Study of the Angular Positioning of a Rotating Object with Neural Model Reference Control CONSTANTIN VOLOŞENCU Department of Automation and Applied Informatics "Politehnica" University Timişoara Timişoara, 300223, ROMANIA Abstract - The study in the paper is placed in the broad context of research for increasing the efficiency of motion control. The purpose of the paper is to make a comparative analysis of the neural model reference control with the linear control for angular positioning of mechanical parts. The structure of the neural model reference control system and its design are presented. Transient characteristics obtained are compared from the point of view of their control efficiency criteria. The differences in performance criteria between the control methods studied are small. Keywords: Motion control, linear control, deep learning, neural network model reference control, neural identification. Received: March 24, 2021. Revised: September 20, 2021. Accepted: October 7, 2021. Published: October 23, 2021. 1. Introduction The paper presents the results of a research study in the field of motion control applied to positioning of an object with known moment of inertia, in the rotational motion around an axis. The purpose of the work is the analysis of methods for positioning based on neural network model reference control, compared with the conventional linear control method. The significance of the study is that it shows the efficiency of these methods, compared to each other. An example of position control of a vehicle subjected to unknown conditions using sliding mode and optimal control is presented in [1]. In [2] a study on the optimization of motion control in automatic machines, robots and multi-body systems is presented. In [3] some examples of the application of intelligent control techniques in motion control are presented. Conventional position control is done using as actuators electric machines, driven by cascade control systems, with internal current control loop, over which overlap a speed control loop and an external position control loop. This is the natural way for control. The current and speed control loops must respond as quickly as possible. And the control of the position must be done asymptotically, aperiodically with zero overshoot. In this paper, a heavy object is taken into consideration. The actuator inertia is not taken in consideration, because it has a very small time constant, compared to the moment of inertia of the object. Neural networks bring learning and training in control. The paper presents, in section 2, preliminary information related to the mathematical model of positioning process, the conventional linear position control, its transient characteristics and performance criteria. The third section presents the position control method based on neural network model reference control. A neural model of the process is developed based on neural identification of the motion model, testing and validation. The neural controller is also trained. The methods were modelled and simulated in Matlab/Simulink. The results that can be obtained with these method are presented in section 4. The characteristics obtained by simulations are compared, analyzed, and discussed. The main contribution of the paper can be summarized as a comparative analysis of two position control methods: conventional linear control and neural model reference control, with application in the particular case of a heavy object in rotational motion at variable angular positions. The behavior of the system with neural predictive control is analyzed. The results are compared with those obtained in the case of linear control. The analyzed methods ensure good performance criteria: zero control error, reduced response time, zero overshoot. The performance criteria differences between the control methods are small. 2. Preliminaries 2.1. Motion process It is considered to adjust the position θ of the object with the moment of inertia J in the rotational motion with angular velocity ω. The rotational movement takes place in the presence of friction. The equations of motion are:        dt d k M dt d J f (1) where M is mechanical torque and kf is coefficient of friction. A speed sensor is used, considered as a first-order delay element with a time constant TTω. The values of the parameters considered are: J = 450 kg.m 2 , kf = 120 kg.m 2 /s and TTω = 0.12 s, and maximum values: Mm = 1000 Nm, ωm = 0.3 rad/s and θm = 180 o . 2.2. Linear Control System A closed-loop, cascading position control system is selected, as the reference control system. In this system, the speed in the inner loop and the position in the outer loop are WSEAS TRANSACTIONS on COMPUTERS DOI: 10.37394/23205.2021.20.25 Constantin Voloşencu E-ISSN: 2224-2872 234 Volume 20, 2021