Distributed Central Pattern Generator Control for a Serpentine Robot

We have built a biologically and neurally inspired autonomous mobile robotic worm. The main aim of the project is to demonstrate elegant motion on a robot with a large number of degrees of freedom (DOFs) under the control of a simple distributed neural system as found in many animals’ spinal cord. Our robot consists of individually controlled segments that exhibit Central-Pattern-Generator (CPG) -driven biomorphic motion. An important aspect of the project is to achieve a level of modularity while closely mimicking the neural control of e.g., the lamprey. This paper presents our robotic platform and the distributed CPG control algorithms. We will mainly focus on the architecture of the initial system and on future developments, and also report some preliminary experimental results. Motivation Among the vast body of work related to serpentine robots (see [3] for an extensive overview, and [4][6]), the issues of modularity and robustness as well as distributed biologically plausible control have seldom been addressed in a single study. The motivation behind our WormBot project [2] is to demonstrate elegant robust robotic motion based on simple, yet biologically plausible design principles in a high-DOF system. We investigate in the motion generated by multiple 1-DOF segments that are individually controlled by local CPGs, but achieve overall motion stability through shortand long-range couplings. A robotic platform to evaluate motion in such a system is not commercially available. Thus the focus of the project reported here is twofold: On the one hand, we developed the serpentine robot WormBot, which allows us to explore issues pertaining to the control of a high number of DOFs, modularity and inter-module communication. We aim Paulina Varshavskaya AI Laboratory Massachusetts Institute of Technology 200 Technology Square USA Cambridge, MA 02139