Abstract
The Interval Type-2 Fuzzy Logic Controller (IT2FLC) is an advanced version of Type-1 Fuzzy Logic Controller (T1FLC) that improves the control strategies by using the advantage of its footprint of uncertainty of the Fuzzy Membership Function (MF). Numerous experimental investigations have shown the superiority of IT2FLC over T1FLC, particularly in high level of uncertainties and nonlinearities. Nevertheless, the systematic design of IT2FLCs remains an attractive problem because of the difficulty in finding the parameters associated with IT2FLCs. In this study, a novel application of Cuckoo Search (CS) algorithm in the design of an optimized cascade Interval Type-2 Fuzzy Proportional Integral Derivative Controller (IT2FPIDC) is presented. The PID gains and the parameters of the antecedent MFs of IT2FPIDC are optimized using CS algorithm. Considering the higher number of parameters to be optimized in cascade IT2FPIDC, the CS method was employed due to its high convergence speed and less computational cost. The proposed CS based cascade optimized IT2FPIDC is compared with CS-based Type-1 Fuzzy Proportional Integral Derivative Controller (T1FPIDC). The present research presents a new application of proposed CS based cascade optimized IT2FPIDC for the balancing control and trajectory tracking control of the Furuta pendulum (FP) which is a nonlinear, non-minimum phase and unstable system. Furthermore, the disturbance rejection ability of the proposed controller is analyzed. The proposed control strategies are evaluated on FP produced by Quanser through numerous experimentations in the real world as well as simulation. The performance characteristic considered for these controllers are settling time (ts), steady state error (Ess), rise time (tr) and maximum overshoot (Mp). Both the simulation and real world experiments results demonstrated the robustness and effectiveness of the proposed CS based IT2FPIDC with respect to parameter variation, noise effects and load disturbances, with an improved performance of 27.1%, 5.7% and 20.8% for tr, ts and Mp respectively over its CS based T1FPIDC counterpart.