2024 Conference on Robot Learning Workshop - CoRL Workshop, Munich (Germany). 09 November 2024
Summary:
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 viceversa. 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.
Spanish layman's summary:
Este trabajo muestra que las políticas de locomoción entrenadas con gradientes analíticos en un simulador diferenciable pueden transferirse a robots reales, gracias a un modelo de contacto suave que equilibra la información de gradiente y la precisión física, logrando una locomoción cuadrúpeda de simulación a real.
English layman's summary:
This work shows that locomotion policies trained with analytic gradients in a differentiable simulator can transfer to real robots, enabled by a smooth contact model that balances gradient informativeness and physical accuracy—achieving sim-to-real quadruped locomotion.
Publication date: 2024-11-09.
Citation:
J. Bagajo, C. Schwarke, V. Klemm, I. Georgiev, J.P. Sleiman, J. Tordesillas Torres, DiffSim2Real: Deploying Quadrupedal Locomotion Policies Purely Trained in Differentiable Simulation, 2024 Conference on Robot Learning Workshop - CoRL Workshop, Munich (Germany). 09 November 2024.
Download this paper