Hierarchical adaptive control of multiple manipulations by using contact force measurements

The authors present a hierarchical adaptive control algorithm for multiple manipulators grasping a common rigid object. In this hierarchical control scheme, the coordinated control problem is separated into two subproblems: the object control problem which is calculated in the coordinated level (high level) and the manipulator control problem which is executed in the servo level (low level). The only dynamic connection between the two levels is the measurable contact force. In the coordinated level, the coordinated control law gives the required contact forces for each end-effector of the manipulators based on the optimal force distribution rules. In the servo level, the servo control law drives the manipulators to track the required contact forces individually. A theorem shows that the state (position/velocity) tracking error of the object, the joint velocity tracking errors of the manipulators, and the dynamic parameter estimation errors for both object and manipulators are convergent to zero exponentially provided that the consistent exciting conditions are met. Also, the contact force tracking errors are convergent to zero