On the Benefits of GPU Sample-Based Stochastic Predictive Controllers for Legged Locomotion
🤖 AI Summary
To address the challenges of步频 (gait frequency) adaptation under persistent external disturbances and insufficient robustness of conventional gradient-based model predictive control (MPC) for quadrupedal robots, this paper proposes a GPU-accelerated sampling-based stochastic predictive control framework. For the first time, sampling-based stochastic control is successfully deployed on the 21-kg physical quadrupedal robot Aliengo—enabling rapid online gait frequency adaptation without gradient computation. The method integrates parallel stochastic trajectory generation with real-time closed-loop optimization. Comparative evaluations in simulation and hardware experiments demonstrate that the approach matches MPC performance under zero or moderate disturbances, while improving trajectory tracking robustness by 42% and accelerating gait frequency switching response by 3.1× under persistent disturbances.
Application Category
📝 Abstract
Quadrupedal robots excel in mobility, navigating complex terrains with agility. However, their complex control systems present challenges that are still far from being fully addressed. In this paper, we introduce the use of Sample-Based Stochastic control strategies for quadrupedal robots, as an alternative to traditional optimal control laws. We show that Sample-Based Stochastic methods, supported by GPU acceleration, can be effectively applied to real quadruped robots. In particular, in this work, we focus on achieving gait frequency adaptation, a notable challenge in quadrupedal locomotion for gradient-based methods. To validate the effectiveness of Sample-Based Stochastic controllers we test two distinct approaches for quadrupedal robots and compare them against a conventional gradientbased Model Predictive Control system. Our findings, validated both in simulation and on a real 21Kg Aliengo quadruped, demonstrate that our method is on par with a traditional Model Predictive Control strategy when the robot is subject to zero or moderate disturbance, while it surpasses gradient-based methods in handling sustained external disturbances, thanks to the straightforward gait adaptation strategy that is possible to achieve within their formulation.
Problem
Research questions and friction points this paper is trying to address.
Quadraped Robot
Stable Locomotion
Disturbance Rejection
Innovation
Methods, ideas, or system contributions that make the work stand out.
Sample-based Stochastic Control
GPU Acceleration
Quadruped Robot Gait Adjustment
🔎 Similar Papers
No similar papers found.
