Dynamic walking: a step forward?

After a long time as a junior partner in robotics, walking has seen an explosion in interest and achievements in the last 15 years as computational power unimaginable until recently has been brought to bear on its problems. At the same time we have seen advances in other related fields: non-linear dynamics has helped us understand and model the robotic systems that we are studying; zoology has provided details of the mechanics of animal locomotion to aid in the design of our robots; and neuroscience has developed a deep understanding of how rhythmic movements like walking occur in the natural world, and this has helped us formulate new ideas for robotic controllers. In this chapter we shall look at the history of walking research. We shall follow how it has progressed from the study of body parts and the construction of clockwork automata to the understanding of neural rhythm generators and the control of robotic somersaulting. We shall see that it has become a focal point of interdisciplinary research between the biological sciences and robotics, and we shall examine what can be gained from this work. We shall also look at what commercial uses walking robots are currently being put to and whether they hold out hope for the future. We shall see to which applications modern walking robots can be applied, and we shall examine what prospects there are for research in these areas.