Type-2 Fuzzy Logic Controller Design Using A Real-Time Pso Algorithm Applied To "Irobot Create" Robot

Type-2 Fuzzy Logic has emerged recently in Artificial Intelligence as a promising technique in control and decision making. According to the literature, type-2 fuzzy logic controllers (FLC) have been successfully applied in several applications. However, the big complexity in conceiving a controller consists fundamentally in fixing the membership functions of the fuzzy parameters, specifically on how to fix the footprint of uncertainties associated to antecedent and consequent parts, which depends strongly on the current used application. In the case of real-robot navigation, modelling a fuzzy controller is quite tricky, given the presence of several types of uncertainties. In addition, at the level of autonomous navigation, several challenges may arise : how to carefully avoid obstacles, keep the vehicle safe, avoid slipping and most importantly reach the target. In this paper, we proposed an adaptive learning type-2 FLC for a robot navigation. From an initial space which represents a feasible solution, we used the PSO to find a smooth solution with less oscillations. The designed structure achieved good performance shown in the "iRobot Create" robot.