In this paper we describe an algorithm that allows mobile robots to predict the reaction of dynamic obstacles. The assumption is that dynamic obstacles are likely to be another mobile robot equipped with reactive collision avoidance capabilities. Therefore, a Heuristic Guidance Module (HGM) is presented that uses Kalman filter estimates from distance measurements to predict how a dynamic obstacle is reacting to the possibility of a collision with a mobile robot. The HGM then calculates utilities for the various reactive options that the mobile robot is capable of undertaking. Results show that the prediction method is effective, and that the HGM can efficiently handle a case when a mobile robot and a dynamic obstacle are on mirrored paths.

This content is only available via PDF.
You do not currently have access to this content.