Online decision of foot placement using singular LQ preview regulation

Abstract

In this paper, we introduce an online decision method of foot placement for legged robots that manages unknown external forces. It consists of a fast trajectory generation method and an optimization method of foot placement. The fast trajectory generation method is based on an explicit solution for singular LQ preview regulation problem of an inverted pendulum model. By using the regulator problem, the conditions of target ZMP trajectory that will not make the COM trajectory diverge and a fast generation method of the COM trajectory which satisfies the target ZMP trajectory are obtained. An online optimization of foot placement is realized by using the fast trajectory generation method. An experiment of the online decision method of foot placement that manages unknown external force shows the performance of the proposed method.