Autonomous Navigation and Path Planning For Ground Vehicle in Indoor Environment

摘要

In this paper, implementation of a vision based robust and inexpensive autonomous navigation system has been discussed. A system which enables determination of optimal path in real time between initial position and desired position in a cluttered environment with help of stereo vision has been presented. Stereo vision has been used for obstacle identification followed by range estimation through generation of disparity map obtained for each frame. Furthermore, upon object identification and calculation of corresponding distance to each object landmark identification is done for building map of the environment for the robot. Following this, implementation of path planning algorithms on the map help in generating waypoints for autonomous traversal with ease. This proposed system enables navigation of a robot based on dynamic mapping of environment and path planning in complex scenarios through stereo vision in real time.