Obstacle avoidance for quadrotor using improved method based on optical flow

摘要

In this paper, we proposed an improved method based on optical flow to help a low-cost quadrotor avoid obstacles. The method is consist of improved balanced strategy and fast Time-to-Collision (TTC). The improved balanced strategy is designed for the quadrotor to achieve 6 degree of freedom movement. We use a learning method based on optical flow data when the vehicle can avoid the obstacle successfully. The fast TTC method helps extract the relative depth information of the environment. We also introduced a monitoring strategies to supervise our method in real time. All these information is combined to avoid the obstacle. Our method has been demonstrated to work well in corridor environment.