FARO: Feasibility-Aware Robot Motion Optimization

2026-07-20

AI Summary

This work addresses the challenge of rapidly generating novel, feasible, and dynamically consistent loco-manipulation behaviors for humanoid robots operating in unknown environments. To this end, the authors propose a nested kino-dynamic planning framework that, for the first time, leverages large language models (LLMs) to sample contact sequences. This approach integrates feasibility-guided tree search with a reinforcement learning (RL) controller to enable efficient trajectory generation and validation. The resulting trajectories exhibit high dynamic consistency and can be stably tracked by the RL controller. Evaluated in complex, high-dimensional scenarios, the method significantly improves both the efficiency and quality of motion planning and has been successfully deployed on a real-world humanoid robot system.

Abstract

Fast planning of novel behaviors in unseen scenarios remains a fundamental challenge in robotics. The high-dimensional, hybrid, and underactuated nature of humanoid loco-manipulation continues to hinder the realization of this goal. In this paper, we address this challenge by proposing a nested kino-dynamic framework for rapid feasibility checking and dynamically consistent trajectory generation given a candidate contact sequence. By integrating this module with a feasibility-guided tree search and a Large Language Model (LLM)-based contact plan sampling strategy, we demonstrate that the proposed framework can substantially improve the search process. Furthermore, we show that the generated trajectories can be tracked using a reinforcement learning (RL)-based controller and show that the resulting trajectories are of sufficiently high quality for execution in real-world loco-manipulation scenarios. A supplementary video is available at: https://youtu.be/R6qCHoCormQ.

Problem

• robot motion planning
• loco-manipulation
• feasibility checking
• humanoid robotics
• trajectory generation

Innovation

• kino-dynamic planning
• feasibility-aware optimization
• humanoid loco-manipulation
• LLM-based contact planning
• reinforcement learning control