no code implementations • 28 Nov 2023 • Yusuke Miyashita, Dimitris Gahtidis, Colin La, Jeremy Rabinowicz, Jurgen Leitner
In this paper, we propose the use of generative artificial intelligence (AI) to improve zero-shot performance of a pre-trained policy by altering observations during inference.
no code implementations • 6 Mar 2021 • Brendan Tidd, Akansel Cosgun, Jurgen Leitner, Nicolas Hudson
While we show the feasibility of our approach in simulation, the difference in performance between simulated and real world scenarios highlight the difficulty of direct sim-to-real transfer for deep reinforcement learning policies.
no code implementations • 23 Jan 2021 • Brendan Tidd, Nicolas Hudson, Akansel Cosgun, Jurgen Leitner
Dynamic platforms that operate over many unique terrain conditions typically require many behaviours.
no code implementations • 1 Nov 2020 • Brendan Tidd, Nicolas Hudson, Akansel Cosgun, Jurgen Leitner
Legged robots often use separate control policiesthat are highly engineered for traversing difficult terrain suchas stairs, gaps, and steps, where switching between policies isonly possible when the robot is in a region that is commonto adjacent controllers.