A CBR based Pure Reactive Layer for Autonomous Robot Navigation

C. Urdiales, E.J. Pérez, F. Sandoval, J. Vázquez-Salceda, and M. Sànchez-Marrè (Spain)


reactive navigation, autonomous robot, case-based reaso ning


This paper presents a new reactive layer using Case-Based Reasoning (CBR) for a hybrid navigation architecture al ready including a deliberative layer that performs all global planning. Thus, the reactive layer only relies on immediate information about sensors readings, robot heading and re lative position to the goal. The proposed layer allows effi cient navigation in cluttered environments without signifi cant oscillations. It can easily be trained to deal with hard situations like navigating between two close obstacles. The layer has been succesfully tested in real environments.

Important Links:

Go Back