Search papers, labs, and topics across Lattice.
This paper introduces HODAgent, a System-2 embodied agent designed for humanoid robots that enhances their ability to interact responsively in service settings by integrating intent recognition, task planning, execution, and outcome verification within a semi-duplex architecture. The agent's hierarchical memory and shared interface facilitate coherent interactions and task management, allowing it to adapt to new requests while in motion. Experimental results demonstrate that HODAgent significantly outperforms baseline models, achieving high success rates in both simulated environments and on physical robots, indicating its effectiveness for real-world applications in humanoid robotics.
HODAgent's unified architecture enables humanoid robots to adaptively manage service tasks in real-time, achieving up to 92% success rates on physical platforms.
We propose HODAgent, a System-2 embodied agent for humanoid robots in service settings, addressing situated intent, responsive execution, task revision, and outcome verification. Its semi-duplex architecture integrates an Env-Interactor, Planner, Executor, and hierarchical Memory to maintain coherent interaction, planning, and task state during service episodes. This allows handling new requests during motion, retaining progress, revising actions, and grounding closure in execution outcomes. A shared interface connects simulation and physical robots (Unitree G1), isolating platform-specific control. In an interactive simulation with 164 cases, HODAgent achieves 84.8% and 91.5% Joint Success under two VLM backbones, outperforming baselines by 9.8 and 18.9 points. On physical robots, pass rates are 92% (atomic), 72% (composite), and 63.3% (complete tasks). On multiple embodied benchmarks, it improves over baselines by 0.7-9.0 points. Results show a unified System-2 agent enables adaptive humanoid service across simulation and reality.
Tsinghua AI