Reference speed generation method and system based on lookup table and storage medium

Abstract

The invention discloses a reference speed generation method and system based on a lookup table and a storage medium, the method comprises the following steps: S1, lookup table calibration: using speed-acceleration information of a vehicle in a manual driving mode to calibrate vehicle accelerations at different vehicle speeds, and using the vehicle accelerations as an acceleration lookup table AT; s2, speed-limiting smoothing: by means of forward traversal and posterior traversal, smoothing an original SV curve generated by combining the highest speeds of the vehicle under different lane curvatures by using the acceleration lookup table AT obtained in the step S1; and S3, forward simulation: based on the smoothed SV curve obtained in the step S2, using the acceleration lookup table AT obtained in the step S1, and obtaining a reference speed curve through forward simulation. The method provided by the invention not only ensures the optimality of the reference speed and solves the problem of time consumption, but also enhances the continuity / stability between frames.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile robots, and in particular to a reference speed generation method, system and storage medium based on a lookup table. Background Art

[0002] As an important function of the motion planning subsystem, the speed planning module of the autonomous driving system directly affects the execution quality of the downstream longitudinal control layer. Inappropriate target speed and acceleration will exceed the limit of vehicle motion and affect the safety and comfort experience of the driver and passengers. At present, the speed planning module is basically divided into two parts. The first part generates the reference speed, and the second part uses the reference speed as a guide for optimization.

[0003] In the document "Baidu Apollo EM Motion Planner", an optimal speed search method based on a mileage-time diagram (ST diagram) is provided. This method searches and generates the optimal speed in a two-dimensional space with the horizontal axis being the time axis and the vertical axis being the mileage axis.

[0004] This method mainly includes three parts: sampling, evaluation, and optimal speed selection:

[0005] (1) Sampling: The time axis and mileage axis are sampled uniformly or sparsely or densely through thresholds, and the two-dimensional space is divided into many small spaces. Each small space is a node. The node includes information such as mileage, speed, time, acceleration, jerk, cost, and previous node.

[0006] (2) Evaluation: All nodes obtained by sampling, starting from the first node, use the cost function to calculate the cost of each node, and update the minimum cost of each node while calculating until all the cost calculations of the last layer of nodes are completed. In this part, the cost of the node, the previous node and other information will be updated for subsequent selection. When updating the cost of each node, the nodes will be screened according to the acceleration range, and the cost calculation will not be performed on the nodes that do not meet the requirements. The cost calculation includes speed cost, acceleration cost, and jerk cost.

[0007] When updating a node, determine whether the next node is within the speed range or acceleration range based on the current node. If not, skip the node and do not calculate the cost; if so, update the cost.

[0008] (3) Optimal speed selection: Select the node with the lowest cost from the last layer of nodes, trace back the information of each node, and obtain an optimal speed passing through n nodes as the reference speed for speed optimization.

[0009] However, in the prior art, it is necessary to balance the optimality and time consumption. The higher the sampling resolution of the S-axis and the T-axis, the more accurate the reference speed, but the resulting time consumption increases exponentially. Summary of the invention

[0010] In view of the above technical problems, the technical solution of the present invention discloses a method for generating a reference speed for speed planning optimization through a lookup table. Under the premise of ensuring the optimality of the reference speed, the method of the present invention adopts a speed-acceleration lookup table to replace the search process, thereby greatly improving the time consumption problem and enhancing the continuity / stability between frames.

[0011] In order to achieve the above-mentioned purpose, the technical solution of the present invention provides a reference speed generation method based on a lookup table, which includes the following steps: S1 lookup table calibration: using the speed-acceleration information of the vehicle in the manual driving mode, calibrate the vehicle acceleration at different vehicle speeds as the acceleration lookup table AT; S2 speed limit smoothing: through forward traversal and post-item traversal, use the acceleration lookup table AT obtained in step S1 to smooth the original SV curve generated by the maximum speed combination of the vehicle under different lane curvatures; S3 forward simulation: based on the smoothed SV curve obtained in step S2, use the acceleration lookup table AT obtained in step S1 to obtain the reference speed curve through forward simulation.

