Smooth transition method of unmanned driving planning path and terminal
By inserting the intermediate point in the unmanned driving planning path and using Bezier curve smoothing, combining obstacle expansion layer detection and gradient descent processing, the path is optimized, and the problems of path dissolution and insufficient obstacle avoidance are solved, achieving safer and smoother path planning.
Patent Information
- Application Number
- CN202510630089.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-16
- Publication Date
- 2025-08-12
AI Technical Summary
There are problems of sharp turn between straight lines in the existing unmanned driving planning paths, resulting in poor path planning and insufficient obstacle avoidance capabilities, and environmental interference has a great impact on the path planning results.
By inserting the intermediate point in each segment of the path, smoothing with the Bezier curve, and collision detection with the expansion layer of the obstacle, combining gradient descent and congenital value calculations, the low-generation value path points are deleted, and the path is optimized.
It improves the safety, fluency and feasibility of the path, and improves the autonomous path planning capabilities of driverless vehicles in different environments.
Smart Images
Figure CN120467374A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of unmanned driving path planning, and in particular to a smooth transition method and terminal for an unmanned driving planned path. Background Art
[0002] Using various path planning algorithms, such as Dijkstra's, A*, D-star, and RRT, to determine a global path based on obstacles, the paths they generate often suffer from sharp turns (large changes in curvature) between straight lines. This can lead to issues such as inconsistent path planning, insufficient obstacle avoidance, and the impact of environmental interference on path planning results. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to provide a smooth transition method and terminal for unmanned driving planning path, which can select a relatively optimal path to improve the smoothness of the path and enhance the ability of unmanned driving to avoid obstacles.
[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is: A smooth transition method for an unmanned driving planned path, comprising the steps of: S1. Obtain a generated unmanned driving planned path, insert a preset number of intermediate points between the two endpoints of each line segment in the unmanned driving planned path, and obtain a first smooth path by connecting the intermediate points; S2. Selecting a preset number of control points for each path segment in the first smooth path, and applying the control points to a Bezier curve to obtain a second smooth path; S3, establishing an expansion layer for obstacles within the path planning area, performing collision detection on the second smooth path, and if it is detected that the second smooth path intersects with the expansion layer, returning to step S2 to reselect a preset number of control points and regenerate the second smooth path; S4. Performing a gradient descent process on the second smooth path to obtain a third smooth path; S5. Calculate cost values for the path points in the third smooth path and delete the path points whose cost values are lower than a threshold to obtain a final smooth path.
[0005] In order to solve the above technical problems, another technical solution adopted by the present invention is: A terminal for smooth transition of an unmanned driving planned path includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, each step of the above-mentioned method for smooth transition of an unmanned driving planned path is implemented.
[0006] The beneficial effects of the present invention are as follows: for an already generated unmanned driving planned path, an intermediate point is first inserted into each line segment of the path, thereby replacing long line segments with short line segments to obtain a first smooth path; multiple control points are then selected from the first smooth path and smoothed using a Bezier curve to obtain a second smooth path; the second smooth path is then adaptively updated by performing collision detection with the expansion layer of the obstacle; the second smooth path is then smoothed using gradient descent, and finally path points with lower cost values are deleted to obtain the final smooth path. The smooth path obtained in this way has improved path safety, fluency, and feasibility, and can better implement autonomous path planning for unmanned vehicles in different environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Figure 1 This is a flow chart of a smooth transition method for an unmanned driving planning path according to an embodiment of the present invention; Figure 2 A schematic diagram of a smooth transition terminal for planning an unmanned driving path according to an embodiment of the present invention; Description of labels: 1. A smooth transition terminal for unmanned driving path planning; 2. Memory; 3. Processor. DETAILED DESCRIPTION
[0008] To illustrate the technical content, achieved objectives and effects of the present invention in detail, the following description is given in conjunction with the embodiments and accompanying drawings.
[0009] Please refer to Figure 1 , an embodiment of the present invention provides a method for smooth transition of an unmanned driving planned path, comprising the steps of: S1. Obtain a generated unmanned driving planned path, insert a preset number of intermediate points between the two endpoints of each line segment in the unmanned driving planned path, and obtain a first smooth path by connecting the intermediate points; S2. Selecting a preset number of control points for each path segment in the first smooth path, and applying the control points to a Bezier curve to obtain a second smooth path; S3, establishing an expansion layer for obstacles within the path planning area, performing collision detection on the second smooth path, and if it is detected that the second smooth path intersects with the expansion layer, returning to step S2 to reselect a preset number of control points and regenerate the second smooth path; S4. Performing a gradient descent process on the second smooth path to obtain a third smooth path; S5. Calculate cost values for the path points in the third smooth path and delete the path points whose cost values are lower than a threshold to obtain a final smooth path.
