacent lanes includes:

[0020] Obtaining the centerline point of the adjacent lane;

[0021] Traversing each of the center line points, taking the position of the vehicle as the starting point and the center line point as the end point, and calculating the interpolation coordinates from the starting point to the end point according to a preset rotation angle and number of rotations;

[0022] Each of the starting point, the interpolation coordinates and the end point are connected respectively to generate a plurality of lane change trajectories.

[0023] As an improvement of the above solution, after performing obstacle collision detection on each reference trajectory, the method further includes:

[0024] When all reference trajectories have obstacle collisions, a plurality of lane change trajectories are generated according to lane parameters of adjacent lanes of the target trajectory.

[0025] As an improvement of the above solution, the method further includes:

[0026] When there is no adjacent lane at the location of the vehicle, or when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory does not allow lane change, the optimal trajectory is used as the local path of the vehicle.

[0027] To achieve the above object, an embodiment of the present invention further provides a local path planning device, comprising:

[0028] A reference trajectory generation module, used to generate a plurality of reference trajectories based on the target trajectory of the vehicle;

[0029] An obstacle collision detection module, used for performing obstacle collision detection on each of the reference trajectories;

[0030] A candidate trajectory selection module, used to select a candidate trajectory without obstacle collision from a number of reference trajectories;

[0031] an optimal trajectory acquisition module, used for calculating the loss function value of each candidate trajectory when the number of the candidate trajectories is greater than 1, and selecting the optimal trajectory from the candidate trajectories according to the loss function value; wherein the loss function value is used to evaluate the trajectory performance of the candidate trajectories;

[0032] A lane change trajectory generating module, configured to generate a plurality of lane change trajectories according to lane parameters of the adjacent lane when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory allows lane change;

[0033] The local path generation module is used to select a lane changing trajectory without obstacle collision and with the shortest path from the plurality of lane changing trajectories as the local path of the vehicle.

[0034] To achieve the above objectives, an embodiment of the present invention also provides a local path planning device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the local path planning method described in any of the above embodiments.

[0035] To achieve the above objectives, an embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the local path planning method described in any of the above embodiments.

[0036] Compared with the prior art, the local path planning method, device, equipment and storage medium disclosed in the present invention generate several reference trajectories based on the target trajectory of vehicle driving, perform obstacle collision detection and loss function value comparison on these reference trajectories in turn, and then screen out the optimal trajectory, introduce a lane changing mechanism, and when this optimal trajectory meets the lane changing conditions, generate several lane changing trajectories, which are more in line with the driving behavior of the vehicle on structured roads. Finally, the lane changing trajectory with no obstacle collision and the shortest path is selected from the several lane changing trajectories as the local path of the vehicle, which can meet the requirements of vehicle constraints and vehicle driving scenarios, and can provide the optimal local path according to the position of the vehicle, thereby improving the vehicle traffic efficiency.

[0037] In addition, a loss function is designed to evaluate the trajectory performance of the candidate trajectory. The trajectory evaluation value and the minimum distance value between the candidate trajectory and the obstacle are introduced to construct the loss function. The trajectory evaluation value takes into account the lateral trajectory, longitudinal trajectory and trajectory consistency evaluation. The evaluation of the lateral trajectory and the longitudinal trajectory can make the selected optimal trajectory conform to the actual driving scenario of the vehicle, and the trajectory consistency is evaluated to ensure the continuity of the front and rear planned routes, prevent sudden changes in the planned route, and achieve path comfort and smoothness. In addition, the loss function takes into account that the optimal path will not be close to the obstacle, and maintain a certain offset distance to ensure obstacle avoidance safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of a local path planning method provided by an embodiment of the present invention;

[0039] Figure 2 It is a reference trajectory effect diagram of frenet coordinate system sampling provided by an embodiment of the present invention;

[0040] Figure 3 It is a local path planning effect diagram provided by an embodiment of the present invention;

[0041] Figure 4 It is another local path planning effect diagram provided by an embodiment of the present invention;

[0042] Figure 5 is another flow chart of a local path planning method provided by an embodiment of the present invention;

[0043] Figure 6is a structural block diagram of a local path planning device provided by an embodiment of the present invention;

[0044] Figure 7 It is a structural block diagram of a local path planning device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] See also Figure 1 , Figure 1 : is a flow chart of a local path planning method provided by an embodiment of the present invention, wherein the local path planning method comprises:

[0047] S1. Generate several reference trajectories based on the target trajectory of the vehicle;

[0048] S2. performing obstacle collision detection on each of the reference trajectories to select a candidate trajectory without obstacle collision from a plurality of reference trajectories;

[0049] S3. When the number of the candidate trajectories is greater than 1, calculating the loss function value of each candidate trajectory, and selecting the optimal trajectory from the candidate trajectories according to the loss function value;

[0050] S4. When there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory allows lane change, generate a plurality of lane change trajectories according to lane parameters of the adjacent lane;

[0051] S5. Selecting a lane changing trajectory without obstacle collision and with the shortest path from the plurality of lane changing trajectories as a local path of the vehicle.

