An RRT That Uses MSR-Equivalence for Solving the Self-Reconfiguration Task in Lattice Modular Robots

Abstract

A modular self-reconfigurable robot can change its configuration to effectively adapt itself to perform a wide range of tasks. The Modular Self-reconfigurable Robot (MSR) is constructed from a collection of connected modules that can move relative to each other and change the system's shape. This research presents a novel sliding Triangular Lattice Modular Robot (STRIMOR). We use a Rapidly-exploring random tree (RRT*) algorithm to solve the self-reconfiguration (SR) problem. This algorithm is computationally expensive due to the configuration space dimension. In order to increase the number of modules we define ‘Topological Configuration’ which better describes the configuration's unique properties. This new approach significantly reduces the tree size by representing a set of MSR-equivalence configurations as a single topological configuration. Furthermore, we examine the advantage of moving several module-groups in a single movement and choose their number by several density functions. Our improvements are presented using a set of simulated SR problems with different complexity levels. On top of that, We designed and fabricated the STRIMOR's dual-type modules and discuss how such a design simplifies the system's mechanical complexity.