[0010] From the above description, it can be seen that the beneficial effects of the present invention are as follows: for the generated unmanned driving planning path, first, an intermediate point is inserted into each line segment of the path, thereby replacing the long line segment with a short line segment to obtain a first smooth path; then, multiple control points are selected from the first smooth path, and smoothed by a Bezier curve to obtain a second smooth path, and the second smooth path is subjected to collision detection with the expansion layer of the obstacle, in this way, the second smooth path is adaptively updated; then, the second smooth path is smoothed by gradient descent, and finally, the path points with lower cost values in the path are deleted to obtain the final smooth path. The smooth path obtained in this way has improved path safety, fluency and feasibility, and can better realize autonomous path planning of unmanned vehicles in different environments.
[0011] Furthermore, step S1 includes: Obtaining a generated unmanned driving planned path, where the unmanned driving planned path is composed of line segments; A preset number of intermediate points are inserted between the two endpoints of each line segment in the unmanned driving planned path using linear interpolation, and all the intermediate points are sequentially connected using straight lines to obtain a first smooth path.
[0012] As can be seen from the above description, by replacing the continuous straight line segments in the path with shorter line segments, the generated path can be made smoother.
[0013] Furthermore, step S2 includes: In the first smooth path, continuous straight line segments are used as a path segment, and a preset number of control points are selected in each path segment, where the distances between adjacent control points are the same; The control points in each path segment are input into a Bezier curve algorithm to obtain a smooth path segment, and all the smooth path segments are connected to obtain a second smooth path.
[0014] From the above description, it can be seen that for each path segment of the first smooth path, some control points can be selected and a smooth path can be calculated using a Bezier curve algorithm. In this way, the path can be adjusted to a curve shape, thereby improving the smoothness of the automatically planned path.
[0015] Furthermore, step S4 includes: defining a cost function, the cost function being used to calculate the distance between a path point of the second smooth path and an obstacle; By calculating the gradient of the cost function, moving the path points of the second smooth path whose distance to the obstacle is less than the first distance in the opposite direction of the gradient; Interpolation is performed on the path points whose distances to the obstacle are greater than the first distance and less than the second distance using an interpolation method.
[0016] As can be seen from the above description, by calculating the gradient of the cost function, we can determine how to adjust the waypoints on the path at a given location to minimize the cost function. By updating the position of the waypoints in a direction away from obstacles, the cost function is gradually reduced, thereby improving the feasibility of automatically planned paths.
[0017] Furthermore, step S5 includes: For each path point in the third smooth path, calculating a cost value of the path point according to the path length and curvature of the path point; If the cost value of the path point is lower than a threshold, the previous path point and the next path point of the path point are connected, and the path point is deleted; After traversing and processing all path points, the final smooth path is obtained.
[0018] As can be seen from the above description, by deleting path points with lower cost values, unnecessary path points can be removed, thereby reducing the complexity of the path and further improving the smoothness of the path.
[0019] Please refer to Figure 2 Another embodiment of the present invention provides a smooth transition terminal for an unmanned driving planned path, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the various steps of the above-mentioned smooth transition method for an unmanned driving planned path are implemented.
[0020] The above-mentioned method and terminal for smooth transition of unmanned driving planned paths are suitable for selecting a relatively optimal path when planning a path for unmanned driving, thereby improving the smoothness of the path and the ability of the unmanned driving to avoid obstacles. The following is an explanation through specific implementation methods: Example 1 Please refer to Figure 1 , a smooth transition method for unmanned driving planning path, comprising the steps of: S1. Obtain a generated unmanned driving planned path, insert a preset number of intermediate points between two endpoints of each line segment in the unmanned driving planned path, and obtain a first smooth path by connecting the intermediate points.
[0021] Step S1 is specifically as follows: Obtaining a generated unmanned driving planned path, where the unmanned driving planned path is composed of line segments; A preset number of intermediate points are inserted between the two endpoints of each line segment in the unmanned driving planned path using linear interpolation, and all the intermediate points are sequentially connected using straight lines to obtain a first smooth path.
[0022] Specifically, a grid-based approach is first used to model the environment. By introducing path smoothness and distance heuristics into the heuristic function, as well as obstacle heuristics into the path transition probability, the path optimization and obstacle avoidance capabilities are improved. The extracted optimal path key nodes are then connected to form the planned path for the autonomous driving system.
[0023] Smooth the generated path. This is done by replacing consecutive straight line segments with shorter ones. Specifically, for each segment, interpolate some intermediate points between its two endpoints. These points can be calculated using an interpolation method (such as linear interpolation) to make the generated path smoother.
[0024] S2. Select a preset number of control points for each path segment in the first smooth path, and apply the control points to a Bezier curve to obtain a second smooth path.
