A Framework for Adaptive and Goal-Driven Behaviour Control of Multi-robot Systems

Robot and multi-robot systems are leaving the friendly well-structured world of automation and are facing the challenges of a dynamic world. Such uncertain conditions call for a high degree of robustness and adaptivity for individual robots as well as for the organization of multi-robot systems. This corresponds to the concepts of self-adaptation and self-organization. Robots adapting to the dynamic environment still have to pursue their given tasks or goals. In order to address the requirements of creating adaptive and goal-driven multi-robot systems, it is necessary to combine existing goal-directed planning and decision-making approaches with self-adaptation and self-organization mechanisms. This work addresses this challenge with a new hybrid approach integrated into a common robot framework, combining symbolic planning with reactive behaviour networks, machine learning, and the pattern-based selection of suitable mechanisms. On that account it brings together the advantages of bottom-up and top-down oriented approaches.