# Reference Free Platform Adaptive Locomotion for Quadrupedal Robots using a Dynamics Conditioned Policy

Source: [Reference Free Platform Adaptive Locomotion for Quadrupedal Robots using a Dynamics Conditioned Policy](<https://www.sota2.com/research/paper/68fa79ec1e209131cea7fff2>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T06:56:50.936Z

Authors: David Rytz, Suyoung Choi, Wanming Yu, Wolfgang Merkt, Jemin Hwangbo, Ioannis Havoutis (1) \_\_INSTITUTION\_6\_\_ Dynamic Robot Systems, Oxford Robotics Institute, University of Oxford, (2) RaiLab, Department of Mechanical Engineering, KAIST)

Published: 2025-05-21T21:48:51.000Z

## Abstract

This article presents Platform Adaptive Locomotion (PAL), a unified control method for quadrupedal robots with different morphologies and dynamics. We leverage deep reinforcement learning to train a single locomotion policy on procedurally generated robots. The policy maps proprioceptive robot state information and base velocity commands into desired joint actuation targets, which are conditioned using a latent embedding of the temporally local system dynamics. We explore two conditioning strategies - one using a GRU-based dynamics encoder and another using a morphology-based property estimator - and show that morphology-aware conditioning outperforms temporal dynamics encoding regarding velocity task tracking for our hardware test on ANYmal C. Our results demonstrate that both approaches achieve robust zero-shot transfer across multiple unseen simulated quadrupeds. Furthermore, we demonstrate the need for careful robot reference modelling during training: exposing the policy to a diverse set of robot morphologies and dynamics leads to improved generalization, reducing the velocity tracking error by up to 30% compared to the baseline method. Despite PAL not surpassing the best-performing reference-free controller in all cases, our analysis uncovers critical design choices and informs improvements to the state of the art.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Velocity tracking on ANYmal-D Real-world (physical hardware trials)](<https://www.sota2.com/research/sota/velocity-tracking-on-anymal-d-real-world-physical-hardware-trials.md>): Velocity tracking; ANYmal-D Real-world (physical hardware trials); eyaw (rad/s): 0.2199.

- [Velocity command tracking and estimation on ANYmal Hardware Flat Terrain (test)](<https://www.sota2.com/research/sota/velocity-command-tracking-and-estimation-on-anymal-hardware-flat-terrain-test.md>): Velocity command tracking and estimation; ANYmal Hardware Flat Terrain (test); RMSE vx: 0.0987.

- [Velocity command tracking and estimation on ANYmal Hardware (Rough Terrain) (test)](<https://www.sota2.com/research/sota/velocity-command-tracking-and-estimation-on-anymal-hardware-rough-terrain-test.md>): Velocity command tracking and estimation; ANYmal Hardware (Rough Terrain) (test); RMSE Linear Velocity (x): 0.0983.

- [Base velocity estimation on ANYmal Hardware (walking test)](<https://www.sota2.com/research/sota/base-velocity-estimation-on-anymal-hardware-walking-test.md>): Base velocity estimation; ANYmal Hardware (walking test); RMSE (X): 0.103.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2505.16042>)

[PDF](<https://arxiv.org/pdf/2505.16042.pdf>)