[0025] Step S2 is specifically as follows: In the first smooth path, continuous straight line segments are used as a path segment, and a preset number of control points are selected in each path segment, where the distances between adjacent control points are the same; The control points in each path segment are input into a Bezier curve algorithm to obtain a smooth path segment, and all the smooth path segments are connected to obtain a second smooth path.
[0026] Specifically, a Bezier curve is used to further smooth the path. A Bezier curve is a mathematical curve whose shape is defined by control points. By selecting some control points and applying the Bezier curve to the path, the path curve can be further smoothed. Specifically, for each path segment, a number of control points are selected and a Bezier curve algorithm is used to calculate a smoothed path.
[0027] S3. Establish an expansion layer for obstacles in the path planning area, perform collision detection on the second smooth path, and if it is detected that the second smooth path intersects with the expansion layer, return to step S2 to reselect a preset number of control points and regenerate the second smooth path.
[0028] Specifically, after the Bezier curve is smoothed, the second smoothed path may collide with obstacles. To solve this problem, an expansion layer can be added to the obstacle. This means adding an additional area around the obstacle so that the path can bypass this area, thus avoiding collision with the obstacle.
[0029] The size of the expansion layer can be determined based on the specific situation. A common approach is to expand the boundary points of the obstacle outward by a certain distance. Different shapes can be used to define the expansion boundary, such as a circle or polygon to describe the shape of the expansion layer.
[0030] Once the obstacle has been expanded, a collision detection algorithm can be used to determine if the path intersects the obstacle. If the path intersects the expansion layer, the path needs to be adjusted to avoid the collision. This can be achieved by replanning the path or adjusting the control points in the path.
[0031] S4. Perform gradient descent processing on the second smooth path to obtain a third smooth path.
[0032] Step S4 is specifically as follows: A cost function is defined, where the cost function is used to calculate the distance between the path point of the second smooth path and the obstacle; the smaller the distance, the greater the cost, indicating that the path is closer to the obstacle.
[0033] By calculating the gradient of the cost function, moving the path points of the second smooth path whose distance to the obstacle is less than the first distance in the opposite direction of the gradient; Interpolation is performed on the path points whose distances to the obstacle are greater than the first distance and less than the second distance using an interpolation method.
[0034] Specifically, using the gradient descent method, vertices are moved and inserted to make them farther away from obstacles, and the path points near obstacles are made denser.
[0035] By calculating the gradient of the cost function, we can determine how to adjust the path points on the path to minimize the cost function at a given location. We update the path points in the opposite direction of the gradient, gradually reducing the cost function. This process is repeated until the specified termination condition is met (e.g., reaching the maximum number of iterations or the cost function converges).
[0036] S5. Calculate cost values for the path points in the third smooth path and delete the path points whose cost values are lower than a threshold to obtain a final smooth path.
[0037] Step S5 is specifically as follows: For each path point in the third smooth path, calculating a cost value of the path point according to the path length and curvature of the path point; If the cost value of the path point is lower than a threshold, the previous path point and the next path point of the path point are connected, and the path point is deleted; After traversing and processing all path points, the final smooth path is obtained.
[0038] In this embodiment, a cost-based shortcutting method is used to remove unnecessary path points. A cost value is calculated for each path point. This cost value can be calculated based on evaluation metrics at the path point (e.g., path length, curvature, etc.). The distance between the path point and obstacles should be considered when calculating the cost value. The smaller the cost value, the lower the cost of the path passing through that point.
[0039] Use a heuristic algorithm (such as a greedy algorithm) to traverse all path points on the path, starting from the starting point, and check the cost of each path point one by one. If the cost of a path point is small, it can be considered unnecessary and can be directly deleted from the path, and the previous and next points can be connected.
[0040] This process is repeated until the entire path is traversed or a predetermined termination condition is reached. The resulting simplified path removes unnecessary path points, thereby reducing the complexity of the path.
[0041] Example 2 Please refer to Figure 2 A smooth transition terminal 1 for an unmanned driving planned path includes a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3. When the processor 3 executes the computer program, each step of a smooth transition method for an unmanned driving planned path in embodiment 1 is implemented.
[0042] In summary, the present invention provides a smooth transition method and terminal for an unmanned driving planned path. For an already generated unmanned driving planned path, an intermediate point is first inserted into each line segment of the path, thereby replacing long line segments with short line segments to obtain a first smooth path; multiple control points are then selected from the first smooth path, and smoothed using a Bezier curve to obtain a second smooth path. The second smooth path is subjected to collision detection with the expansion layer of the obstacle, and the second smooth path is adaptively updated in this way; the second smooth path is then smoothed using gradient descent, and finally path points with lower cost values are deleted in the path to obtain the final smooth path. The smooth path obtained in this way has improved path safety, fluency, and feasibility, and can better realize autonomous path planning for unmanned vehicles in different environments.