[0052] It is worth noting that the local path planning method described in the embodiment of the present invention is implemented by a cloud platform, which can obtain vehicle information and perform local path planning, and then send the planned local path to the vehicle, instructing the vehicle to travel along the local path.

[0053] Specifically, in step S1, the target trajectory of the vehicle is used to generate a plurality of reference trajectories, including: constructing a reference coordinate system with the target trajectory as a reference line, projecting the target trajectory onto the reference line to obtain a plurality of reference points; sampling the reference points based on the vehicle's speed parameters, the expected road width, and the trajectory planning duration parameters to generate a plurality of reference trajectories.

[0054] First, configure the local path planning parameters, which include the vehicle's speed parameters, expected road width, and trajectory planning duration parameters. The speed parameters include the maximum speed MAX_SPEED, the maximum acceleration MAX_ACCEL, the target speed sampling interval D_T_S, the target speed TARGET_SPEED, and the current speed CUR_SPEED; the target speed is the speed that the vehicle needs to reach in this planning. The expected road width includes the maximum road width MAX_WIDTH and the road sampling width D_WIDTH; wherein, the maximum road width is not the lane width, but the path width planned or expected by the local planning algorithm. The maximum lane width can be similar to the actual lane width of the lane where the vehicle is located, or the total width of the three lanes on the left and right sides of the vehicle and the vehicle itself can also be set. In this way, the planned path will not allow the vehicle to cross several lanes from the current lane. Driving across multiple lanes continuously is relatively dangerous; the trajectory planning duration parameters include: planning prediction maximum time MAXT, planning prediction minimum time MINT, planning prediction sampling interval DT; wherein, the planning prediction maximum / minimum time indicates how long it takes to plan the path. The longer the planning prediction time, the longer the reference trajectory.

[0055] Secondly, the target trajectory of the vehicle is obtained according to the current position of the vehicle. This target trajectory is the trajectory that the vehicle is planned to travel in the global path navigation of the vehicle. The point column data of the grid obstacles around the vehicle are obtained according to the current position of the vehicle. The obstacles refer to the traffic participants in the vehicle's driving process, such as vehicles, pedestrians, cone barrels, etc., which are all considered obstacles. The grid obstacles are generated using the grid method. The specific process of generating obstacles using the grid method can refer to the existing technology and will not be repeated here.

[0056] Furthermore, a reference coordinate system is constructed with the target trajectory as a reference line. The reference coordinate system is a frenet coordinate system. The trajectory points of the target trajectory are projected onto the reference line to obtain reference points. The direction along the reference line is the longitudinal axis, and the direction perpendicular to the reference line is the transverse axis. The frenet coordinate system is established based on the target trajectory. In this coordinate system, any point can be regarded as a linear combination of a transverse unit vector and a longitudinal unit vector. For example, see Figure 2 , Figure 2 is a reference trajectory effect diagram of the frenet coordinate system sampling provided by an embodiment of the present invention, Figure 2 The solid circle in the middle is the reference point obtained by projecting the trajectory point of the target trajectory onto the reference line, and the lines on both sides of this target trajectory are the reference trajectories obtained after sampling the reference point.

[0057] Finally, the reference points are sampled based on the vehicle's speed parameters, expected road width, and trajectory planning duration parameters to construct a reference trajectory. The method for forming a reference trajectory is as follows: in the range of [-MAX_WIDTH, MAX_WIDTH], sampling is performed incrementally according to D_WIDTH to obtain a road width sampling set; on the basis of the road width sampling set, the planned prediction time [MINT, MAXT] is used as the range, and sampling is performed incrementally according to the planned sampling interval DT to obtain a prediction time sampling set; on the basis of the prediction time sampling set, in the range of [CUR_SPEED, TARGET_SPEED], a speed sampling set is obtained according to the D_T_S sampling interval, wherein the target speed TARGET_SPEED cannot exceed the maximum speed MAX_SPEED, and the acceleration of the target speed TARGET_SPEED and the current speed CUR_SPEED cannot exceed the maximum acceleration MAX_ACCEL. The reference points of the target trajectory are first-order, second-order, and third-order derivatives are taken according to the road width sampling values ​​in the road width sampling set, the predicted time sampling values ​​in the predicted time sampling set, and the speed sampling values ​​in the speed sampling set to form a plurality of reference trajectories.

