Showing 1–1 of 1 results for author: Bagajo, J

DiffSim2Real: Deploying Quadrupedal Locomotion Policies Purely Trained in Differentiable Simulation

Authors: Joshua Bagajo, Clemens Schwarke, Victor Klemm, Ignat Georgiev, Jean-Pierre Sleiman, Jesus Tordesillas, Animesh Garg, Marco Hutter

Abstract: Differentiable simulators provide analytic gradients, enabling more sample-efficient learning algorithms and paving the way for data intensive learning tasks such as learning from images. In this work, we demonstrate that locomotion policies trained with analytic gradients from a differentiable simulator can be successfully transferred to the real world. Typically, simulators that offer informativ… ▽ More Differentiable simulators provide analytic gradients, enabling more sample-efficient learning algorithms and paving the way for data intensive learning tasks such as learning from images. In this work, we demonstrate that locomotion policies trained with analytic gradients from a differentiable simulator can be successfully transferred to the real world. Typically, simulators that offer informative gradients lack the physical accuracy needed for sim-to-real transfer, and vice-versa. A key factor in our success is a smooth contact model that combines informative gradients with physical accuracy, ensuring effective transfer of learned behaviors. To the best of our knowledge, this is the first time a real quadrupedal robot is able to locomote after training exclusively in a differentiable simulation. △ Less

Submitted 4 November, 2024; originally announced November 2024.

Comments: Presented at the CoRL 2024 Workshop 'Differentiable Optimization Everywhere'

ACM Class: I.2.9; I.6.5