Date of Original Version
Abstract or Description
Kinematic motion planning often requires a notion of distance between configurations. Euclidean distances on a parameter space are easy to compute, but can drastically distort the effort required to change configuration. Here, we present a framework for characterizing this distortion, based on principles adopted from the cartographic community, and a method for transforming configuration coordinates to better represent actuation costs. As a demonstration of this approach, we derive a true configuration distance metric for an important class of locomoting systems: low Reynolds number swimmers. Applying our cartographic coordinate transformation to these systems both provides intuition for previous numerical results, and allows direct geometric comparison between systems with heterogeneous morphology.
Proceedings of Robotics: Science and Systems.