[0012] Further, in step S2, the current speed limit point information s is obtained according to the original SV curve. curr 、v curr , calculate and update the next speed limit point information s according to the acceleration lookup table next 、v next and the previous speed limit point information prev 、v prev , and then perform speed limit smoothing on the original SV curve as follows:

[0013] S21 forward traversal: traverse all speed limit points from front to back, if v next >v curr , then get the acceleration a accel =AT(v curr ),renew

[0014] S22 backward traversal: traverse all speed limit points from back to front, if v prev >v curr , then update

[0015] Furthermore, in step S3, the current frame planning is divided into N steps, with a resolution of Δt between each step, and the k-th step state s of the vehicle is obtained. k、v k and the k+1th step state s k+1 、v k+1 , and then follow the steps below to obtain the reference speed curve:

[0016] S31: Get acceleration a k =AT(v k ), calculate v k+1 =v k +a k Δt, s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 ;

[0017] S32: According to s k+1 Interpolate to obtain the corresponding speed limit v on the smoothed SV curve max ;

[0018] S33: If v k+1 >v max , then update v k+1 =v max , s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 .

[0019] Furthermore, in step S2, the lane maximum speed curve is obtained from the map information, and the curvature speed curve is calculated according to the lane curvature k, wherein the speeds under different curvatures are calculated by the designed maximum lateral centripetal acceleration, and the original SV curve is generated by sorting them according to the size of S and taking the smaller V combination under the same S.

[0020] Further, in step S3, the k-th step state s is obtained according to the current state of the vehicle. k 、v k , and calculate the k+1th step state s according to the kth step state and the smoothed SV curve k+1 、v k+1 .

[0021] Further, in step S3, the range of the resolution Δt is 0.1s-0.5s.

[0022] The technical solution of the present invention also provides a reference speed generation system based on a lookup table, which includes the following modules: a lookup table calibration module, which uses the speed-acceleration information of the vehicle in the manual driving mode to calibrate the vehicle acceleration at different vehicle speeds as an acceleration lookup table AT; a speed limit smoothing module, which uses the acceleration lookup table AT obtained by the lookup table calibration module to smooth the original SV curve generated by the maximum speed combination of the vehicle under different lane curvatures through forward traversal and backward traversal; a forward simulation module, which uses the acceleration lookup table AT obtained by the lookup table calibration module to obtain a reference speed curve through forward simulation based on the smoothed SV curve obtained by the speed limit smoothing module.

[0023] The technical solution of the present invention also provides a computer-readable storage medium containing a computer program, which is used to execute the method as described above when the computer program is executed by one or more processors. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 It is a flowchart of a reference speed generation method based on a lookup table of the present invention;

[0026] Figure 2 Schematic diagram of the lane speed limit curve of the present invention.

[0027] Figure 3 It is a schematic diagram of generating the reference speed curve of the present invention. DETAILED DESCRIPTION

[0028] 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.

[0029] In an embodiment of the invention, a reference speed generation method based on a lookup table is provided, which includes the following steps: S1 lookup table calibration: using the speed-acceleration information of the vehicle in the manual driving mode, calibrate the vehicle acceleration at different vehicle speeds as the acceleration lookup table AT; S2 speed limit smoothing: through forward traversal and post-traversal, use the acceleration lookup table AT obtained in step S1 to smooth the original SV curve generated by the maximum speed combination of the vehicle under different lane curvatures; S3 forward simulation: based on the smoothed SV curve obtained in step S2, use the acceleration lookup table AT obtained in step S1 to obtain the reference speed curve through forward simulation.

