Hierarchische hybride Planung für mobile Roboter

To act autonomously and robustly in complex environments, mobile robots require task planning to adapt to the given situation. With the growing capabilities of robot hardware and software this is becoming increasingly important. Yet, only few integrated robotic systems exist. This thesis attempts to make a next step towards robust plan-based robot control. For this, it first describes how the HTN-Planner SHOP2 [117] can be used in a robot control architecture that aims to improve its performance based on experiences. Based on that, means to increase the systems robustness by integrating the planner with other components are discussed. The environments and tasks of mobile robots contain various forms of knowledge like information about temporal requirements or resources that should be reasoned about in order to achieve a safe and robust behaviour. However, this cannot be fully used by planners like SHOP2. Therefore, this thesis presents the hierarchical hybrid planner CHIMP which combines the advantages of hierarchical planning and hybrid planning with different forms of knowledge as a MetaCSP [28, 104]. Furthermore, CHIMP’s plans can contain actions that can be executed in parallel, and additional goal tasks can be inserted into an existing plan during its execution.