Departmental Seminar: (Machine) Learning Cheetah Locomotion

Title: (Machine) Learning Cheetah Locomotion

Abstract: Animals are extremely agile. However, we are still far from understanding exactly how they can manoeuvre so rapidly. This is crucial for understanding predator/prey interactions, in terms of the success (or failure) of hunts, the associated energetic costs and injury risks to animals. This understanding will enable a comprehensive view of locomotion, which has natural implications for biomechanics, ecology, evolutionary biology, and will be equally important to fast legged robots if they are to move at high-speed outside the safety of the laboratory.
In this talk I will discuss my lab’s efforts towards the challenging problem of understanding the locomotion of the cheetah, the pinnacle of manoeuvrability. We do this by through the lens computing by developing a broad range of systems and techniques including multi-body modelling, feedback control, computer vision & deep learning, physical robots, as well as trajectory optimisation. In summary, our research on cheetah locomotion provides valuable insights into animal agility and has far-reaching implications for biomechanics, ecology, evolutionary biology, and the development of high-speed legged robots.