On sensor-based roadmap: a framework for motion planning for a manipulator arm in unknown environments

We present a general framework for sensor-based motion planning for an articulated robot arm with many degrees of freedom. We assume a sensor that measures distances in physical space (a range sensor). Inspired by recent work in classical motion planning, the crux of our framework is to represent the connectivity of the configuration space (C-space) with a graph structure, the roadmap. This roadmap is then incrementally built. We have performed some simulations that show the promise of our approach.