Modeling humanoid robots

The notion of humanoid robot was introduced in the early 1970s to describe mobile robots with anthropomorphic characteristics, as opposed to industrial robots attached to a fixed base. The aim of humanoid robotics research is to approach human locomotion performance as closely as possible. Since the early work of Ichirô Katô and his colleagues at Tokyo's Waseda University, remarkable progress has been made, particularly in studying the stability of a humanoid robot, and its walking and running gaits. Nevertheless, the performance of a humanoid robot still needs to be significantly improved before it can rival that of a human being. Friction in joints and mechanical transmissions is not negligible, whereas it is virtually non-existent in humans. The humanoid robot's autonomy, power-to-weight ratio, shock resistance, optimal mass distribution, locomotion over rough terrain, the ability to perform safe tasks in collaboration with a human being, and its ability to evolve in an obstacle-ridden environment are all challenges that remain relevant today. Modeling these robots is an essential aspect of research in this field, thanks in particular to highly accurate physical models that take into account contact phenomena and robot dynamics.