THE NEXT CHALLENGES IN BIO-INSPIRED ROBOTICS

Nature has been providing a number of inspirations to engineers and scientists in the history of robotics. Many important concepts and theories ranging from mechanical designs to control theories and computational algorithms were developed on the basis of our careful observations and experimentations of biological systems. One of the important lessons from the previous bio-inspired robotics research is that the adequate abstraction of mechanisms in nature gives us substantial impact in the technological development of advanced mechatronic systems. As one of such important design principles originated in recent bio-inspired robotics research, there has been an increasing interest in the use of mechanical dynamics in control of robot behaviors. For example, the mechanical dynamics was nicely exploited to demonstrate many different kinds of animal-like complex behaviors such as walking, running, hopping, swimming, flying, and dancing. There were two important lessons that we learned form this line of research. First, it was identified that the use of mechanical dynamics significantly improves motor control in terms of energetic costs, which is beneficial for mobile robots in particular. And second, the proper designs of mechanical dynamics can provide self-stability to maintain robot trajectories against undesired deviations, which is also known as "mechanical intelligence". Although this approach has shown a number of impressive demonstrations, some additional breakthroughs seem to be necessary toward high-impact applications. In general, one of the fundamental questions is how to scale up the complexity of the systems that rely on passivity-based control. The passivity-based dynamic walking control is, for example, still limited to relatively simple task environments, and it is still not fully understood how we could scale up. Essentially, when a robot exploits more mechanical dynamics, the system has to suffer from more design and control parameters to be calibrated or optimized in order to achieve sensible motion control. From this perspective, in this talk, we will discuss the next high-priority challenges and potential technological solutions in the field of bio-inspired robotics in order to achieve significant breakthroughs toward practical applications.