Learning Hybrid Behavior Planning for Autonomous Loco-manipulation

Enabling humanoid robots to autonomously perform hybrid locomotion and manipulation in diverse environments can be beneficial for long-horizon tasks such as factory material handling, household chores, and work assistance. This requires extensive exploitation of intrinsic motion capabilities, extraction of affordances from rich environmental information, and planning of physical interaction behaviors. Despite recent progress has demonstrated impressive humanoid whole-body control abilities, they struggle to achieve versatility and adaptability for new tasks. In this work, we propose HYPERmotion, a framework that learns, selects and plans behaviors based on tasks in different scenarios. We combine reinforcement learning with whole-body model predictive control to generate motion for 37 actuated joints and create a motion library to store the learned skills. We further apply the online planning and reasoning features of the large language models (LLMs) to complex loco-manipulation tasks, constructing a hierarchical task graph that comprises a series of primitive behaviors to bridge lower-level execution with higher-level planning. By leveraging the interaction of distilled spatial geometry and 2D observation with a visual language model (VLM) to ground knowledge into a robotic morphology selector to choose appropriate actions in single- or dual-arm, legged or wheeled locomotion. Experiments in both simulated and real-world settings show that learned whole-body skills can efficiently adapt to new tasks, demonstrating high autonomy from free-text instructions in unstructured scenes.