Search papers, labs, and topics across Lattice.
This paper introduces DynaWM, a dynamics-aware representation learning framework designed to improve the ability of bipedal-wheeled robots to traverse complex stair terrains. By incorporating a world model as a regularizer and a momentum target encoder for stable knowledge transfer, DynaWM enhances terrain encoding and preserves comprehensive geometry, leading to better adaptability and smoother locomotion. Experimental results show that DynaWM significantly outperforms existing methods in both simulation and real-world applications, enabling robots to navigate diverse continuous stairs effectively.
Bipedal-wheeled robots can now traverse complex stair terrains with unprecedented smoothness and adaptability, thanks to a novel dynamics-aware framework.
Recent advances in control have enabled bipedal-wheeled robots to traverse slopes and single-step obstacles, yet long staircase traversal remains challenging as current teacher-student frameworks suffer from weakened dynamics-aware representations and incomplete terrain geometry encoding. To bridge this gap, we propose DynaWM, a dynamics-aware representation learning framework. To enhance terrain encoding capability and enable transparent assessment, we introduce a world model as a regularizer to enforce forward-dynamics awareness, preserving comprehensive terrain geometry while facilitating hierarchical encoding visualization. To stabilize knowledge transfer, we employ a momentum target encoder to provide consistent distillation targets, preventing dimensional collapse from non-stationary teacher updates. Evaluation of the learned representations through Principal Component Analysis (PCA) visualization and quantitative metrics reveals that our encoder hierarchically captures terrain geometry with higher terrain encoding capability, leading to enhanced terrain adaptability and motion smoothness. Experimental results in simulation and real hardware demonstrate that our method achieves superior terrain adaptability and motion smoothness, enabling bipedal-wheeled robots to overcome diverse continuous stairs, as shown in Fig. 1.
