
Learning Domain Randomization Distributions for Transfer of Locomotion Policies
Domain randomization (DR) is a successful technique for learning robust ...
read it

AutoTuned SimtoReal Transfer
Policies trained in simulation often fail when transferred to the real w...
read it

Assessing Transferability from Simulation to Reality for Reinforcement Learning
Learning robot control policies from physics simulations is of great int...
read it

Policy Transfer via Kinematic Domain Randomization and Adaptation
Transferring reinforcement learning policies trained in physics simulati...
read it

Learning Transferable Policies for Monocular Reactive MAV Control
The ability to transfer knowledge gained in previous tasks into new cont...
read it

BayesSim: adaptive domain randomization via probabilistic inference for robotics simulators
We introduce BayesSim, a framework for robotics simulations allowing a f...
read it

Stochastic Grounded Action Transformation for Robot Learning in Simulation
Robot control policies learned in simulation do not often transfer well ...
read it
Bayesian Domain Randomization for SimtoReal Transfer
When learning policies for robot control, the realworld data required is typically prohibitively expensive to acquire, so learning in simulation is a popular strategy. Unfortunately, such polices are often not transferable to the real world due to a mismatch between the simulation and reality, called 'reality gap'. Domain randomization methods tackle this problem by randomizing the physics simulator (source domain) according to a distribution over domain parameters during training in order to obtain more robust policies that are able to overcome the reality gap. Most domain randomization approaches sample the domain parameters from a fixed distribution. This solution is suboptimal in the context of simtoreal transferability, since it yields policies that have been trained without explicitly optimizing for the reward on the real system (target domain). Additionally, a fixed distribution assumes there is prior knowledge about the uncertainty over the domain parameters. Thus, we propose Bayesian Domain Randomization (BayRn), a black box simtoreal algorithm that solves tasks efficiently by adapting the domain parameter distribution during learning by sampling the realworld target domain. BayRn utilizes Bayesian optimization to search the space of source domain distribution parameters which produce a policy that maximizes the realword objective, allowing for adaptive distributions during policy optimization. We experimentally validate the proposed approach by comparing against two baseline methods on a nonlinear underactuated swingup task. Our results show that BayRn is capable to perform direct simtoreal transfer, while significantly reducing the required prior knowledge.
READ FULL TEXT
Comments
There are no comments yet.