Position control of a DC motor system for tracking periodic reference inputs in a data driven paradigm

Abstract

Control techniques over the decades have evolved from the various aspects of Model-Based Control (MBC) to Data Driven Control (DDC). In stark contrast to the model based paradigm which is targeted at addressing the fundamental physics driving the system and intends to freely determine the process transfer function. The data-driven approach instead connotes to ascertaining the process parameters of a system, void of a specified architecture by measuring the input and output data. The present investigation targets a validatory execution of a simple feedback DDC architecture on the position control of a DC motor module. The input-output data obtained from an onboard potentiometer is logged to the host PC using a low cost acquisition set-up and the corresponding model structure is identified using Matlab System Identification Toolbox. Based on the identified plant model, an appropriately tuned PID scheme is proposed that can represent the original hardware response with acceptable fidelity. The ability of the proposed control scheme is augmented by the introduction of standard repetitive control strategy in order to reduce Steady-State tracking errors of the system while negotiating periodic inputs. The experimental results demonstrate the effectiveness of the proposed scheme in offering a highly accurate asymptotic tracking ability.