融合危险度模糊分级的UAV动态避障方法

UAV dynamic obstacle avoidance method based on fuzzy classification of risk

针对无人机(UAV)在多障碍环境中面对运动障碍的反应式避障问题,为了在反应避障决策时能够朝向周围障碍更远、更少的方向进行避障,提出了一种融合障碍危险度模糊分级的三维速度障碍避障方法。通过三维速度障碍法,给出探测域内各运动障碍的碰撞预测。选取坐标系三轴方向作为相互独立的可用避障加速度方向,在速度及加速度限制下求解三轴方向上避障加速度方向以及数值区间。通过模糊算法,给出了周围其他静止障碍的危险分级方法,提出优化评价函数对计算各方向上的最优加速度,给出优化避障指令。仿真结果表明,针对复杂环境,UAV在反应避障的同时考虑了其他障碍的影响,相比原算法拥有了更高的避障成功率和更少的避障次数。

Aiming at the problem of reactive obstacle avoidance of UAV facing speed obstacles in a multi obstacle environment,in order to avoid obstacles farther and in less directions towards the surrounding obstacles in response to obstacle avoidance decisions,a three⁃dimensional speed obstacle avoidance method based on fuzzy classification of obstacle risk is proposed.The geometric analysis of collision prediction for each speed obstacle in the detection domain is given by using the three⁃dimensional speed obstacle method.The three⁃axis direction of the coordinate system is selected as the mutually independent available obstacle avoidance acceleration direction,and the obstacle avoidance acceleration direction and numerical interval in the three⁃axis direction are solved under the restriction of speed and acceleration.Through the fuzzy algorithm,the danger classification method of other obstacles around is given,the optimal evaluation function is proposed to calculate the optimal acceleration in all directions,and the optimal obstacle avoidance command is given.The simulation results show that,for complex environments,UAV considers the influence of other obstacles while responding to obstacle avoidance,and has higher success rate of obstacle avoidance and fewer times of obstacle avoidance than the original algorithm.