Fast path planning method, device and storage medium adapting to unknown dynamic space

Abstract

A fast path planning method adapting to unknown dynamic space belongs to the field of automatic driving. The steps of the method include: on the basis of the grid map, considering the size constraints of the planned vehicle, an extended map is established; based on the extended map, the convex vertices of the impassable area are obtained; the convex vertices are used as nodes, and the path planning algorithm is adaptively selected to obtain Take the rough path; according to the characteristics of the local environment, consider the trajectory following control ability, correct the rough path, and quickly form a smooth and optimal local path, which can be realized by the vehicle forming control commands. The present invention can greatly reduce the number of search space points, thereby realizing rapid planning; it can adaptively select a planning algorithm suitable for the characteristics of the map, ensuring the efficiency of path planning and the optimality of the path; the problem of trajectory realization is considered in path planning, which is convenient for loading It implements path following control.

Description

technical field [0001] The invention relates to automatic driving technology, in particular to a fast path planning method, device and storage medium adapting to unknown dynamic space. Background technique [0002] Autonomous driving technology has been developed for decades, and the development of path planning technology for urban road environments is relatively mature. Under the premise of different road conditions, different planning algorithms can be selected to match and adapt to the environmental characteristics. In the face of an unknown dynamic environment, the map is extremely insufficient, and the size and complexity of the map area cannot be judged in advance, making a single path planning algorithm incompetent. In addition, in the face of an unknown environment, the accuracy of the navigation algorithm and the error of the sensor will introduce certain errors into the pose estimation of the vehicle. At the same time, if the trajectory following control and pat...

AI Technical Summary

Problems solved by technology

In the face of an unknown dynamic environment, the map is extremely insufficient, and the size and complexity of the map area cannot be judged in advance, making it difficult for a single path planning algorithm

In addition, in the face of an unknown environment, the accuracy of the navigation algorithm and the error of the sensor will introduce certain errors into the pose estimation of the vehicle.

At the same time, if the trajectory following control and path planning are separated, under the influence of high uncertainty in the environment, the performance constraints of the vehicle may cause the consequences that the path cannot be realized

Therefore, if the influence of trajectory following control is considered in path planning, the uncertainty of vehicle positioning and attitude determination and environmental uncertainty will further increase the difficulty of path planning

Images