Optimal Fuzzy Logic Controller based on Invasive Weed Optimization for a Single-link Flexible Manipulator System
Hyreil A Kasdirin, Adrian Leong Wai Chen, M Ruzaini Hashim
Universiti Teknikal Malaysia Melaka
Abstract
Nowadays, a single-link flexible robot arm is used extensively in a lot of industries. This is because it can manage duties that are hard or repetitive work in varieties of production tasks. Therefore, in managing to operate like a human hand, this flexible manipulator system is needed to be extremely precise. Hence, this paper proposed an optimal fuzzy logic controller (FLC) by improving the accuracy and repeatability of a proportional-integral-derivative-like fuzzy logic controller (PID-FLC) for a single-link flexible manipulator system. In this strategy, the proposed controller is optimized by an invasive weed optimization (IWO) algorithm. To simulate such system, a PID-type fuzzy logic controller will be designed in the MATLAB Simulink environment. In the control strategy, the PID acts as scaling gain for the FLC in order to normalize inputs and output and to damp oscillations of the system. The performance of PID-FLC is tuned and optimized by IWO algorithm by discovering the best value of membership function for fuzzy system. The effectiveness of the proposed strategy is analyzed and evaluated by comparing the performance of PID-FLC with PD-like FLC and PID-like FLC without tuning mechanism from IWO algorithm. The comparative studies are validated in terms of rise time, overshoot and undershoot, settling time and steady-state error of the single-link manipulator movement. As a result, the proposed control strategy only needs only less than 30 number of iterations with less time-consuming and lower cost function in order to generate the optimum performance. Furthermore, PID-FLC shows better performance by improving 11% of lower overshoot at first signal, 27.9% of smaller undershoot at second signal and 0.8% lower overshoot at third signal compared to PD-FLC with IWO. PID-FLC also achieved better overall performance than PID-FLC without IWO in terms of rise time, settling time, overshoot and undershoot and steady-state error. However, there are some future works to be done which are comparing the performance of IWO with other biological-based algorithm and apply this optimized control system to the real model of robot arm.
Keywords: Fuzzy Logic Controller, Invasive Weed Optimization, Single-link Flexible Manipulator
Topic: Electrical Engineering