Dynamics Aware Quadrupedal Locomotion via Intrinsic Dynamics Head

TL;DR

This paper introduces a training framework for quadrupedal robots that incorporates an Intrinsic Dynamics Head to improve stability, efficiency, and transferability of locomotion policies from simulation to real robots.

Contribution

It proposes a novel method to learn and tune physical dynamics within the control policy, enhancing quadrupedal locomotion performance and sim-to-real transfer.

Findings

Achieved 16.8% torque efficiency improvement in real robots.

Demonstrated 18.6% increase in action rate.

Reduced mechanical power consumption by 12.8%.

Abstract

Quadrupedal locomotion plays a critical role in enabling agile, versatile movement across complex terrains. Understanding and estimating the underlying physical dynamics are essential for achieving efficient and stable quadrupedal locomotion. We propose a novel training framework for quadrupedal locomotion that enables the Control Policy to understand and reason about physical dynamics. In simulation, we concurrently train an Intrinsic Dynamics (ID) Head that learns state-to-torque dynamics alongside the Control Policy, and we define a dynamics reward enabled by the ID Head that encourages the Policy toward more predictable dynamical behavior. We also provide a mechanism to tune the learned dynamics in the resulting Policy by controlling the training coefficients of the ID Head. Our simulation experiments show that this mechanism drives convergence to better optima across a wide range of standard quadrupedal locomotion rewards, yielding more efficient and smoother policies. Our real-robot experiments demonstrate sim-to-real transfer of these improvements, with significant gains in torque efficiency (16.8%), action rate (18.6%), and mechanical power (12.8%), while improving safe torque occupancy by 6.4%.