A CBR System for Autonomous Robot Navigation

In this paper we propose the use of case-based reasoning techniques to improve the navigation of an autonomous robot in unknown semistructured environments. At the moment the current goal is to identify problematic situations (such as dead ends or obstacle layouts that the robot is not able to avoid) and take the proper actions in order to avoid them. As the first steps we propose a similarity function to retrieve similar past cases. We integrate a CBR agent into an existing multiagent navigation system in order to evaluate the performance of the CBR system. The results obtained through simulation show that the new system not only prevents the robot from getting blocked in certain situations, but also improves the performance in terms of time and distance of the path taken to reach the target.