Versatile Locomotion by Integrating Ankle, Hip, Stepping, and Height Variation Strategies

Abstract

Stable walking in real-world environments is a challenging task for humanoid robots, especially when considering the dynamic disturbances, e.g., caused by external perturbations that may be encountered during locomotion. The varying nature of disturbance necessitates high adaptability.In this paper, we propose an enhanced Nonlinear Model Predictive Control (NMPC) approach for robust and adaptable walking -we term it versatile locomotion, by limiting both the Center of Pressure (CoP) and Divergent Component of Motion (DCM) movements. Due to utilization of the Nonlinear Inverted Pendulum plus Flywheel model, the robot is endowed with the capabilities of CoP manipulation (if equipped with finitesized feet), step location adjustment, upper body rotation, and vertical height variation. Considering the feasibility constraints, especially the usage of relaxed CoP constraints, the NMPC scheme is established as a Quadratically Constrained Quadratic Programming problem, which is solved efficiently by Sequential Quadratic Programming with enhanced solvability. Simulation experiments demonstrate the effectiveness of our method to recruit optimal hybrid strategies in order to realize versatile locomotion, for the robot with finite-sized or point feet.

Robust Locomotion Exploiting Multiple Balance Strategies: An Observer-Based Cascaded Model Predictive Control Approach

“…a) Slack Constraints on CoP Movement: To guarantee walking stability, CoP should be restricted within the support polygon. Inheriting from [19], we adopt slack constraints…”

“…To alleviate this limitation, [18] proposed an NMPC scheme, which was formulated as a nonconvex quadratically constrained quadratic programming (QCQP) problem. And an enhanced version could be found in [19]. Although the nonconvex QCQP could be solved efficiently by sequential quadratic programming (SQP), the global convergence was hardly guaranteed in [18] and [19].…”

“…And an enhanced version could be found in [19]. Although the nonconvex QCQP could be solved efficiently by sequential quadratic programming (SQP), the global convergence was hardly guaranteed in [18] and [19].…”

“…Another drawback is that most of the above works were only validated by simulations (e.g., [5], [6], [13], [16]- [19]) or run in an open-loop manner (e.g., [14] and [15]), where the problem of state estimation was ignored. In the presence of dynamic disturbances, including modeling errors such as those caused by the usage of a simplified dynamic model (e.g., the LIP model [1]- [3], [20]) and external perturbations such as those caused by external forces, estimating robot states is a nontrivial problem, especially when considering the height variation and the change of angular momentum.…”

“…In this article, we propose a disturbance observer-based MPC scheme for locomotion control, aiming to exploit ankle, stepping, hip, and height variation strategies for maintaining balance. Differing from [17]- [19] that all the four strategies were integrated into one single MPC, this work adopts a two-layer cascaded structure. Specifically, in the first layer, an NMPC, which is formulated as a QCQP, computes 3-D optimal CoM trajectory while modulating the step locations.…”