Majlesi Journal of Electrical Engineering Vol. 14, No. 2, June 2020 81 Novel m-PSO Optimized LQR Control Design for Flexible Link Manipulator: An Experimental Validation Naveen Kumar 1* , Jyoti Ohri 1 1- Department of Electrical Engineering, NIT Kurukshetra, India. Email: naveenvermaindia@gmail.com(Corresponding author) Received: August 2019 Revised: November 2019 Accepted: January 2020 ABSTRACT: Recently, robotic manipulators are the key industry requirement. These have find the importance to enhance the productivity as well as accuracy. Furthermore, industries are also moving towards the use of Flexible Link Manipulator (FLM) owing to their unique characteristics i.e. light weight, high speed operations, and the larger workspace. The FLM system has flexibility of link that causes vibrations and oscillations which affect adversary to the performance of robotic arm. The performance of FLM system is measured w.r.t. minimum error and oscillations in trajectory tracking. In this research paper, an attempt has been made to overcome the complications of FLM system. A full state feedback Linear Quadratic Regulator (LQR), is designed for FLM. It is observed that the designed controller can enhance the accuracy of the robotic arm, while reducing oscillations and vibrations. In addition, to enhance the performance of controller and to reduce the hassle in terms of selecting the parameter of Q matrix in LQR, modified particle swarm optimization (m- PSO) is used. The effectiveness of designed controller is simulated in MATLAB. Further, the validation of designed controller is tested on hardware FLM device. The results obtained from the simulation and hardware are compared. . KEYWORDS: PSO, FLM, LQR, Vibration, Tracking Error. 1. INTRODUCTION Robotics field involves the application of diverse disciplines such as physical, static and dynamic properties of materials, control theory, electronics, vision and signal processing, computer science. A robotic manipulator is basically a mechanical arm designed to work the similar task as of human arm. These are used for various industrial applications to perform repeated task precisely. These are consist of number of links and joints, rigid as well as flexible [1]. The Flexible Link Manipulator (FLM), received great attention in the past few decade among the researchers. For industrial application, it has shown various advantages over rigid manipulator in terms of light weight, high speed, low inertia, lower energy consumption and large work space [2]. The FLM has complex dynamic structure compared to rigid link and hence it becomes a difficult task for control engineers to design a control law. In case of industrial applications, preciseness about the given tasks, are always desirable. Therefore, control of position and oscillation becomes an important performance aspect, whereas in case of FLM, flexibility and the vibration presented in the link itself, leads to oscillations in the output. Hence, a control law is necessary to design with the objective to track the desired position with minimum or zero oscillations. Various authors have proposed different control strategies to follow the desired position or trajectories precisely. In [3], LMI and SMC based control law is design for FLM. PI and Fuzzy logic based controller for flexible joint are designed in [4]. Various other control strategies given in literature are variable structure control [5], optimal control [6], adaptive control [7], robust control [8], and intelligent based neural control [9], [10] etc. Among them, optimal control method is chosen in this work. In this, a full state feedback system is designed to find the gain of state feedback control. Simulation based LQR method has been designed for FLM in [11]. Furthermore, in LQR, Q matrix parameter selection is always a hectic task. In literature, it has been chosen based on research experience available and fined tuned. Therefore, for better tracking and stability, optimization algorithms may be employed such as Genetic algorithm (GA) [12], [13], Particle Swarm Optimization (PSO) [14], Ant Colony Optimization (ACO) [15] etc. In these, PSO has proved its capability over wide range of applications, in case of simulation as well as experimental work [16], [17]. Further, this PSO tuned LQR optimal control law is rarely used for flexible link manipulator simulation which can help largely to find the suitable parameter of Q matrix in LQR. It can enhance the tracking ability along with stability of FLM.