Abstract
The real-time force control of a quadruped robot involves the optimization of an underdetermined force system subjected to both equality and inequality constraints. A new method for optimal force distribution for the legs of a quadruped robot is presented in this paper. It is characterized by transforming the friction constraints from the nonlinear inequalities into the combination of linear equalities and linear inequalities, by eliminating the linear equality constraints from the original problem to reduce the problem size, and by solving a quadratic optimization problem to meet the needs for quality of solution. The technique is compared with the existing QP Method and Analytical Method to show its superior performance in terms of the problem size, quality of solution and the scape application. The effectiveness of the proposed method is illustrated by giving some simulations of optimal foot force distribution for a quadruped robot.
