ADEPT: Adaptive Diffusion Environment for Policy Transfer Sim-to-Real

TL;DR

ADEPT introduces an adaptive diffusion-based environment generation method for zero-shot sim-to-real policy transfer, enhancing training diversity and robustness in complex, unstructured environments.

Contribution

It leverages Denoising Diffusion Probabilistic Models to dynamically generate diverse environments tailored to policy performance, advancing environment generation for reinforcement learning.

Findings

ADEPT outperforms procedural and natural environments in navigation tasks.

The method improves policy robustness and transferability.

A new multi-layer map representation accelerates environment benchmarking.

Abstract

Model-free reinforcement learning has emerged as a powerful method for developing robust robot control policies capable of navigating through complex and unstructured environments. The effectiveness of these methods hinges on two essential elements: (1) the use of massively parallel physics simulations to expedite policy training, and (2) an environment generator tasked with crafting sufficiently challenging yet attainable environments to facilitate continuous policy improvement. Existing methods of outdoor environment generation often rely on heuristics constrained by a set of parameters, limiting the diversity and realism. In this work, we introduce ADEPT, a novel \textbf{A}daptive \textbf{D}iffusion \textbf{E}nvironment for \textbf{P}olicy \textbf{T}ransfer in the zero-shot sim-to-real fashion that leverages Denoising Diffusion Probabilistic Models to dynamically expand existing training environments by adding more diverse and complex environments adaptive to the current policy. ADEPT guides the diffusion model's generation process through initial noise optimization, blending noise-corrupted environments from existing training environments weighted by the policy's performance in each corresponding environment. By manipulating the noise corruption level, ADEPT seamlessly transitions between generating similar environments for policy fine-tuning and novel ones to expand training diversity. To benchmark ADEPT in off-road navigation, we propose a fast and effective multi-layer map representation for wild environment generation. Our experiments show that the policy trained by ADEPT outperforms both procedural generated and natural environments, along with popular navigation methods.