Rethinking Sim2Real: Lower Fidelity Simulation Leads to Higher Sim2Real Transfer in Navigation

Joanne Truong, Max Rudolph, Naoki Harrison Yokoyama, Sonia Chernova, Dhruv Batra, Akshara Rai

Keywords: Sim2Real, Deep Reinforcement Learning, Visual-Based Navigation

TL;DR: Sim2real transfer of robots may be improved not by increasing but by decreasing simulation fidelity; we should instead prioritize simulation speed for tasks that can be represented with abstract action spaces.

Abstract: If we want to train robots in simulation before deploying them in reality, it seems natural and almost self-evident to presume that reducing the sim2real gap involves creating simulators of increasing fidelity (since reality is what it is). We challenge this assumption and present a contrary hypothesis -- sim2real transfer of robots may be improved with lower (not higher) fidelity simulation. We conduct a systematic large-scale evaluation of this hypothesis on the problem of visual navigation -- in the real world, and on 2 different simulators (Habitat and iGibson) using 3 different robots (A1, AlienGo, Spot). Our results show that, contrary to expectation, adding fidelity does not help with learning; performance is poor due to slow simulation speed (preventing large-scale learning) and overfitting to inaccuracies in simulation physics. Instead, building simple models of the robot motion using real-world data can improve learning and generalization.

Student First Author: yes

Website: http://www.joannetruong.com/projects/kin2dyn.html

Code: https://github.com/joannetruong/habitat-lab/tree/kin2dyn