Control of multimanipulator systems-trajectory tracking, load distribution, internal force control, and decentralized architecture

The author proposes a coordinated control law for a multimanipulator system performing parts-matching tasks. This control law enables the manipulators to perform the preplanned parts-matching maneuver while the entire parts-matching system is driven to follow a desired path. Manipulators are essentially treated as six-degree-of-freedom actuators with some nonlinear dynamics, which exert a set of contact forces on the object so that trajectory tracking is achieved and the desired internal force is realized. When the parts-matching system consists of only a single object, the control law degenerates to an expression that will drive a group of manipulators transporting a single object. A load-sharing scheme minimizes the weighted norm of the force applied to the object. In this way, a heavily weighted direction tends to get less load. This scheme does not require a force sensor. The author also discusses the choosing of the weighting factor and shows that the proposed control law can be implemented in a decentralized fashion