The constrained control of force and position by multilink manipulators

The efficiency of a manipulator with actuators in each single joint can be improved considerably (up to a factor of 2) by adding actuators that act over more than a single joint. Even though this creates redundancy the efficient control of force and position of the manipulator end point induces a constraint that eliminates (part of) the redundancy. It is demonstrated that an efficient control requires that monoarticular actuators contribute positive work to the total work done by the manipulator. Without biarticular actuators this requirement cannot be met and some monoarticular actuators will dissipate work delivered by other actuators. The role of biarticular actuators is to transfer the mechanical output (work or torque) of some monoarticular actuators to another joint