RI(RUI): Motion Planning Algorithms for Self-Reconfigurable Robotic Systems

Vassar College, Poughkeepsie NY

Investigators

Abstract

The goal of the proposed research is to develop deterministic motion-planning algorithms to accomplish system-wide reconfiguration of connected clusters of mobile, modular robots, including undergraduate students in all aspects of algorithm development. The first part of this project involves the incremental development and analysis of a suite of algorithms that together comprise a complete deterministic motion planner. The planner is ?complete? in the sense that it will reconfigure a system of n modular, mobile robots, from a connected initial configuration into any equal-sized connected goal configuration. We seek a distributed solution to this problem that is reasonably efficient and provably near-optimal, in terms of number of robot moves and overall reconfiguration time. We have already made progress on the development, analysis, and simulation of algorithms that form the basis for a complete planner for a robotic system composed of hexagonal robots and have also begun to develop algorithms that are resilient to the presence of obstacles in the reconfiguration environment. In the second part of this project, we will focus on making our algorithms adaptive to uncertainty and different module types. The algorithms will be verified experimentally using a number of different visualization tools.