AnyCar to Anywhere: Learning Universal Dynamics Model for Agile and Adaptive Mobility

TL;DR

AnyCar is a transformer-based universal dynamics model that enables agile, adaptive control of various wheeled robots, outperforming specialist models through robust training, real-world fine-tuning, and few-shot and zero-shot generalization.

Contribution

We introduce AnyCar, a novel generalist dynamics model for agile wheeled robot control, capable of adapting to diverse vehicles and terrains with minimal tuning.

Findings

Outperforms specialist models by up to 54% in real-world tests.

Enables few-shot and zero-shot generalization across diverse vehicles.

Supports robust adaptation to different terrains and large state estimation errors.

Abstract

Recent works in the robot learning community have successfully introduced generalist models capable of controlling various robot embodiments across a wide range of tasks, such as navigation and locomotion. However, achieving agile control, which pushes the limits of robotic performance, still relies on specialist models that require extensive parameter tuning. To leverage generalist-model adaptability and flexibility while achieving specialist-level agility, we propose AnyCar, a transformer-based generalist dynamics model designed for agile control of various wheeled robots. To collect training data, we unify multiple simulators and leverage different physics backends to simulate vehicles with diverse sizes, scales, and physical properties across various terrains. With robust training and real-world fine-tuning, our model enables precise adaptation to different vehicles, even in the wild and under large state estimation errors. In real-world experiments, AnyCar shows both few-shot and zero-shot generalization across a wide range of vehicles and environments, where our model, combined with a sampling-based MPC, outperforms specialist models by up to 54%. These results represent a key step toward building a foundation model for agile wheeled robot control. We will also open-source our framework to support further research.