Multi-optimization of PID controller parameters using stochastic search techniques for rotary inverted pendulum system

Abstract

Proportional-Integral-Derivative (PID) controller is a well-known controller in various aspects of industrial automation due to its simplicity and effectiveness in design and implementation of industrial applications. However, it has been difficult to tune up PID controller gains accurately because of complexity and nonlinearity of industrial plants. Therefore, the selection of controller parameters are usually complex and sometimes are selected via trial and error and from designer’s intuitive and experience, resulting in less optimal performance. The aim of this project is to analyze and formulate multi-optimization methods for the parameters of the PID controller for controlling the angular displacement of pendulum and arm of Rotational Inverted Pendulum (RIP) system. In this project, the RIP system is chosen due to it is known as an inherent nonlinear system which can be a good prospect for control engineering and MATLAB has been also used to simulate and verify the mathematical model. The performance of the PID controller is evaluated and compared when the parameters are automatically optimized using the Model Reference Adaptive Control (MRAC) concept and stochastic algorithms such as Particle Swarm Optimization (PSO) and Differential Evolution (DE) algorithms for satisfying the main goal which is balancing the pendulum in the vertical position . Finally, the results demonstrated the robustness and effectiveness of the designed PID controller by proposed stochastic algorithms in terms of easy implementation, computational cost, complexity and effectiveness. As a conclusion, these proposed stochastic search techniques can be considered as systematic and effective ways to control the various nonlinear industrial plants.