Connectionist approach for cognitive map learning and navigation based on spatio-temporal experiences
一种基于连接机制和时空经验的认知地图学习与导航方法

A connectionist method is proposed for mobile robot, which lacks a priori environmental model and global localization information, to learn goal directed cognitive map from its own spatio temporal experiences. Temporal sequence processing network (TSPN), which is constructed at run time, provides compact representations of history perceptive information, transforms spatial knowledge into cell firing characteristics and retrieves them in later runs to guide the robot. The navigation system integrating TSPN and a reactive safeguard module performs dynamic landmark and heading detection, route learning and collision free real time navigation in noisy environments. The simulation and real world experiments demonstrate the effectiveness and flexibility of the system.