Iterative optimal control synthesis for nonlinear switching systems

Abstract

This paper introduces computational control solutions to the constrained optimal control problem for nonlinear switching systems. The proposed approaches enable direct incorporation of the system's full-scale dynamics and also allow the system to operate with multiple irregular (distinct) switches. The optimal control signal is synthesized by iteratively solving a sequence of linearly constrained quadratic programs. The effectiveness of the proposed framework is illustrated on the control synthesis of different legged robot models.