Reduced-Order Model Guided Contact-Implicit Model Predictive Control for Humanoid Locomotion

2025-02-21

AI Summary

Humanoid robots face significant challenges in achieving robust, real-time gait control over complex terrain due to the high-dimensional, nonlinear, and hybrid nature of their dynamics modeling and optimization. This paper proposes HLIP-CI-MPC, a hierarchical cooperative architecture: an upper layer employs the Hybrid Linear Inverted Pendulum (HLIP) to generate robust nominal gaits, while a lower layer utilizes Contact-Implicit Model Predictive Control (CI-MPC) to close the loop on full-body motion planning and dynamic contact-sequence optimization. Crucially, this work establishes the first tightly coupled integration of HLIP-based gait generation with CI-MPC, enabling automatic adaptation—without manual tuning—to contact-mode transitions, external disturbances, and model/state uncertainties. Evaluated on a 24-DOF Achilles simulation platform, the system runs online at 50 Hz and demonstrates stable locomotion across rough terrain, rapid recovery from perturbations, obstacle interaction, and resilience under multi-source uncertainties—achieving a favorable trade-off between computational efficiency and full-body dynamical fidelity.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation

Abstract

Humanoid robots have great potential for real-world applications due to their ability to operate in environments built for humans, but their deployment is hindered by the challenge of controlling their underlying high-dimensional nonlinear hybrid dynamics. While reduced-order models like the Hybrid Linear Inverted Pendulum (HLIP) are simple and computationally efficient, they lose whole-body expressiveness. Meanwhile, recent advances in Contact-Implicit Model Predictive Control (CI-MPC) enable robots to plan through multiple hybrid contact modes, but remain vulnerable to local minima and require significant tuning. We propose a control framework that combines the strengths of HLIP and CI-MPC. The reduced-order model generates a nominal gait, while CI-MPC manages the whole-body dynamics and modifies the contact schedule as needed. We demonstrate the effectiveness of this approach in simulation with a novel 24 degree-of-freedom humanoid robot: Achilles. Our proposed framework achieves rough terrain walking, disturbance recovery, robustness under model and state uncertainty, and allows the robot to interact with obstacles in the environment, all while running online in real-time at 50 Hz.

Problem

• Control high-dimensional nonlinear hybrid dynamics in humanoid robots.
• Combine HLIP for gait generation with CI-MPC for dynamics management.
• Achieve robust real-time locomotion and interaction in complex environments.

Innovation

• Combines HLIP and CI-MPC
• Manages whole-body dynamics
• Runs online at 50 Hz