A Self-Adaptive Inertia Hybrid Control of a Fully Constrained Cable-Driven Parallel Robot

摘要

A Cable-Driven Parallel Robot (CDPR) is a special type of parallel robot actuated by elastic cables instead of rigid links, which brings uncertainties of model nonlinearities and structural friction depending on the payload variation. A precise motion with minimum control effort against an unknown payload has been a challenging task. To conquer it, we propose an adaptive hybrid control approach for the fully-constrained CDPR with an unknown payload in this paper. A Cartesian space position control and a joint space cable force control are designed through inertia estimation to encounter the effect of the uncertain payload. And then, an artificial neural network (ANN) is employed to estimate the end-effector force in Cartesian space for precise force control and consecutive uncertainties compensation. The stability of the proposed controller is proven based on the Lyapunov method and the efficacy is verified by the experiments. As for feasibility results, the proposed adaptive hybrid control-based cable force estimation was evaluated on the Mini CDPR. Compared to the traditional control scheme, the tracking error was decreased approximately 56% and the utilized tension was significantly reduced for the given trajectory tracking test. Furthermore, the investigation of free movements of the helix trajectory and pick-and-place tasks showed that the proposed hybrid adaptive control can control the position and cable force altogether while being able to estimate the inertia of the end-effector. Note to Practitioners—The study was motivated by the control problem of adjusting cable tensions of a cable-driven parallel robot (CDPR) against varying payloads at the end-effector. Several advanced control approaches were proposed for but limited to CDPR model parameter estimation and did not consider varying payload. This paper suggests a practical methodology capable of both self-adaptation to unknown payload inertia and consequently improving tracking control performance for a CDPR. In this paper, we derived an adaptive control model and proved stability to ensure safety. In order to encounter unmodeled uncertainties such as pulley friction and cable characteristics, we utilized an artificial neural network (ANN) trained by experiments. The experimental results could confirm that the proposed scheme can estimate the payload inertia and generate minimized cable tension while maintaining good trajectory tracking performance. In future research, we will show more practical validation for a scale-up CDPR with heavy objects toward logistics or pick-and-place application in practice.