Action planning model for autonomous mobile robots

Abstract

The main challenge for autonomous mobile robots is to interpret a new situation correctly and to react appropriately in unknown environments. Autonomous mission completion in new territory requires flexible, adaptable, complex, but also fast action planning. For most applications behavior selection cannot be simply rule-based, requiring more sophisticated evaluation systems delivering appropriate information and interpretation. This approach shows a solution using three control systems of different complexity, time lines and flexibility. One of the key modules of the behavioral software architecture schedules competitive actions proposed by the different control systems. The design has been tested in simulation, where mobile robots armed with different behavioral software models can compete in different game-like situations. The simulation has shown that cooperation between agents emerges to accomplish different situations.