An approach to motion planning for mobile manipulation

Increasing the flexibility of robot systems has been one of the main objectives of robotics research in recent years, There are various approaches to achieve this: increasing fault-tolerance on different levels, the development of sophisticated sensors, grippers and manipulators, and the use of learning techniques. But the field of operation of such smart systems is still limited to a rather small area. Mobile manipulation is another facet of increasing the flexibility of a robot system. By exploiting the mobility of a platform, the dexterous workspace of manipulators can be considerably increased. One main research topic on mobile manipulation is the decomposition of the motion of the tool-center-point into manipulator motion and platform motion. In this paper a new online approach to this decomposition is presented which allows the consideration of complex obstacles and multiple manipulators mounted on the platform. Simulation studies verify the suitability of the approach. Some of the realized algorithms have been applied to a real mobile two-arm system