[0058] Exemplarily, the first-order, second-order, and third-order derivatives of the reference point of the target trajectory are performed as follows:

[0059] The first derivative usually represents velocity and can be described by the following formula:

[0060]

[0061]

[0062] Wherein, L_dot represents the first-order lateral derivative, L is the lateral displacement, t is the time, which is obtained from the predicted time sampling value in the predicted time sampling set; S_dot represents the first-order longitudinal derivative, S is the longitudinal displacement. The longitudinal displacement and the lateral displacement need to be within the range of the road width sampling set, and the speed obtained by the derivative needs to be within the range of the speed sampling set.

[0063] The second derivative represents acceleration and can be described by the following formula:

[0064]

[0065]

[0066] Among them, L_ddot represents the second-order transverse derivative, that is, L_ddot is the derivative of L_dot with respect to time t; S_ddot represents the second-order longitudinal derivative, that is, S_ddot is the derivative of S_dot with respect to time t.

[0067] The third-order derivative represents acceleration (Jerk), which can be described by the following formula:

[0068]

[0069] Among them, S_dddot represents the third-order longitudinal derivative, that is, S_dddot is the derivative of S_ddot with respect to time t.

[0070] Thus, after taking first-order, second-order, and third-order derivatives of the reference points of the target trajectory, a plurality of reference trajectories satisfying the above formulas (1) to (5) are obtained.

[0071] Specifically, in step S2, obstacle collision detection is performed on each of the reference trajectories to select a candidate trajectory without obstacle collision from a plurality of reference trajectories.

[0072] Exemplarily, all reference tracks are traversed, collision detection is performed with obstacles around the vehicle, and reference tracks with collision are filtered. The specific process of collision detection is as follows: traverse all reference points of each reference track, and generate a rectangular frame AA based on the length and width of the vehicle body at this point; traverse the four corner points of each grid obstacle, and determine whether the corner point is within the rectangular frame AA. If so, it is determined that there is an obstacle collision on the reference track, and if not, it is determined that there is no obstacle collision on the reference track; traverse all reference tracks and detect whether there is a collision on each reference track.

[0073] Specifically, in step S3, when the number of the candidate trajectories is greater than 1, the loss function value of each candidate trajectory is calculated, and the optimal trajectory is selected from the candidate trajectories according to the loss function value.

[0074] Exemplarily, the loss function value is calculated by a loss function, and the input parameters of the loss function include: the product of the trajectory evaluation value and its corresponding weight value, the ratio of the minimum distance value between the candidate trajectory and the obstacle and its weight value; wherein the trajectory evaluation value includes a lateral trajectory evaluation value, a longitudinal trajectory evaluation value and a trajectory consistency evaluation value. The larger the loss function value, the worse the trajectory performance, indicating that this candidate trajectory is more likely to collide. Therefore, after calculating the loss function value of each candidate trajectory, the candidate trajectory with the smallest loss function value is selected as the optimal trajectory. If there are two or more identical minimum loss function values, the candidate trajectory corresponding to one of the loss function values ​​can be selected as the optimal trajectory.

[0075] Exemplarily, the loss function satisfies the following formula:

[0076] cost=KLAT*costd+KLON*costs+KCON*consistency+KOBS / minobstacledis(6);

[0077] Among them, cost is the loss function value; KLAT is the weight value of the lateral trajectory evaluation value; costd is the lateral trajectory evaluation value; KLON is the weight value of the longitudinal trajectory evaluation value; costs is the longitudinal trajectory evaluation value; KCIN is the weight value of the trajectory consistency evaluation value; consistency is the trajectory consistency evaluation value; KOBS is the weight value of the minimum distance value between the candidate trajectory and the obstacle; minobstacledis is the minimum distance value between the candidate trajectory and the obstacle.

[0078] Specifically, the lateral trajectory evaluation value costd is calculated according to the path planning duration and distance of the candidate trajectory, and the weight value KLAT of the lateral trajectory evaluation value is a preset first parameter value, which can be set according to an empirical value and is not specifically limited here.

