基于改进向量场直方图算法的无人机动态避障策略  

Dynamic Obstacle Avoidance of UAV Based on Improved Vector Field Histogram Algorithm

作　　者：符小卫[1] 吴迪 支辰元 Fu Xiaowei;Wu Di;Zhi Chenyuan(Northwestern Polytechnical University,Xi’an 710072,China)

机构地区：[1]西北工业大学,陕西西安710072

出　　处：《航空科学技术》2023年第9期100-109,共10页Aeronautical Science & Technology

基　　金：航空科学基金(2020Z023053001)。

摘　　要：目前传统向量场直方图(VFH)算法存在易陷入局部陷阱的缺陷,本文提出了基于陷阱检测机制与动态阈值更新策略的改进VFH算法,更加符合局部未知环境下无人机路径规划的要求,并针对复杂未知场景中无人机避障问题,提出了基于A^(*)算法和改进VFH算法的避障算法。首先,无人机根据全局已知障碍物信息,基于A^(*)算法构建目标航路点;其次,在目标航路点不可达的情况下,无人机根据运动状态与激光雷达探测到的地形信息,基于改进向量场直方图算法进行局部规划。在局部规划中,针对传统VFH算法存在的缺陷进行了改进:针对传统VFH算法的无记忆性导致在一些特殊场景中易陷入局部陷阱,本文提出陷阱检测机制的VFH算法,动态选择历史信息增强向量场直方图算法的记忆性,无人机可自主检测陷阱并及时跳出;针对向量场直方图算法的阈值敏感性问题,设计了动态阈值更新策略,使得无人机能够在复杂或稀疏的障碍物环境中,动态平衡避障安全性和抵达目标的时效性。最后,通过对比仿真验证了算法的有效性,为传统VFH算法易陷入局部陷阱的缺陷提供了一种解决方法。The traditional Vector Field Histogram(VFH)algorithm is easy to fall into local traps.Therefore,this paper proposed an improved VFH algorithm based on trap detection mechanism and dynamic threshold updating strategy,which is more conducive to path planning of UAV in local unknown environment;Aiming at the problem of UAV obstacle avoidance in unknown complex environment,an obstacle avoidance algorithm based on A^(*)algorithm and improved VFH algorithm is proposed.Firstly,the UAV constructs the waypoint to the target based on A^(*)algorithm according to the information of the global known obstacles.Secondly,in the case that the target waypoint is unreachable,the UAV performs local planning based on the improved VFH algorithm according to its own motion state and the obstacles information detected by the lidar.In the local planning,the following improvements have been made to the traditional VFH algorithm:Because the traditional VFH algorithm has no memory,it is easy to fall into local traps in some special scenarios.This paper proposed a trap detection mechanism,which dynamically selects historical information to supplement the memory of the VFH algorithm,and the UAV can autonomously detect the trap and jump out in time.In this paper,a dynamic threshold update strategy is designed for the sensitivity threshold of the VFH algorithm,so that the UAV can dynamically balance the safety of obstacle avoidance and the timeliness of reaching the target in a complex or sparse obstacle environment.Finally,the effectiveness of the algorithm is verified by simulation,which provides a solution to the problem that the traditional VFH algorithm is easy to fall into local traps.

关 键 词：旋翼无人机 自主避障 A^(*)算法 向量场直方图算法 路径规划 

分 类 号：V279[航空宇航科学与技术—飞行器设计]