Online vs. Offline Adaptive Domain Randomization Benchmark
This provides a benchmark for researchers and practitioners in robotics and reinforcement learning to evaluate adaptive domain randomization methods, though it is incremental as it compares existing methods rather than introducing new ones.
The authors tackled the lack of thorough comparison between adaptive domain randomization methods for transferring reinforcement learning policies from simulation to reality, by creating an open benchmark for offline and online methods. They found that online methods are limited by policy quality during iteration, while offline methods can fail with open-loop commands in simulation.
Physics simulators have shown great promise for conveniently learning reinforcement learning policies in safe, unconstrained environments. However, transferring the acquired knowledge to the real world can be challenging due to the reality gap. To this end, several methods have been recently proposed to automatically tune simulator parameters with posterior distributions given real data, for use with domain randomization at training time. These approaches have been shown to work for various robotic tasks under different settings and assumptions. Nevertheless, existing literature lacks a thorough comparison of existing adaptive domain randomization methods with respect to transfer performance and real-data efficiency. In this work, we present an open benchmark for both offline and online methods (SimOpt, BayRn, DROID, DROPO), to shed light on which are most suitable for each setting and task at hand. We found that online methods are limited by the quality of the currently learned policy for the next iteration, while offline methods may sometimes fail when replaying trajectories in simulation with open-loop commands. The code used will be released at https://github.com/gabrieletiboni/adr-benchmark.