PhD Position: Inverse Reinforcement Learning of Cheetah Locomotion

The job advert seeks a student to investigate the neuromechanics of legged manoeuvrability in the wild, specifically focusing on the cheetah as a model animal. The goal is to utilize inverse reinforcement learning to understand how animals negotiate trade-offs between competing requirements during highspeed manoeuvres. The study will investigate the best reward function to maximize for control of legged robotic systems inspired by animal motion. Inverse Reinforcement Learning (IRL) will be utilized to produce a reward function based on observed optimal motions. The successful student will investigate the application of IRL to analyse the mechanical system behaviour and rapid acceleration, as well as the application of cheetah reward functions to quadruped robots. This research aims to provide insights into the dynamics and control of legged locomotion in animals, which can inform the development of
robust and agile robotic systems. This project is fully funded by Mathworks.