Autonomous decentralized approach in the PTP control of redundant manipulators

A new control strategy is proposed for the point-to-point (PTP) control of redundant multi-link manipulators. On the basis of this strategy, two control schemes are introduced where each manipulator joint is considered as an individual unit. Scheme I provides an autonomous decentralized control law, the execution of each joint requires only information on its own position, its final goal position and the position of the end-effector, therefore no interactions exist among joints and the system is robust against partial faults in the joints. In Scheme II, although each joint executes its action autonomously, in each step, only one joint rotates. A criteria is employed to decide the order of the execution so that a roughly straight-line end-effector trajectory can be automatically generated without the need of providing intermediate targets. With this scheme, it is also expected that a closed joint motion is generated from any closed Cartesian trajectory, thus no intermediate re-orientation manoeuvres are required during a repeated manipulation