Overview

We present Sym2Real, a fully data-driven framework that provides a principled way to train low-level adaptive controllers in a highly data-efficient manner. Using only about 10 trajectories, we achieve robust control of both a quadrotor and a racecar in the real world, without expert knowledge or simulation tuning. Our approach achieves this data efficiency by bringing symbolic regression to real-world robotics by addressing key challenges that prevent its direct application, including noise sensitivity and model degradation that lead to unsafe control. Our key observation is that the underlying physics is often shared for a system no matter the internal or external changes. Hence, we strategically combine low-fidelity simulation data with targeted real-world residual learning. Through experimental validation on quadrotor and racecar platforms, we demonstrate consistent data-efficient adaptation across 6 out-of-distribution sim2sim scenarios and successful sim2real transfer across 5 real-world conditions.

Video (Click to YouTube)

Paper

Check out our paper linked here.

Codebase and Supplementary Materials

Check out our codebase at https://github.com/generalroboticslab/sym2real. Discovered symbolic equations of the drone and the car can be found here and here.

Citation

@misc{lee2025sym2realsymbolicdynamicsresidual,
      title={Sym2Real: Symbolic Dynamics with Residual Learning for Data-Efficient Adaptive Control}, 
      author={Easop Lee and Samuel A. Moore and Boyuan Chen},
      year={2025},
      eprint={2509.15412},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2509.15412}, 
}  

Acknowledgment

This work was supported by ARO under award W911NF2410405 and DARPA TIAMAT program under award HR00112490419.

Contact

If you have any questions, please feel free to contact Easop Lee.