[0079] Exemplarily, the lateral trajectory evaluation value costd satisfies the following formula:

[0080] costd = w J J t (d(t))+w d d 2 +w t T (7);

[0081] Among them, J t (d(t)) is the change amplitude within the sampling time, which is used to evaluate comfort, t is the sampling time; d is the distance between the starting point and the end point of the candidate trajectory, w d d 2 It can represent the indicator that the candidate trajectory deviates from the center line of the lane; T is the path planning time, w t T can represent the index of vehicle driving efficiency; w J 、w d 、w t is the preset weight value.

[0082] Specifically, the longitudinal trajectory evaluation value costs is calculated based on the path planning duration of the candidate trajectory and the vehicle speed difference parameter, and the longitudinal trajectory evaluation value weight KLON is calculated based on the target speed and current speed of the vehicle. The vehicle speed difference parameter includes the target configuration speed and the set expected speed.

[0083] Exemplarily, the longitudinal trajectory evaluation value costs satisfies the following calculation formula:

[0084] costs = w J J t (d(t))+w s (sl -s s ) 2 +w t T (8);

[0085] Among them, s l The target configuration speed for the vehicle when traveling on this candidate trajectory; s s To set the desired speed; w s (s l -s s ) 2 It is used to characterize the difference between the target configuration speed and the set expected speed; w s is the preset weight value.

[0086] Exemplarily, the weight value KLON of the longitudinal trajectory evaluation value satisfies the following calculation formula:

[0087] KLON=kl*|S d -S c | (9);

[0088] Among them, kl is a fixed value, which can be given according to experience; S d is the target speed of the vehicle; S c is the current speed of the vehicle.

[0089] Specifically, the trajectory consistency evaluation value consistency is calculated based on the candidate trajectory and the average curvature and curve offset value of the local path calculated last time. It is worth noting that curvature is the rotation rate of the tangent direction angle of a point on the curve to the arc length, which is defined by differentiation, indicating the degree of deviation of the curve from the straight line, and mathematically indicating the numerical value of the curvature of the curve at a certain point. The greater the curvature, the greater the curvature of the curve. In the embodiment of the present invention, the curvature of all trajectory points in the local path is calculated, and then the average curvature of the local path can be obtained after summing and averaging. Since the present invention plans a local path, it is equivalent to planning a path once every time the vehicle walks a certain distance, and the distance walked is the local path planned each time. Therefore, after walking this distance, it is necessary to plan the next local path. In this process, it is necessary to use the average curvature and curve offset value of the local path planned last time as reference factors and apply them to this local path planning. It can be understood that if this path planning is the first planning, there is no "average curvature and curve offset value of the local path calculated last time" at this time, so the trajectory consistency evaluation value consistency can be set to 0. The weight value KCON of the trajectory consistency evaluation value is a preset second parameter value, and the second parameter value can be set according to an empirical value and is not specifically limited here.

[0090] Exemplarily, the trajectory consistency evaluation value consistency satisfies the following calculation formula:

[0091] consistency=fabs(avgks-preavgks)+fabs(avgd-preavgd) (10);

[0092] Among them, fabs() is a function for finding absolute values; avgks is the average curvature of the candidate trajectory, which can be obtained by calculating the curvature of all trajectory points in the candidate path and then summing and averaging them; preavgks is the average curvature of the local path calculated last time; avgd is the average offset value of the candidate trajectory, and preavgd is the average offset value of the local path calculated last time. The consistency indicator is intended to ensure the continuity of the previous and next planned routes and prevent sudden changes in the planned routes, which may cause the vehicle to suddenly turn or accelerate at a large angle.

[0093] Specifically, the weight value KOBS of the minimum distance value is calculated according to the number of drivable lanes of the candidate path and a fixed value.

[0094] Exemplarily, the minimum distance value minobstacledis between the candidate trajectory and the obstacle satisfies the following formula:

[0095]

[0096] Among them, Wx i is the horizontal coordinate of a point in the reference point column of the candidate trajectory; i OBSx is the ordinate of a point in the reference point column of the candidate trajectory; j is the horizontal coordinate of a point in the obstacle; OBSy j is the ordinate of a point in the obstacle.

[0097] Exemplarily, the weight value KOBS of the minimum distance between the candidate trajectory and the obstacle satisfies the following calculation formula:

[0098] KOBS = ko / n (12);

[0099] Among them, kO is a fixed value that can be set according to the empirical value; n is the number of drivable lanes under the current candidate trajectory, and KOBS is used to adjust the standardization of the planning so that the candidate trajectory will not be too close to the obstacle.

[0100] Furthermore, when the number of the candidate trajectories is 1, there is no need to calculate the loss function value of this candidate trajectory, but this candidate trajectory is directly taken as the optimal trajectory.

