Biologically Inspired Motor Control for Underactuated Robots: Trends and Challenges

There has been an increasing interest in the study of underactuated robotics toward significant improvement of behavioral performances in robotic systems. The previous studies on underactuated robots have demonstrated energy efficient, maneuverable, and robust motion control by properly exploiting passive dynamics. Despite these fascinating characteristics in underactuated systems, however, there are a few particularly important challenges which are related to the control of nonlinear dynamics derived from physical interactions with the environment. Because passive dynamics induces nonlinear system-environment interactions, it is highly difficult to develop control architectures for the precise control of the systems, and thus they are able to exhibit only limited variations of behavior patterns. In order to tackle with the challenges, this article will discuss three important research directions that we have been pursuing, i.e. mechanical feedback for self-stability, design optimization for behavioral diversity, and control optimization for precise motor control and planning.