Performance Optimization of DC 054B-2 Motor with MATLAB-Based Linear Quadratic Regulator (LQR) and Linear Quadratic Tracker (LQT) Method Approach for Control Engineering Applications

Abstract

An accurate system model is essential as a basis for designing a reliable control system. This research aims to design an online system identification technique based on MATLAB to develop system models of Single Input Single Output (SISO), Single Input Multi Output (SIMO), Multi Input Single Output (MISO), and Multi Input Multi Output (MIMO). As a case study, the DC motor type 054B-2 was used to simulate SISO, SIMO, MISO, and MIMO systems in the first and second orders. In the implementation of a control system, external interference in the form of noise is often a major challenge, which can lead to high overshoot and affect system stability. Therefore, the system identification technique is designed to consider the influence of noise, using a MATLAB Simulink-based approach. This research also utilizes the Linear Quadratic Regulator (LQR) method as an optimal control technique that has been widely applied in industry, robotics, and various other engineering fields. The advantage of LQR lies in its ability to provide optimal solutions for systems defined in the context space, as expressed by Yul and Nazaruddin (2018). In addition, the Linear Quadratic Tracker (LQT) method is applied as an alternative, with the ability to follow (tracking) the reference path provided through system input. This method is particularly relevant for applications where tracking reference signals is a priority, as explained by Andria et al. (2014). The simulation results show that the LQR-based approach is able to reduce overshoot and improve system stability, while LQT provides high accuracy in dynamic signal tracking. This research has made a significant contribution to the development of optimal control techniques for DC motor-based applications, as well as opening up opportunities for further application in the field of modern control engineering.