A load distribution scheme for multi-arm coordinating robots

Load distribution algorithms are proposed along the line of nonlinear programming. In a certain sense, each algorithm obtained provides an optimal force distribution of K coordinating manipulators in which the force magnitude of each actuator is assumed to be bounded. To be specific, the objective function adopted is the p-norm of the joint forces as the p-norm is differentiable and approaches to the infinity norm as p to infinity , meaning that, with a sufficiently large p, constraints on joint force magnitude can implicitly be incorporated into the objective function so that one actually deals with an unconstrained optimization problem. Moreover, it is shown that if the optimal distribution algorithm is applied only to the load components greater than a given threshold, the p-norm type of objective function is strictly convex that enables one to use the classic Newton's method to find the solution with an order-two convergence rate.<<ETX>>