A dense obstacle avoidance algorithm for UAVs based on safe flight corridor

基于安全飞行走廊的无人机密集障碍规避算法

¹ School of Automation, Northwestern Polytechnical University, Xi'an 710072, China
² School of Electronics and Information, Northwestern Polytechnical University, Xi'an 710072, China
³ Shenyang Aircraft Design Research Institute, Shenyang 110035, China
⁴ Xi'an Modern Control Technology Research Institute, Xi'an 710065, China

Received: 31 March 2022

Abstract

Aiming at the problem of autonomous obstacle avoidance of fixed-wing UAVs in a complex, dense and multi-obstacle environment, a path planning algorithm for fixed-wing UAVs based on a safe flight corridor is proposed. The difficulty of avoiding dense obstacles lies in the choice of obstacle circumvention and traversal: although circumvention is safer, the flight cost is greater; although the traversal cost is lower, the safety threat is higher. How to quickly solve the optimal path is the core issue. This paper firstly defines a safe flight corridor innovatively based on the maneuvering characteristics of fixed-wing UAVs and the Dubins curves. By comprehensively considering UAV flight safety and flight costs, an obstacle threat evaluation function is constructed. Secondly, in view of the computational complexity caused by the dense obstacles, an obstacle clustering algorithm based on obstacle density is proposed, and the nonlinear evaluation function in a high dynamic environment is quickly approximated by Monte Carlo sampling method. Finally, simulations verify the effectiveness of the proposed algorithm in solving dense obstacle avoidance for fixed-wing UAVs.

摘要

针对固定翼无人机在复杂密集多障碍物环境中的自主避障问题, 提出了一种基于安全飞行走廊的固定翼无人机路径规划算法。密集障碍规避的难点在于障碍绕行与穿行的选择: 绕行虽然更安全, 但飞行成本更大; 穿行虽然飞行成本更低, 但安全威胁较高, 如何快速求解最优路径是其中的核心问题。创新性地根据固定翼无人机的机动特性与Dubins曲线定义了安全飞行走廊, 综合考虑无人机飞行安全与飞行成本, 构建了障碍威胁评价函数; 针对障碍物密集造成的计算复杂问题, 提出了基于障碍物密度的障碍聚类算法, 并通过蒙特卡洛采样法实现了高动态环境下的非线性评价函数快速近似求解; 通过仿真验证了所提算法对于解决固定翼无人机密集障碍规避的有效性。