In this paper we address the problem of path planning in an unknown environment with an aerial robot. The main goal is to safely follow the planned trajectory by avoiding obstacles. The proposed approach is suitable for aerial vehicles equipped with 3D sensors, such as LiDARs. It performs obstacle avoidance in real time and on an on-board computer. We present a novel algorithm based on the conventional Artifcial Potential Field (APF) that corrects the planned trajectory to avoid obstacles. To this end, our modifed algorithm uses a rotation-based component to avoid local minima. The smooth trajectory following, achieved with the MPC tracker, allows us to quickly change and re-plan the UAV trajectory. Comparative experiments in simulation have shown that our approach solves local minima problems in trajectory planning and generates more effcient paths to avoid potential collisions with static obstacles compared to the original APF method.

翻译：本文研究了空中机器人在未知环境中的路径规划问题，主要目标是通过避障安全地跟踪规划轨迹。该方案适用于配备激光雷达等三维传感器的飞行器，能在机载计算机上实时执行避障功能。我们提出了一种基于传统人工势场法（APF）的新算法，通过修正规划轨迹来规避障碍物。为此，改进算法采用旋转分量避免局部极小值问题。结合MPC跟踪器实现的平滑轨迹跟踪，使得无人机能够快速调整并重新规划轨迹。仿真对比实验表明，与原始APF方法相比，该方法解决了轨迹规划中的局部极小值问题，生成了更高效的路径以避免与静态障碍物发生潜在碰撞。