keyword
dynamics randomization
Dynamics randomization is a machine learning and robotics technique where the physical and mechanical parameters of a simulated environment are randomly perturbed across training iterations to help a control policy generalize to the physical world. During the simulation process, variables such as object mass, surface friction, damping, actuator latency, joint limits, and sensor noise are varied within defined ranges. Exposing the learning algorithm to a diverse spectrum of physical dynamics prevents it from overfitting to the idealized or inaccurate physics of a specific simulation engine. Consequently, the resulting control policy develops robust or adaptive behaviors that can accommodate discrepancies between simulated models and real hardware, effectively bridging the reality gap during simulation-to-real deployment.
2 items
MetaMorph: Learning Universal Controllers with Transformers
Agrim Gupta, Linxi Fan, Surya Ganguli, Li Fei-Fei
Why you should read this
Develops MetaMorph, a Transformer architecture that conditions control policies on robot morphology to enable universal motor control with zero-shot generalization across unseen modular robot designs and tasks.
Multiple domains like vision, natural language, and audio are witnessing tremendous progress by leveraging Transformers for large scale pre-training followed by task specific fine tuning. In contrast, in robotics we primarily train a single robot for a single task. However, modular robot systems now allow for the flexible combination of general-purpose building blocks into task optimized morphologies. However, given the exponentially large number of possible robot morphologies, training a controller for each new design is impractical. In this work, we propose MetaMorph, a Transformer based approach to learn a universal controller over a modular robot design space. MetaMorph is based on the insight that robot morphology is just another modality on which we can condition the output of a Transformer. Through extensive experiments we demonstrate that large scale pre-training on a variety of robot morphologies results in policies with combinatorial generalization capabilities, including zero shot generalization to unseen robot morphologies. We further demonstrate that our pre-trained policy can be used for sample-efficient transfer to completely new robot morphologies and tasks.
Added
2026-09-26
Sim-to-Real Transfer of Robotic Control with Dynamics Randomization
Xue Bin Peng, Marcin Andrychowicz, Wojciech Zaremba, Pieter Abbeel
Why you should read this
Demonstrates that randomizing physical dynamics during simulation training enables robotic control policies to transfer directly to physical hardware without requiring real-world data or exact model calibration.
Simulations are attractive environments for training agents as they provide an abundant source of data and alleviate certain safety concerns during the training process. But the behaviours developed by agents in simulation are often specific to the characteristics of the simulator. Due to modeling error, strategies that are successful in simulation may not transfer to their real world counterparts. In this paper, we demonstrate a simple method to bridge this "reality gap". By randomizing the dynamics of the simulator during training, we are able to develop policies that are capable of adapting to very different dynamics, including ones that differ significantly from the dynamics on which the policies were trained. This adaptivity enables the policies to generalize to the dynamics of the real world without any training on the physical system. Our approach is demonstrated on an object pushing task using a robotic arm. Despite being trained exclusively in simulation, our policies are able to maintain a similar level of performance when deployed on a real robot, reliably moving an object to a desired location from random initial configurations. We explore the impact of various design decisions and show that the resulting policies are robust to significant calibration error.
Added
2026-09-18