Reconfiguration planning for modular self-reconfigurable robots

Modular Self-reconfigurable (MSR) robotics is a new technology with the potential to significantly expand the domain of robot applications. The key characteristic of MSR robots is their ability to change configuration automatically, enabling them to adapt their shape to suit multiple, changing tasks. The ability to self-reconfigure translates into great versatility, making MSR systems true general-purpose robots. However, self-reconfiguration introduces a difficult planning problem as the number of modules (and degrees of freedom) in a robot increases. In addition, different hardware designs require strikingly different approaches to reconfiguration. To date, few solutions have been proposed, and no practical algorithms exist yet for some categories of MSR robots. This thesis presents a new planning architecture for a category of MSR robots known as Chain Reconfiguration MSR robots. These are characterized by the fact that, in order to reconfigure, modules move in groups or “chains”. This is in contrast to other MSR robots, where reconfiguration happens as a series of individual module motions. Chain Reconfiguration introduces several planning requirements: solving the kinematics (forward and inverse) of groups of inter-connected modules, maintaining overall stability and avoiding collisions as the module chains move. The Reconfiguration Planner in this thesis is specifically suited to Chain Reconfiguration MSR robots, being the first of its kind. The method proposed is based on the integration of two complementary components: a Connectivity Planner and a Motion Planner. The Connectivity Planner operates in a topological domain, generating an ordered sequence of module connectivity changes. This sequence is passed to the Motion Planner, which computes feasible motion trajectories, accounting for geometry and physics, for the modules involved in each connectivity step. Combining the two planners results in a complete Reconfiguration Planner. The effectiveness of the planning strategy was tested using graphical simulations of PolyBot, an MSR robot developed at the Xerox Palo Alto Research Center. The examples presented span a range of useful manipulation and locomotion configurations, and show various PolyBot robots completing reconfiguration while avoiding external obstacles and self-collisions, and maintaining stability. Performance issues and limitations are discussed, along with recommendations for future research.