Learning control for a biped walking robot with a trunk

The authors propose learning control methods for compensative trunk motion for a biped walking robot that has a trunk, based on the ZMP (zero moment point) stability criterion, for the cases of the ZMP being inside the stable region and outside the stable region, respectively. They have developed a biped walking robot with a ZMP measurement system and a support device. The results of computer simulation and learning control experiments confirm the convergence of the learning methods and the change of the convergence rate with the change of the weight coefficient. The learning control experiments for the case of the ZMP being outside the stable region show that even though the walking state of the robot itself changes, with the support of a human and by its learning with the ZMP and the support force, stable walking even without the support of a human ultimately is realized.