基于拉普拉斯人工势场的无人机避障控制

Path planning and obstacle avoidance for UAVbased on Laplacian potential field

针对无人机在复杂环境中的自主导航问题,设计基于拉普拉斯方程的人工势场,并以人工势场为基础进行无人机的运动控制。对地图中阻碍边界与目标边界分别设置初始条件,采用边界元法求解拉普拉斯方程。由此建立调和势场,它对障碍物形态适应程度高,不存在局部极小点,并可快速计算内部任一点的梯度。以构建的梯度场作为参考速度,采用线性二次型高斯控制器进行速度追踪,从而完成无人机的姿态控制。采用Airsim仿真平台进行不同场景下的模拟实验,结果表明算法可以有效引导无人机完成路径规划与避障控制,且在遭遇意外干扰时具备良好的调整适应能力。

We propose a method for autonomous navigation of unmanned aerial vehicles(UAVs)in cluttered environment utilizing Laplacian potential field.With suitable boundary conditions for obstacles and goal in the map,Laplace’s equations are solved by boundary element method(BEM).Then we establish artificial potential field which is capable of handling obstacles with complex configuration and guarantees nonexistence of local minima.Since such a kind of potential field has easy access to the gradient of arbitrary point in the map,a linear quadratic Gaussian controller(LQG)is designed for velocity tracking to help complete attitude control.The simulation is conducted on Airsim platform with different scenarioes,and the results illustrate that this method is valid for path planning and obstacle avoidance and has adaptability to unexpected disturbance.