基于预探索的动态自适应3DVFH+无人机避障算法

Dynamic adaptive 3DVFH+ UAV obstacle avoidance algorithm based on pre-exploration

为解决无人机在未知复杂环境下的自主避障规划问题,同时考虑到旋翼无人机平台硬件资源稀缺的问题,提出了一种基于预探索的动态自适应3DVFH+算法。根据无人机飞行进度以及障碍物信息,动态调整代价函数的自适应权重因子;通过对无人机视野范围的预探索,构建了搜索树,并对障碍物及无人机模型进行了建模研究,同步实现了无人机在复杂环境中的可视化分析。结果表明,动态自适应代价函数消除了3DVFH+算法的死区现象;预探索机制解决了算法的局部最优解问题,规划航线缩短37.15%,所需时间减少24.48%。同时搭建了复杂度较高的森林障碍物进行验证分析,相较于使用八叉树框架建图的3DVFH+算法执行相同避障规划任务时节省内存可达25.26%。

In order to solve the problem of autonomous obstacle avoidance planning of UAV in unknown complex environment,and considering the scarcity of hardware resources of rotor UAV platform,a dynamic adaptive 3DVFH+ algorithm based on pre-exploration is proposed.According to the UAV flight schedule and obstacle information,the adaptive weight factor of the cost function is dynamically adjusted;By pre-exploring the visual field of UAV,a search tree is constructed.The obstacle and UAV models are modeled,and the visual analysis of UAV in complex environment is realized synchronously.The results show that the dynamic adaptive cost function eliminates the dead time of 3DVFH+ algorithm;The pre-exploration mechanism solves the local optimal solution problem of the algorithm.The planned route is shortened by 37.15% and the time required is reduced by 24.48%.At the same time,a forest obstacle with high complexity is built for verification and analysis.Compared with 3DVFH + algorithm using octree frame to build maps,it can save 25.26% memory when performing the same obstacle avoidance planning task.