Online walking control system for biped robot with optimized learning mechanism: an experimental study

摘要

This paper solves the problem of dynamic balance control for a biped robot with good energy efficiency. A new multi-joints control scheme is proposed, which contains a single hidden layer feedback neural networks mechanism. To improve the control performance, an optimized learning algorithm is integrated in our scheme, in which the interval type-2 FLSs are developed to deduce the learning weights of training data. Finally, experiment results are provided to validate the proposed method.