A Novel Adaptive Fuzzy Inference System for Mobile Robot Navigation

The Fuzzy hybridization technique for intelligent systems have become of research interests in a variety of research areas over the past decade. There are limitations faced by all popular fuzzy systems architectures when they are applied to applications with a large number of inputs (more than three). The present paper proposes a novel adaptive fuzzy inference system for multi-sensors mobile robot navigation. A novel fuzzy inference system is constructed by the automatic generation of membership functions (MFs) and formed a minimal numbers of rules using hybrid fuzzy clustering algorithm (Combination of Fuzzy C-means and Subtractive clustering algorithm) and the modified apriori algorithm, respectively. A modified apriori algorithm is utilized to count the number of common elements from the clusters and to obtain a minimal set of decision rules based on input-output datasets. The generated modified adaptive fuzzy inference system is then adjusted by the least square method and the gradient descent algorithm towards better performance with a minimal set of rules. The proposed algorithm is able to reduce the number of rules which increases exponentially when more input variables are involved. The performance is compared with other existing approaches in an application of mobile robot navigation and shown to be very competitive and improved results.