Journal of Mechanical Engineering Research and Developments ISSN: 1024-1752 CODEN: JERDFO Vol. 43, No. 7, pp. 255-270 Published Year 2020 255 Non-linear PID Controller for Trajectory Tracking of a Differential Drive Mobile Robot Umar Zangina †,‡† , Salinda Buyamin †* , Mohamad Shukri Zainal Abidin † , Mohd Saiful Azimi Mahmud † , Hameedah Sahib Hasan †,‡† † School of Electrical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor, Malaysia ‡ Ministry of Higher Education and Scientific Research, Al Furat Al Awsat Technical University, Iraq. ‡† Research Fellow l @ sokoto energy research centre Usmanu Danfodiyo University sokoto, Nigeria *Corresponding author E-mail: salinda@fke.utm.my ABSTRACT: The application of differential drive robots has grown from scientific research to broader industrial and commercial purposes. In order to Navigate the robot in difficult terrains, it must be well equipped with a robust controller with good path tracking ability and general stability. Typically, the wheeled mobile robot (WMR) can essentially be kinematically controlled by defining a route and determining the traveling time, speed and direction to get from one place to another. However, by ignoring the dynamic model of the robot, a purely kinematic model approach has been revealed to produce unrealistic results at higher speeds and loads. As a consequence, there are significant limitations to the applicability of solely kinematic systems to mobile robotics and hence, in recent years, there has been a trend towards the application of dynamic modelling. In this study, a simple but effective solution for the path tracking problem of a mobile robot using a PID controller is proposed. The method adopted is a trial and error technique with six tuning parameters for the robot to track a desired trajectory. The final mathematical derivation for a nonholonomic differential drive mobile robot was computationally simulated using MATLAB for both kinematic and dynamic models respectively. The controller was used to overcome the nonlinearity of the reference trajectory tracking as well as the speed of the DC motor adjustments. In order to evaluate the performance of the developed robot controller, tests were also carried out for different trajectories in terms of the initial and final conditions. The results show that the developed PID controller is responsive enough to be able to speed up when required to match the reference trajectory. KEYWORDS: differential drive, trajectory tracking, mobile robot, PID controller and dynamic model. INTRODUCTION Mobile robots are robots that can move autonomously from one particular predefined location to another. They possess the exceptional features of stirring around without restrictions within a known workspace contrasting with the majority of industrial robots that can only move in a definite workspace [1]. This flexibility makes them more suitable for a considerably better performance in terms of applications in structured, semi-structured and unstructured environments. Differential drive wheeled mobile robot is the most commonly used mobile robot and it comprises of a robot platform having two fixed powered wheels attached to its left and right sides. Both wheels are independently driven and one or more passive castor wheels are used for stability and balancing [2]. The robot moves straightforward or backward if the wheels rotate at the same speed. It tracks a curved route along the arc of an instantaneous circle if one wheel is running faster than the other; and if the two wheels are rotating at equal speeds in opposite directions, the robot turns about the midpoint of the two driving wheels [2,3]. From the viewpoint of control design, an essential characteristic of mobile robots is the non-square proportions of the models. However, this makes the task of designing their controller simpler since they generally entertain only two variables (linear and angular speed) in modelling the movement while employing three variables to model their positions [4]. The position variables are the x and y coordinates on a plane and the angle the robot forms with the horizontal x axis. Therefore, there are fewer control variables than variables to be controlled [5]. The simplest and most fundamental control of a robot in an environment can be achieved with a point-to-point method (classic control), as the path of the states between the initial and final states are not critical. The second option is to control brought to you by CORE View metadata, citation and similar papers at core.ac.uk provided by Universiti Teknologi Malaysia Institutional Repository