[0030] Further, in step S2, the current speed limit point information s is obtained according to the original SV curve. curr 、v curr , calculate and update the next speed limit point information s according to the acceleration lookup table next 、v next and the previous speed limit point information prev 、v prev , and then perform speed limit smoothing on the original SV curve as follows:

[0031] S21 forward traversal: traverse all speed limit points from front to back, if v next >v curr , then get the acceleration a accel =AT(v curr ),renew

[0032] S22 backward traversal: traverse all speed limit points from back to front, if v prev >v curr , then update

[0033] Furthermore, in step S3, the current frame planning is divided into N steps, with a resolution of Δt between each step, and the k-th step state s of the vehicle is obtained. k 、v k and the k+1th step state s k+1 、v k+1 , and then follow the steps below to obtain the reference speed curve:

[0034] S31: Get acceleration a k =AT(v k ), calculate v k+1 =v k +a k Δt, s k+1 =s k +v k ·Δt+0.5·ak ·Δt 2 ;

[0035] S32: According to s k+1 Interpolate to obtain the corresponding speed limit v on the smoothed SV curve max ;

[0036] S33: If v k+1 >v max , then update v k+1 =v max , s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 .

[0037] Furthermore, in step S2, the lane maximum speed curve is obtained from the map information, and the curvature speed curve is calculated according to the lane curvature k, wherein the speeds under different curvatures are calculated by the designed maximum lateral centripetal acceleration, and the original SV curve is generated by sorting them according to the size of S and taking the smaller V combination under the same S.

[0038] Further, in step S3, the k-th step state s is obtained according to the current state of the vehicle. k 、v k , and calculate the k+1th step state s according to the kth step state and the smoothed SV curve k+1 、v k+1 .

[0039] Further, in step S3, the range of the resolution Δt is 0.1s-0.5s.

[0040] In another embodiment of the present invention, a reference speed generation system based on a lookup table is also provided, which includes the following modules: a lookup table calibration module, which uses the speed-acceleration information of the vehicle in the manual driving mode to calibrate the vehicle acceleration at different vehicle speeds as an acceleration lookup table AT; a speed limit smoothing module, which uses the acceleration lookup table AT obtained by the lookup table calibration module to smooth the original SV curve generated by the maximum speed combination of the vehicle under different lane curvatures through forward traversal and backward traversal; a forward simulation module, which uses the acceleration lookup table AT obtained by the lookup table calibration module to obtain a reference speed curve through forward simulation based on the smoothed SV curve obtained by the speed limit smoothing module.

[0041] In other embodiments of the present invention, a computer-readable storage medium containing a computer program is provided. When the computer program is executed by one or more processors, the computer program is used to perform the method as described above. Specific embodiments

[0043] See also Figure 1 The method of the present invention mainly includes three parts: velocity-acceleration lookup table calibration, speed limit calculation and smoothing, and forward simulation:

[0044] (1) Velocity-acceleration lookup table

[0045] Using the speed-acceleration information of the vehicle in manual driving mode, the acceleration of the vehicle at different speeds is calibrated through a large amount of data and used as an acceleration lookup table (an accel table, referred to as AT); a constant is used as a deceleration (a decel).

[0046] (2) Speed ​​limit calculation and smoothing

[0047] a. Speed ​​limit calculation:

[0048] According to different s (mileage) positions (s>=0 and increasing, s interval is between 0.5 and 1.0m), get the lane speed from the map (such as Figure 2 The blue curve in Figure 2) and the curvature speed calculated based on the lane curvature k (as Figure 2 The speed under different curvatures is determined by the designed maximum lateral centripetal acceleration (the turning radius of different vehicles is different, which is set to 1.0-1.5 in this embodiment. ) is calculated and an SV (mileage-speed) curve is generated by sorting by s and taking the smaller v combination under the same s.

[0049] b. Speed ​​limit smoothing

[0050] According to the original SV curve, the current speed limit point information s can be obtained curr 、v curr , update the next speed limit point information s according to the accelerometer calculation next 、v next and the previous speed limit point information prev 、v prev .

[0051] i. Forward traversal: traverse all speed limit points from front to back, if v next >v curr , then get the acceleration a accel =AT(v curr ),renew

[0052] ii. Backward traversal: traverse all speed limit points from back to front. If v prev >v curr , then update

[0053] (3) Forward simulation to obtain reference speed (such as Figure 3 The blue curve in

[0054] The current frame planning is divided into N steps, and the resolution between each step is Δt (for example, 0.2s can also be other values ​​between 0.1-0.5s); the kth (=0) step state is obtained according to the current state of the vehicle, and the kth step state s k 、v k , and according to the k-th step state and the smoothed SV curve (such as Figure 3 The red curve in the figure calculates the state s at step k+1 k+1 、v k+1 .

