Efficient Domain-Adaptive Policy Learning via Kernel Representation with Application to Quadrotor Control under Non-Stationary Disturbances

2026-06-11

AI Summary

This work addresses the challenge of efficiently adapting quadrotor control under nonstationary disturbances such as gusts, payload variations, and ground effects. To this end, we propose a kernel-based domain adaptive control method that leverages random Fourier features to generate diverse disturbances during an offline phase, integrated with differentiable simulation and analytical gradient-based optimization. During online operation, kernel parameters are updated in real time via least-squares estimation, and the kernel bandwidth is adaptively tuned to balance modeling expressiveness with computational efficiency. Requiring only 50 seconds of offline training, the approach enables rapid online adaptation and significantly improves trajectory tracking performance under complex disturbances in both high-fidelity simulation and on the Crazyflie hardware platform, effectively narrowing the sim-to-real gap.

Abstract

We present an algorithm for efficient domain-adaptive policy learning via kernel representations. Learning domain-adaptive policies is challenging since it requires an environment representation that is both sufficiently expressive to model complex sim-to-real gaps during offline training, and computationally efficient enough to support rapid online adaptation during deployment. For instance, a quadrotor may encounter time-varying, non-stationary disturbances, such as sudden gusts of wind, payload shifts, or transitions between distinct flight regimes with and without ground effects. To address these challenges, we model unknown disturbances using a differentiable kernel approximation based on random Fourier features. During the offline training phase, we randomly sample kernel coefficients and bandwidth parameters to generate a rich diversity of disturbance profiles. We then optimize the control policy via differentiable simulation with analytical gradients, a process that takes only 50 seconds of training time on an RTX 4090 GPU. During hardware deployment, the policy adapts to non-stationary environments in real time by updating both the kernel coefficients and bandwidth through online least-squares estimation. We evaluate our method on quadrotor trajectory tracking tasks across high-fidelity numerical simulations and hardware experiments using Crazyflie, subjected to various disturbances, including complex aerodynamic effects, wind, ground effects, and payload fluctuations.

Problem

• domain-adaptive policy learning
• non-stationary disturbances
• quadrotor control
• sim-to-real gap
• online adaptation

Innovation

• domain-adaptive policy learning
• kernel representation
• random Fourier features
• differentiable simulation
• online adaptation