The DD&P Robot Control Architecture - A Preliminary Report

The paper presents current results of our work on DD&P, a two layer control architecture for autonomous robots. DD&P comprises a deliberative and a behavior-based part as two peer modules with no hierarchy among these two layers, as sketched in [HJZM98]. Interaction between these control layers is regulated by the structure of the information flow, extending the "classical" sense-model-plan-act principle. The paper stresses two architectural highlights of our approach: The implementation of the "Plan-as-Advice" principle to execute plan operators by the behavior-based part, and the grounding of the symbolic planner description via chronicle recognition.