TransCurriculum: Multi-Dimensional Curriculum Learning for Fast & Stable Locomotion

2026-03-14

AI Summary

This work addresses the challenges of instability in high-speed quadrupedal locomotion and significant sim-to-real transfer gaps, which are exacerbated by the inability of conventional one-dimensional curriculum learning to jointly optimize multidimensional task complexities. To overcome this, the authors propose a novel Transformer-based multidimensional curriculum learning approach that, for the first time, integrates a Transformer into the curriculum generation mechanism to co-adaptively modulate command velocity, terrain difficulty, and domain randomization parameters. By predicting performance from historical reward sequences, the method dynamically expands the curriculum space. Policies trained in Isaac Gym achieve zero-shot transfer to the Unitree Go1 platform, reaching a simulated top speed of 6.3 m/s and a real-world speed of 4.1 m/s—an 18.8% improvement—while reducing sim-to-real performance loss from 27% to 18% and maintaining task success rates between 80% and 90%.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Search and Optimization: Learning to Search
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

High-speed legged locomotion struggles with stability and transfer losses at higher command velocities during deployment. One reason is that most curricula vary difficulty along single axis, for example increase the range of command velocities, terrain difficulty, or domain parameters (e.g. friction or payload mass) using either fixed update rule or instantaneous rewards while ignoring how the history of robot training has evolved. We propose TransCurriculum, a transformer-based multi-dimensional curriculum learning approach for agile quadrupedal locomotion. TransCurriculum adapts to 3 axes, velocity command targets, terrain difficulty, and domain randomization parameters (friction and payload mass). Rather than feeding task reward history directly into the low-level control policy, our formulation exploits it at the curriculum level. A transformer-based teacher retrieves the sequence of rewards and uses it to predict future rewards, success rate, and learning progress to guide expansion of this multidimensional curriculum towards high performing task bins. Finally we validate our approach on the Unitree Go1 robot in simulation (Isaac Gym) and deploy it zero-shot on Go1 hardware. Our TransCurriculum policy achieves a maximum velocity of 6.3 m/s in simulation and outperforms prior curriculum baselines. We tested our TransCurriculum trained policy on terrains (carpets, slopes, tiles, concrete), achieving a forward velocity of 4.1 m/s on carpet surpassing the fastest curriculum methods by 18.8% and achieves maximum zero-shot value among all tested methods. Our multi-dimensional curriculum also reduces the transfer loss to 18% from 27% for command only curriculum, demonstrating the benefits of joint training over velocity, terrain and domain randomization dimension while keeping the task success rate of 80-90% on rigid indoor and outdoor surfaces.

Problem

• legged locomotion
• curriculum learning
• transfer loss
• stability
• multi-dimensional difficulty

Innovation

• TransCurriculum
• multi-dimensional curriculum learning
• transformer-based teacher
• zero-shot transfer
• legged locomotion