A goal-oriented autonomous mapping and exploration system that combines reactive and planned robot navigation. First, a navigation policy is learned through a deep learning framework in a simulated environment. This policy guides an autonomous agent toward a goal while avoiding obstacles. We develop a navigation system that integrates these learned policies into a motion planning stack as a local navigation layer to move the robot to the intermediate targets. A global path planner is used to mitigate the local optimal problem and guide the robot to the global goal. Possible intermediate target locations are extracted from the environment and used as local targets according to the heuristics of the navigation system. The fully autonomous navigation is performed without any prior knowledge, while the mapping is performed as the robot moves through the environment. After the goal is reached, navigation continues between the goal and the point of origin. A plan is first obtained from a path planner on the newly explored map, and navigation is performed over the DRL network.

More information at: https://github.com/reiniscimurs/GDAM

Please take the opportunity to connect and share this video with your friends and family if you find it helpful.