Hybrid Reasoning for Mobile Manipulation based on Object Knowledge

Abstract

Autonomous mobile manipulation in every day environments is a key feature for next generation service robots. However, the diversity of problems that need to be solved by a robot is vast. Detailed knowledge about the environment, surrounding objects, and handling instructions is required. In this paper we combine object-centered hybrid reasoning with reachability analysis for autonomous spatial reasoning on mobile manipulation tasks. Object knowledge is used to distinguish the right place and time to move a mobile platform in order to solve a given task. The methods are evaluated in an experiment with the humanoid robot Rollin’ Justin.