[0101] In an embodiment of the present invention, a loss function is designed to evaluate the trajectory performance of the candidate trajectory, and a trajectory evaluation value and a minimum distance value between the candidate trajectory and the obstacle are introduced to construct the loss function. The trajectory evaluation value takes into account the lateral trajectory, the longitudinal trajectory and the trajectory consistency evaluation. The evaluation of the lateral trajectory and the longitudinal trajectory can make the selected optimal trajectory conform to the actual driving scenario of the vehicle, and the trajectory consistency is evaluated to ensure the continuity of the previous and next planned routes, prevent sudden changes in the planned route, and achieve the comfort and smoothness of the path. In addition, the loss function takes into account that the optimal path will not be close to the obstacle, and maintain a certain offset distance to ensure obstacle avoidance safety.

[0102] Specifically, in step S4, when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory allows lane change, a plurality of lane change trajectories are generated according to the lane parameters of the adjacent lane.

[0103] Exemplarily, after obtaining the optimal trajectory, determine whether the lane change conditions are met. First, determine whether there are multiple lanes at the current position of the vehicle. If not, lane change is not allowed. If so, obtain the center line point column data and lane data of the adjacent lane of the optimal trajectory. The lane data includes lane lines, and the lane lines include solid lines and dashed lines. If the lane line is a dashed line, you can change lanes to the adjacent lane. If the lane line is a solid line, you cannot change lanes to the adjacent lane. If the lane line corresponding to the lane data is a dashed line, it is necessary to further determine whether the distance between the optimal trajectory and the adjacent lane is less than a preset distance threshold. If so, it means that the distance to the adjacent lane is very close and lane change is allowed. At this time, several lane change tracks are generated according to the lane parameters of the adjacent lane, and the optimal trajectory is abandoned. If not, lane change is not allowed. It is worth noting that the distance threshold can be set according to an empirical value, such as one-fourth of the lane width.

[0104] Specifically, the generating a plurality of lane change trajectories according to the lane parameters of the adjacent lanes includes:

[0105] S41, obtaining the centerline point of the adjacent lane;

[0106] S42, traversing each of the center line points, taking the position of the vehicle as the starting point and the center line point as the end point, and calculating the interpolation coordinates from the starting point to the end point according to a preset rotation angle and number of rotations;

[0107] S43, respectively connecting each of the starting point, the interpolation coordinates and the end point to generate a plurality of lane change trajectories.

[0108] Exemplarily, take the centerline point of the adjacent lane, traverse each centerline point, such as using the horn curve as a model to generate several lane change trajectories, take the vehicle's location as the starting point, and represent the coordinates of the starting point as (x0, y0), take the centerline point as the end point, and represent the coordinates of the end point as (x1, y1), and calculate the interpolated coordinates from the starting point to the end point according to the preset rotation angle t and the number of rotations n; wherein, the value of t is 0 to 2π, the unit is radian, and each rotation angle increases by 0.26 radians on the basis of the previous rotation angle, and radians 0 to 2π are converted to angles of 0° to 360°, and each rotation angle increases by 15° on the basis of the previous rotation angle. The interpolated coordinates satisfy:

[0109] x(t)=(1-t / (2π))*x0+(t / (2π))*x1*cos(n*t);

[0110] y(t)=(1-t / (2π))*y0+(t / (2π))*y1*cos(n*t).

[0111] Exemplarily, x(t) is the horizontal coordinate of the interpolation coordinate, y(t) is the vertical coordinate of the interpolation coordinate, and after obtaining the interpolation coordinate, each of the starting points, the interpolation coordinates and the end point are connected respectively to generate a plurality of lane change trajectories.

[0112] Further, when there is no adjacent lane at the location of the vehicle, or when there is an adjacent lane at the location of the vehicle and lane change is not allowed in the adjacent lane of the optimal trajectory, the optimal trajectory is used as the local path of the vehicle.

[0113] Exemplarily, when there is no adjacent lane at the location of the vehicle, it indicates that the vehicle is driving without changing lanes. When there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory does not allow lane change, it indicates that the vehicle is driving without changing lanes. When driving without changing lanes, several lane change trajectories will not be generated based on the optimal trajectory, but the optimal trajectory will be directly used as the local path of the vehicle.

[0114] Furthermore, after performing obstacle collision detection on each of the reference trajectories in step S2, the method further includes: when all reference trajectories have obstacle collision, generating a plurality of lane change trajectories according to lane parameters of adjacent lanes of the target trajectory.

