Adaptive Finite-Time Coordination Control of a Multi-robotic Fiber Placement System With Model Uncertainties and Closed Architecture

Abstract

The coordination and trajectory tracking accuracy of multi-robotic fiber placement systems (MRFPSs) are critical to assure the quality of the fiber placement process. However, the model uncertainties and closed architecture (CA) in industrial robots significantly hinder the system from achieving high performance in coordination and tracking simultaneously. In addition, the convergence rates of the tracking and synchronization errors are also essential performance indicators for the MRFPSs. To improve the three abovementioned performances, this article presents an equivalent model of the CA dynamics based on a radial basis function neural network. Employing this equivalent model, a novel indirect torque control algorithm named adaptive finite-time coordination control (AFCC) is proposed for a MRFPS consisting of two heterogeneous robots. Within the controller, two adaptive laws are designed to handle the uncertainties, and three additional adaptive laws are developed to mitigate the effects of the unknowns in the CA, contact forces, and disturbances. The stability analysis of the AFCC algorithm proves that the errors can converge to zero within a finite time. Furthermore, three experiments show the advantages and practicality of the AFCC algorithm.