Trajectory-Level Automatic Curriculum Learning for Legged Locomotion on Unstructured Terrain

2026-08-17

AI Summary

This study addresses the challenges of poor curriculum generalization and the absence of explicit difficulty scheduling in legged locomotion training on unstructured terrain. We propose a trajectory-level automatic curriculum learning framework that generates tasks directly from terrain maps, achieving adaptive alignment between the curriculum and policy through closed-loop iteration involving a difficulty evaluator and an adaptive sampler. Experimental results demonstrate that this framework improves success rates by 56.3% compared to non-curriculum baselines and outperforms hand-crafted curricula by 18.5% on challenging tasks. Furthermore, it enhances multi-directional traversal capability by 39.74%, effectively overcoming critical difficulties in policy training for complex terrains.

Abstract

Training locomotion policies for complex unstructured terrain requires a curriculum to avoid early exploration failures. However, since unstructured terrain lacks explicit difficulty ordering for curriculum design, existing methods resort to heuristic curricula over parameterized terrains. This abstraction limits generalization, as policies can overadapt to near-fixed perceptual patterns. To address this, we propose \textbf{\ourname{}}, an \textbf{T}rajectory-level \textbf{A}utomatic \textbf{C}urriculum \textbf{L}earning framework that generates training tasks directly from unstructured terrain maps. At each curriculum update, the evaluator learns a difficulty function for the current policy that maps a given trajectory task to a difficulty score. The sampler then proposes new trajectories guided by the learned evaluator as the curriculum for the next policy update. This forms a closed loop in which the curriculum is iteratively matched to the evolving policy. Quantitative and qualitative experiments show that \ourname{} continuously provides effective curricula on unstructured terrain, improving trajectory success rate by \(56.3\%\) over direct training without curriculum. Compared with handcrafted curriculum learning, our method improves success rate by \(18.5\%\) on the hardest terrain tasks and by up to \(39.74\%\) when evaluating traversal from diverse approach directions on the same obstacle type.

Problem

• Legged Locomotion
• Unstructured Terrain
• Curriculum Learning
• Generalization

Innovation

• Automatic Curriculum Learning
• Legged Locomotion
• Unstructured Terrain
• Trajectory-Level Training
• Difficulty Evaluator