[0115] Exemplarily, if collision detection is performed on all reference trajectories and it is found that all reference trajectories will collide with obstacles, then all reference trajectories are abandoned, and the target trajectory is used as a reference. First, it is determined whether there are multiple lanes at the current position of the vehicle. If not, lane change is not allowed. If so, the center line point column data and lane data of the adjacent lane of the target trajectory are obtained. If the lane line is a dotted line, lane change to the adjacent lane is allowed. If the lane line is a solid line, lane change to the adjacent lane is not allowed. If lane change is allowed in the adjacent lane of the target trajectory, it is necessary to further determine whether the distance between the target trajectory and the adjacent lane is less than a preset distance threshold. If so, lane change is allowed. At this time, several lane change tracks are generated according to the lane parameters of the adjacent lane, and the target trajectory is abandoned. If not, lane change is not allowed. It is worth noting that the specific process of generating several lane change tracks according to the lane parameters of the adjacent lane can refer to the above steps S41 to S43.

[0116] Specifically, in step S5, a lane changing trajectory without obstacle collision and with the shortest path is selected from the plurality of lane changing trajectories as the local path of the vehicle.

[0117] Exemplarily, all generated lane change trajectories are subjected to collision detection with obstacles around the vehicle body. If there is no collision, the lane change trajectory is added to the group of candidate trajectories; if there is a collision, the lane change trajectory is filtered; the group of candidate trajectories is sorted according to the trajectory length; the candidate trajectory with the shortest length is taken as the final lane change trajectory, that is, the local path of the vehicle. Figures 3-4 , Figures 3-4 2 are two local path planning effect diagrams provided by the embodiments of the present invention. The solid line in the diagram represents the target trajectory, the white circle represents the obstacle, and the curve composed of black dots is the local path. Figure 3 It can be seen that the target trajectory is in contact with the obstacle corresponding to the horizontal coordinate 35. If the vehicle continues to drive along this target trajectory, it will inevitably collide with this obstacle, while the local path does not have any contact with this obstacle. Figure 4 It can be seen that the target trajectory is very close to the obstacle corresponding to the horizontal coordinate 20 in the figure. If the vehicle drives according to the target trajectory, it is very likely to collide with this obstacle, while the local path is far away from this obstacle and will not collide.

[0118] Further, the entire workflow of the local path planning method described in the embodiment of the present invention can be referred to Figure 5, based on the target trajectory of the vehicle, several reference trajectories are generated, obstacle collision detection is performed on these reference trajectories in turn, and candidate trajectories without obstacle collision are screened. When the number of candidate trajectories is greater than 1, the loss function value of each candidate trajectory is calculated to screen the optimal trajectory from multiple candidate trajectories, and it is judged whether the optimal trajectory has the lane change condition to change to the adjacent lane. If the lane change condition is not met, the optimal trajectory is used as the local path; if the lane change condition is met, several lane change trajectories are generated, and the lane change trajectory without obstacle collision and the shortest path is screened as the local path, and the vehicle is instructed to run according to this local path.

[0119] Compared with the prior art, the local path planning method disclosed in the present invention generates several reference trajectories based on the target trajectory of the vehicle, performs obstacle collision detection and loss function value comparison on these reference trajectories in turn, and then selects the optimal trajectory, introduces a lane changing mechanism, and when this optimal trajectory meets the lane changing conditions, generates several lane changing trajectories, which are more in line with the driving behavior of the vehicle on structured roads. Finally, the lane changing trajectory with no obstacle collision and the shortest path is selected from the several lane changing trajectories as the local path of the vehicle. By generating the lane changing trajectory, it is more in line with the driving behavior of the vehicle on structured roads, can meet the requirements of vehicle constraints and vehicle driving scenarios, and can provide the optimal local path according to the position of the vehicle, thereby improving the vehicle traffic efficiency.

[0120] See also Figure 6 , Figure 6 is a structural block diagram of a local path planning device 100 provided in an embodiment of the present invention, wherein the local path planning device 100 comprises:

[0121] A reference trajectory generation module 11 is used to generate a plurality of reference trajectories based on the target trajectory of the vehicle;

[0122] An obstacle collision detection module 12, configured to perform obstacle collision detection on each of the reference trajectories;

[0123] A candidate trajectory selection module 13, used to select a candidate trajectory without obstacle collision from a plurality of reference trajectories;

[0124] The optimal trajectory acquisition module 14 is used to calculate the loss function value of each candidate trajectory when the number of the candidate trajectories is greater than 1, and select the optimal trajectory from the candidate trajectories according to the loss function value; wherein the loss function value is used to evaluate the trajectory performance of the candidate trajectories;

[0125] A lane change trajectory generating module 15 is used to generate a plurality of lane change trajectories according to lane parameters of the adjacent lane when there is an adjacent lane at the position of the vehicle and the adjacent lane of the optimal trajectory allows lane change;

[0126] The local path generation module 16 is used to select a lane changing trajectory without obstacle collision and with the shortest path from the plurality of lane changing trajectories as the local path of the vehicle.

