Coordination dynamics and model-based neural network synchronous controls for redundantly full-actuated parallel manipulator

Abstract

Redundantly full-actuated parallel manipulator takes number of actuations exceeding its degree of freedom, and actuation coordination makes its basis of stable operation. This paper studies the coordination dynamics of general redundantly full-actuated parallel manipulator and derives coordination dynamics models for driving force coordination and internal force regulation respectively. Associated with coordination dynamics models, two neural network synchronous control methods are proposed for each situation correspondingly. Self-learning synchronous algorithms for those methods are designed additionally. Manipulator 6PUS+UPU is taken as a prototype for co-simulations and experiments. Results reveal that the two methods proposed above could improve actuation coordination and internal force precision of redundantly full-actuated parallel manipulator respectively. This paper provides new dynamics-based control methods for the research and control application of parallel manipulator with redundant actuation.