[0055] a. Get acceleration a k =AT(v k ), calculate v k+1 =v k +a k ·Δt,s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 ;

[0056] b. According to s k+1 Interpolate to obtain the corresponding speed limit v on the smooth SV curve max ;

[0057] c. If v k+1 >v max , then update v k+1 =v max , s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 .

[0058] The reference speed generation method of the present invention not only ensures the optimality of the reference speed and improves the time consumption problem, but also enhances the continuity / stability between frames.

[0059] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it. They cannot be used to limit the protection scope of the present invention. All equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A reference speed generation method based on a lookup table, characterized in that: The steps include: S1 Lookup table calibration: Use the speed-acceleration information of the vehicle in the manual driving mode to calibrate the vehicle acceleration at different vehicle speeds as the acceleration lookup table AT; S2 speed limit smoothing: by forward traversal and backward traversal, the acceleration lookup table AT obtained in step S1 is used to smooth the original SV curve generated by the maximum speed combination of the vehicle under different lane curvatures; S3 forward simulation: Based on the smoothed SV curve obtained in step S2, using the acceleration lookup table AT obtained in step S1, a reference speed curve is obtained through forward simulation.

2. The method according to claim 1, characterized in that: In step S2, according to the original SV curve, the current speed limit point information scurr, vcurr is obtained, and the next speed limit point information snext, vnext and the previous speed limit point information sprev, vprev are calculated and updated according to the acceleration lookup table, and then the original SV curve is speed-limited and smoothed according to the following steps: S21 forward traversal: traverse all speed limit points from front to back. If vnext>vcurr, get the acceleration aaccel=AT(vcurr) and update S22 backward traversal: traverse all speed limit points from back to front, if vprev>vcurr, update 3. The method according to claim 2, characterized in that In step S3, the current frame planning is divided into N steps, with a resolution of Δt between each step, and the k-th step state sk, vk and the k+1-th step state sk of the vehicle are obtained. +1 、vk +1 , and then follow the steps below to obtain the reference speed curve: S31: Get acceleration a k =AT(v k ), calculate v k+1 =v k +a k Δt, s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 ; S32: According to s k+1 Interpolate to obtain the corresponding speed limit v on the smoothed SV curve max ; S33: If v k+1 >v max , then update s k+1 =s k +v k ·Δt+0.5·a k ·Δt 2 .

4. The method according to claim 3, characterized in that: In step S2, the lane maximum speed curve is obtained from the map information, and the curvature speed curve is calculated according to the lane curvature k, wherein the speeds under different curvatures are calculated by the designed maximum lateral centripetal acceleration, and the original SV curve is generated by sorting them according to the size of S and taking the smaller V combination under the same S.

5. The method according to claim 4, characterized in that In step S3, the k-th step state s is obtained according to the current state of the vehicle k 、v k , and calculate the k+1th step state s according to the kth step state and the smoothed SV curve k+1 、v k+1 .

6. The method according to claim 5, characterized in that In step S3, the range of the resolution Δt is 0.1s-0.5s.

7. A reference speed generation system based on a lookup table, characterized in that: Includes the following modules: A lookup table calibration module, used to calibrate the vehicle acceleration at different vehicle speeds using the speed-acceleration information of the vehicle in the manual driving mode as an acceleration lookup table AT; The speed limit smoothing module is used to smooth the original SV curve generated by the combination of the vehicle's maximum speed under different lane curvatures by using the acceleration lookup table AT obtained by the lookup table calibration module through forward traversal and backward traversal; The forward simulation module is used to obtain a reference speed curve through forward simulation based on the smoothed SV curve obtained by the speed limit smoothing module and using the acceleration lookup table AT obtained by the lookup table calibration module.

8. A computer-readable storage medium containing a computer program, characterized in that: When the computer program is executed by one or more processors, it is used to perform the method according to any one of claims 1 to 6.

Images