BRAIN-INSPIRED COGNITIVE MAP BUILDING FOR MOBILE ROBOT

Inspired by the spatial localization mechanism of brain, this work proposes a cognitive map building method for the mobile robot, which can afford robot the cognition and navigation ability in the unknown environment. This work models the grid cells and place cells to afford the robot spatial position information and extracts the state neurons to imitate place cells to represent its different positions, i.e., states. The robot system activates different state neurons when the robot enters different positions. These state neurons connect with each other to construct the episodic memory, and these spatial positions obtained from grid cells and place cells form the topological description of the environment, i.e., cognitive map. Based on the memory-based motion trajectory planning method, the robot can localize itself relative to its experiences in the memory space and anticipate a preferred motion trajectory to the target. The experimental results show that the proposed method can effectively and accurately build the cognitive map and achieve the robotic navigation.