Motion planning of quadrotors in cluttered environments using efficient mapping and spatial-temporal optimization

Engineering notes

Compared to benchmark methods, the proposed method demonstrates superior performance in terms of average speed, trajectory length and smoothness in unknown cluttered environments with dense obstacles. The proposed method improves trajectory smoothness by 69.17%, reduces trajectory length by 2.02%, increases flight speed by 2.25% and decreases flight time by 4.39% compared to the best metrics of benchmark methods.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

The purpose of this study is to develop a method enabling quadrotors to achieve autonomous navigation in complex and obstacle-dense environments, without the need for pre-built maps. First, the authors propose an efficient local mapping method that implements local map sliding without boundary restricted. Then, they generate a safe flight corridor within the local map by applying convex decomposition to a dynamic feasible initial trajectory. Finally, a spatial-temporal trajectory optimization method is proposed, which takes the dynamically feasible trajectory as the initial guess, to ensure safety, smoothness, agility and dynamic feasibility of the quadrotor. Compared to benchmark methods, the proposed method demonstrates superior performance in terms of average speed, trajectory length and smoothness in unknown cluttered environments with dense obstacles. The proposed method improves trajectory smoothness by 69.17%, reduces trajectory length by 2.02%, increases flight speed by 2.25% and decreases flight time by 4.39% compared to the best metrics of benchmark methods. The main contribution of this paper is to propose a motion planner enabling quadrotors to achieve autonomous navigation in complex and obstacle-dense environments. Specifically, the proposed method leverages a one-dimensional array with fixed memory size to facilitate large-scale autonomous flight without boundary constraints. Furthermore, by employing spatiotemporal trajectory optimization, the method effectively addresses challenges such as reduced trajectory quality and dynamic infeasibility caused by constrained free space in highly cluttered environments. Extensive simulation and real-world experiments verify the effectiveness and robustness of this method.

Links and sources

• Official / arXiv page

Comments