Dynamic Coordination of Multiple Robot Arms With Flexible Joints

Abstract

A nonlinear decoupling and linearizing feedback control for dynamic coordination of multiple robot arms handling an ob ject constrained by the environment is considered. A general method is presented that ensures an exact feedback lineariza tion for cases of both rigid and flexible joint arms, as the joint flexibility can cause instability of robotic system control. The method takes the manipulator dynamics and object dynamics into consideration, along with material constraints imposed on the system (mixed contact conditions between arms and object and between object and the environment) and program con straints (specified internal forces, contact forces, and object motion). A method for internal force parameterization, which uses a set of kinematic constraints imposed on the motion of the robot end effectors, is proposed. Control algorithms are developed for redundant flexible joint manipulators designed to have more joints than necessary for given tasks. The appli cation of the control law produces a set of decoupled linear subsystems defined in subspaces that have different physical meaning. The implementation of this method is demonstrated for the case of a system of two planar three-link arms with the end effectors manipulating an object and with different constrained task configurations. Numerical experiments demon strate high-precision tracking capabilities of the method, prove that it can be implemented on fast contemporary computers, and show that robustness to bounded modeling errors and external disturbances can be achieved.