Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experience
This addresses the sim-to-real gap for robotics, enabling more efficient policy transfer without manual tuning, though it is incremental as it builds on existing simulation randomization methods.
The paper tackles the problem of transferring policies from simulation to the real world by adapting simulation randomization using real-world experience, resulting in policies that reliably transfer to different robots in tasks like swing-peg-in-hole and opening a cabinet drawer.
We consider the problem of transferring policies to the real world by training on a distribution of simulated scenarios. Rather than manually tuning the randomization of simulations, we adapt the simulation parameter distribution using a few real world roll-outs interleaved with policy training. In doing so, we are able to change the distribution of simulations to improve the policy transfer by matching the policy behavior in simulation and the real world. We show that policies trained with our method are able to reliably transfer to different robots in two real world tasks: swing-peg-in-hole and opening a cabinet drawer. The video of our experiments can be found at https://sites.google.com/view/simopt