Real-Time Guidance of Quadrotor for Obstacle Mapping Using Vision System
Jongho Park, Youdan Kim · 2015
A real-time guidance algorithm for obstacle mapping is proposed. Stereo vision and ultrasonic sensor are assumed to be mounted on the quadrotor to get the information. Depth map obtained by the stereo vision is utilized to detect and map the obstacle. Mapping an obstacle is performed by utilizing obstacle data point acquired from the depth map. Quadrotor states are used to generate a path to map the obstacle. The quadrotor moves around the obstacle to map the overall shape of the obstacle while avoiding a collision. The quadrotor also performs vertical maneuver if the image plane of the vision system cannot cover the whole figure of the obstacle. Numerical simulations are performed to demonstrate the performance of the proposed guidance algorithm.