Extension Approach for the Behaviour-Based Control System of the Outdoor Robot RAVON

This paper describes the extension of a behaviour-based control system for autonomous outdoor navigation. To perform robust obstacle avoidance the existing stereo vision system is complemented with a 2D laser scanner. Instead of combining raw sensor data, the additional information is integrated exploiting behaviour fusion. The scanner data processing is wrapped into new behaviours, which are added to the existing network. That way no modifications to existing behaviours and interconnections are necessary. Furthermore sensors can be evaluated individually augmenting traceability and testability. The performance of the resulting system is shown in an experiment carried out on the outdoor robot RAVON.