Ankle, hip and stepping strategies for humanoid balance recovery with a single Model Predictive Control scheme

Abstract

While humans are highly efficient in dealing with balance perturbations, current biped humanoid robots are far from showing similar skills. This is mainly due to the limited capacity of current robot controllers to deal with the inherently complex dynamics of biped robots. Though Model Predictive Control schemes have shown improved robustness to perturbations, they still suffer from a few shortcomings such as not considering the upper body inertial effects or nonoptimal step durations. We propose here a Model Predictive Control scheme that specifically addresses these shortcomings and generates human-like responses to perturbations, involving appropriate combinations of ankle, hip and stepping strategies, with automatically adjusted step durations. The emphasis of this paper is on modeling and analyzing the effects of different cost functions and coefficients on the behavior of the controller while leaving real-time implementations and experiments for later work.

Compliant attitude control and stepping strategy for balance recovery with the humanoid COMAN

“…Other approaches for walking stabilization include quick modifications of the CoP reference based on perceived disturbances [21], or posture and force control to simplify the robot walking dynamics [12]. [34] and [1] express stepping strategies within a model predictive control scheme. In [38], a modified version of preview control based on the LIPM is used as a unified way to realize walking pattern generation and stepping for push recovery.…”

Unified Multi-Contact Fall Mitigation Planning for Humanoids via Contact Transition Tree Optimization

“…Prior approaches have either used the internal joints of the robot to resist disturbances (e.g. ankle strategy, hip strategy [1], [2], or inertial shaping [3]) or external contacts (e.g., protective stepping [4], knee contact [5], hand contact [6], [7]). However, these strategies have been considered in isolation and there have been limited attempts to unify multiple strategies into a single fall mitigation system.…”

Adaptive step duration in biped walking: A robust approach to nonlinear constraints

“…On the continuous side, nonlinear MPC schemes have been proposed using Sequential Quadratic Programming (SQP) [5] or Interior Point (IP) methods [6]. On the discrete side, these MPC schemes have been formulated as Mixed Integer Quadratic Programs [7].…”