Robotic Path Planning Based on Episodic-cognitive Map

Inspired by mammal's spatial awareness and navigation capabilities, a new episodic-cognitive map building and path planning method was proposed, used for navigation tasks of mobile robot under the unstructured environment. Combined with characteristic of cognitive map and simulated the formation mechanism of episodic memory in the hippocampus, a novel episodic-cognitive map encapsulated the information of scene perception, state neuron and pose perception was built, realized the real-time, incremental accumulative and updating cognition of the robot to the environment. Based on the episodic-cognitive map, using the minimum distance between events, an algorithm of the event sequence planning was put forward for preferred trajectory choosing. Experimental results showed that the proposed algorithms realized the mobile robot choose the preferred planning path, and using the SIFT-based visual navigation method, the mobile robot can reach the target very well. The method in this paper extended the application of biological cognition theory in the field of robotic planning.