Whole-body Motion Planning – Building Blocks for Intelligent Systems

Humanoid robots have become increasingly sophisticated, both in terms of their movement as well as their sensorial capabilities. This allows one to target for more challenging problems, eventually leading to robotic systems that can perform useful tasks in everyday environments. In this paper, we review some elements we consider to be important for a movement control and planning architecture. We first explain the whole-body control concept, which is the underlying basis for the subsequent elements. We then present a method to determine optimal stance locations with respect to a given task. This is a key element in an action selection scheme that evaluates a set of controllers within a parallel prediction architecture. It allows the robot to quickly react to changing environments. We then review a more global movement planning approach which casts the overall robot movement into an integral optimization problem, and leads to smooth and collision-free movements within interaction time. This scheme is then extended to cover the problem of grasping simple objects.