[0127] Specifically, the loss function value is calculated by a loss function, and the input parameters of the loss function include: the product of the trajectory evaluation value and its corresponding weight value, and the ratio of the minimum distance value between the candidate trajectory and the obstacle and its weight value.

[0128] Specifically, the trajectory evaluation value includes a lateral trajectory evaluation value, a longitudinal trajectory evaluation value and a trajectory consistency evaluation value; wherein, the lateral trajectory evaluation value is calculated according to the path planning time and distance of the candidate trajectory, and the weight value of the lateral trajectory evaluation value is a preset first parameter value; the longitudinal trajectory evaluation value is calculated according to the path planning time and speed difference parameter of the candidate trajectory, and the weight value of the longitudinal trajectory evaluation value is calculated according to the target speed and current speed of the vehicle; the trajectory consistency evaluation value is calculated according to the average curvature and curve offset value of the candidate trajectory and the local path calculated last time, and the weight value of the trajectory consistency evaluation value is a preset second parameter value; the weight value of the minimum distance value is calculated according to the number of drivable lanes of the candidate path and a fixed value.

[0129] Specifically, the reference trajectory generation module 11 is used to: construct a reference coordinate system with the target trajectory as a reference line, project the target trajectory onto the reference line to obtain a number of reference points; sample the reference points based on the vehicle's speed parameters, expected road width, and trajectory planning duration parameters to generate a number of reference trajectories.

[0130] Specifically, the lane change trajectory generation module 15 is used to: obtain the center line point of the adjacent lane; traverse each of the center line points, take the vehicle position as the starting point and the center line point as the end point, and calculate the interpolation coordinates from the starting point to the end point according to a preset rotation angle and number of rotations; respectively connect each of the starting point, the interpolation coordinates and the end point to generate a number of lane change trajectories.

[0131] Specifically, the lane-changing trajectory generating module 15 is further configured to generate a plurality of lane-changing trajectories according to lane parameters of adjacent lanes of the target trajectory when all reference trajectories have an obstacle collision.

[0132] Specifically, the local path generation module 16 is used to: when there is no adjacent lane at the location of the vehicle, or when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory does not allow lane change, use the optimal trajectory as the local path of the vehicle.

[0133] It is worth noting that the working process of each module in the local path planning device 100 described in the embodiment of the present invention can refer to the working process of the local path planning method described in the above embodiment, which will not be repeated here.

[0134] Compared with the prior art, the local path planning device 100 disclosed in the present invention generates a number of reference trajectories based on the target trajectory of the vehicle, performs obstacle collision detection and loss function value comparison on these reference trajectories in turn, and then selects the optimal trajectory, introduces a lane changing mechanism, and when this optimal trajectory meets the lane changing conditions, generates a number of lane changing trajectories that are more in line with the driving behavior of the vehicle on structured roads. Finally, a lane changing trajectory with no obstacle collision and the shortest path is selected from the several lane changing trajectories as the local path of the vehicle. By generating the lane changing trajectory, it is more in line with the driving behavior of the vehicle on structured roads, can meet the requirements of vehicle constraints and vehicle driving scenarios, and can provide the optimal local path according to the position of the vehicle, thereby improving the vehicle traffic efficiency.

[0135] See also Figure 7 , Figure 7 1 is a structural block diagram of a local path planning device 200 provided in an embodiment of the present invention, and the local path planning device 200 includes a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, the steps in the above-mentioned local path planning method embodiments are implemented, such as steps S1 to S5.

[0136] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the local path planning device 200.

[0137] The local path planning device 200 may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will appreciate that the schematic diagram is merely an example of the local path planning device 200 and does not constitute a limitation on the local path planning device 200. The local path planning device 200 may include more or fewer components than shown in the diagram, or may combine certain components, or different components. For example, the local path planning device 200 may also include input and output devices, network access devices, buses, and the like.

[0138] The processor 21 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 21 is the control center of the local path planning device 200, and uses various interfaces and lines to connect various parts of the entire local path planning device 200.

