SimR: Simulation in Robotics

ICRA 2026 Tutorial

“What simulator should I use?” is a common question among robotics researchers and practitioners when starting a new project or exploring a new domain (we now rarely ask “Should I use a simulator?”). This is a particularly challenging question for newcomers to our field. The rise of data-driven methods in robotics has led to surge in simulator development, supported by research labs, industry, and open-source communities. These systems vary widely in their capabilities, strengths, and pain-points. Even with recent advances in simulation and training materials, finding the right level of abstraction for a given problem remains a mixture of art and science. Moreover, finding the right simulator must necessarily consider the post-simulation goals; how will any simulation results transfer to the real world? In this tutorial, we will survey modern simulation techniques and tools for robotics research. We will then train a simple reinforcement learning agent in simulation and “deploy” it to another (simulation) environment. SimR is a tutorial on the fundamentals of simulation for robotics research.