[0043] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's description and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A smooth transition method for unmanned driving planning path, characterized in that: Including steps: S1. Obtain a generated unmanned driving planned path, insert a preset number of intermediate points between the two endpoints of each line segment in the unmanned driving planned path, and obtain a first smooth path by connecting the intermediate points; S2. Selecting a preset number of control points for each path segment in the first smooth path, and applying the control points to a Bezier curve to obtain a second smooth path; S3, establishing an expansion layer for obstacles within the path planning area, performing collision detection on the second smooth path, and if it is detected that the second smooth path intersects with the expansion layer, returning to step S2 to reselect a preset number of control points and regenerate the second smooth path; S4. Performing a gradient descent process on the second smooth path to obtain a third smooth path; S5. Calculate cost values for the path points in the third smooth path and delete the path points whose cost values are lower than a threshold to obtain a final smooth path.
2. The method for smooth transition of an unmanned driving planned path according to claim 1, characterized in that: Step S1 includes: Obtaining a generated unmanned driving planned path, where the unmanned driving planned path is composed of line segments; A preset number of intermediate points are inserted between the two endpoints of each line segment in the unmanned driving planned path using linear interpolation, and all the intermediate points are sequentially connected using straight lines to obtain a first smooth path.
3. The method for smooth transition of an unmanned driving planning path according to claim 1, characterized in that: Step S2 includes: In the first smooth path, continuous straight line segments are used as a path segment, and a preset number of control points are selected in each path segment, where the distances between adjacent control points are the same; The control points in each path segment are input into a Bezier curve algorithm to obtain a smooth path segment, and all the smooth path segments are connected to obtain a second smooth path.
4. The method for smooth transition of an unmanned driving planned path according to claim 1, characterized in that: Step S4 includes: defining a cost function, the cost function being used to calculate the distance between a path point of the second smooth path and an obstacle; By calculating the gradient of the cost function, moving the path points of the second smooth path whose distance to the obstacle is less than the first distance in the opposite direction of the gradient; Interpolation is performed on the path points whose distances to the obstacle are greater than the first distance and less than the second distance using an interpolation method.
5. The method for smooth transition of an unmanned driving planning path according to claim 1, characterized in that: Step S5 includes: For each path point in the third smooth path, calculating a cost value of the path point according to the path length and curvature of the path point; If the cost value of the path point is lower than a threshold, the previous path point and the next path point of the path point are connected, and the path point is deleted; After traversing and processing all path points, the final smooth path is obtained.
6. A smooth transition terminal for unmanned driving planning path, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the following steps are implemented: S1. Obtain a generated unmanned driving planned path, insert a preset number of intermediate points between the two endpoints of each line segment in the unmanned driving planned path, and obtain a first smooth path by connecting the intermediate points; S2. Selecting a preset number of control points for each path segment in the first smooth path, and applying the control points to a Bezier curve to obtain a second smooth path; S3, establishing an expansion layer for obstacles within the path planning area, performing collision detection on the second smooth path, and if it is detected that the second smooth path intersects with the expansion layer, returning to step S2 to reselect a preset number of control points and regenerate the second smooth path; S4. Performing a gradient descent process on the second smooth path to obtain a third smooth path; S5. Calculate cost values for the path points in the third smooth path and delete the path points whose cost values are lower than a threshold to obtain a final smooth path.
7. The smooth transition terminal for unmanned driving planning path according to claim 6, characterized in that: Step S1 includes: Obtaining a generated unmanned driving planned path, where the unmanned driving planned path is composed of line segments; A preset number of intermediate points are inserted between the two endpoints of each line segment in the unmanned driving planned path using linear interpolation, and all the intermediate points are sequentially connected using straight lines to obtain a first smooth path.
8. The smooth transition terminal for unmanned driving planning path according to claim 6, characterized in that: Step S2 includes: In the first smooth path, continuous straight line segments are used as a path segment, and a preset number of control points are selected in each path segment, where the distances between adjacent control points are the same; The control points in each path segment are input into a Bezier curve algorithm to obtain a smooth path segment, and all the smooth path segments are connected to obtain a second smooth path.
9. The smooth transition terminal for unmanned driving planning path according to claim 6, characterized in that: Step S4 includes: defining a cost function, the cost function being used to calculate the distance between a path point of the second smooth path and an obstacle; By calculating the gradient of the cost function, moving the path points of the second smooth path whose distance to the obstacle is less than the first distance in the opposite direction of the gradient; Interpolation is performed on the path points whose distances to the obstacle are greater than the first distance and less than the second distance using an interpolation method.
10. The smooth transition terminal for unmanned driving planning path according to claim 6, characterized in that: Step S5 includes: For each path point in the third smooth path, calculating a cost value of the path point according to the path length and curvature of the path point; If the cost value of the path point is lower than a threshold, the previous path point and the next path point of the path point are connected, and the path point is deleted; After traversing and processing all path points, the final smooth path is obtained.