[0139] The memory 22 can be used to store the computer program and / or module. The processor 21 implements various functions of the local path planning device 200 by running or executing the computer program and / or module stored in the memory 22 and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0140] Wherein, if the module / unit integrated in the local path planning device 200 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 such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 21, the steps of the above-mentioned method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0141] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A local path planning method, characterized in that: include: Generate a number of reference trajectories based on the target trajectory of the vehicle; Performing obstacle collision detection on each of the reference trajectories to select a candidate trajectory without obstacle collision from a plurality of reference trajectories; When the number of the candidate trajectories is greater than 1, a loss function value of each candidate trajectory is calculated, and an optimal trajectory is selected from the candidate trajectories according to the loss function value; wherein the loss function value is used to evaluate the trajectory performance of the candidate trajectories; When there is an adjacent lane at the location of the vehicle, and the adjacent lane of the optimal trajectory allows lane change, generating a plurality of lane change trajectories according to lane parameters of the adjacent lane; A lane changing trajectory without obstacle collision and with the shortest path is selected from the plurality of lane changing trajectories as a local path of the vehicle.

2. The local path planning method according to claim 1, characterized in that: The loss function value is calculated by a loss function, and the input parameters of the loss function include: the product of the trajectory evaluation value and its corresponding weight value, and the ratio of the minimum distance value between the candidate trajectory and the obstacle and its weight value.

3. The local path planning method according to claim 2, characterized in that: The trajectory evaluation value includes a lateral trajectory evaluation value, a longitudinal trajectory evaluation value and a trajectory consistency evaluation value; wherein, The lateral trajectory evaluation value is calculated according to the path planning duration and distance of the candidate trajectory, and the weight value of the lateral trajectory evaluation value is a preset first parameter value; The longitudinal trajectory evaluation value is calculated according to the path planning duration and the vehicle speed difference parameter of the candidate trajectory, and the weight value of the longitudinal trajectory evaluation value is calculated according to the target speed and the current speed of the vehicle; The trajectory consistency evaluation value is calculated according to the candidate trajectory and the average curvature and the curve offset value of the local path calculated last time, and the weight value of the trajectory consistency evaluation value is a preset second parameter value; The weight value of the minimum distance value is calculated based on the number of drivable lanes of the candidate path and a fixed value.

4. The local path planning method according to claim 1, characterized in that: The generating of a plurality of reference trajectories based on the target trajectory of the vehicle includes: A reference coordinate system is constructed with the target trajectory as a reference line, and a plurality of reference points are obtained by projecting the target trajectory onto the reference line; The reference points are sampled based on a speed parameter of the vehicle, an expected road width, and a trajectory planning duration parameter to generate a plurality of reference trajectories.

5. The local path planning method according to claim 1, characterized in that: The generating a plurality of lane change trajectories according to the lane parameters of the adjacent lanes includes: Obtaining the centerline point of the adjacent lane; Traversing each of the center line points, taking the position of the vehicle as the starting point and the center line point as the end point, and calculating the interpolation coordinates from the starting point to the end point according to a preset rotation angle and number of rotations; Each of the starting point, the interpolation coordinates and the end point are connected respectively to generate a plurality of lane change trajectories.

6. The local path planning method according to claim 1, characterized in that: After performing obstacle collision detection on each of the reference trajectories, the method further includes: When all reference trajectories have obstacle collisions, a plurality of lane change trajectories are generated according to lane parameters of adjacent lanes of the target trajectory.

7. The local path planning method according to claim 1, characterized in that: The method further comprises: When there is no adjacent lane at the location of the vehicle, or when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory does not allow lane change, the optimal trajectory is used as the local path of the vehicle.

8. A local path planning device, characterized in that: include: A reference trajectory generation module, used to generate a plurality of reference trajectories based on the target trajectory of the vehicle; An obstacle collision detection module, used for performing obstacle collision detection on each of the reference trajectories; A candidate trajectory selection module, used to select a candidate trajectory without obstacle collision from a number of reference trajectories; an optimal trajectory acquisition module, used for calculating the loss function value of each candidate trajectory when the number of the candidate trajectories is greater than 1, and selecting the optimal trajectory from the candidate trajectories according to the loss function value; wherein the loss function value is used to evaluate the trajectory performance of the candidate trajectories; A lane change trajectory generating module, configured to generate a plurality of lane change trajectories according to lane parameters of the adjacent lane when there is an adjacent lane at the location of the vehicle and the adjacent lane of the optimal trajectory allows lane change; The local path generation module is used to select a lane changing trajectory without obstacle collision and with the shortest path from the plurality of lane changing trajectories as the local path of the vehicle.

9. A local path planning device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the local path planning method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the local path planning method according to any one of claims 1 to 7.

Images