Task-based method for determining topology of reconfigurable modular robot

Reconfigurable robot supplies more choices for flexible manufacturing system, but the complicated tasks, environments and robot modular that can assemble in infinite form are greatly difficult problems for the finalization of robot topology. Genetic algorithms (GA) and iteration algorithms (IA) that in the two layers' search are used in the optimum design of the combination of the robot topology. GA is used to do the evolutionary design for the robot topology, meeting the reach ability in the workspace, the obstacle proximity, the linear and angular distances, the dexterity in the end, the dynamics need of DOF joint, expected number of involved module and other performance criteria. IA is used to resolve inverse kinematics and compute the reach ability in the work point in the space. According to the assemble modular robot and the idea of the modular, the formulas of velocity, acceleration and other essential dynamics equations can be derived. For the purpose of automated modeling, the dynamics problems are solved by means of an iteration compensative method. The highly adaptable topology that obtained in the search using the professional experiments is modified to meet requests of the actual work. Finally, a practical example is given.