Multi-sensors data fusion through fuzzy clustering and predictive tools

Sensory data are generally associated with imprecision and uncertainty, and consequently, it becomes difficult to extract useful information from them. The problem becomes even more difficult to handle, when the data are collected using multiple sensors. Realizing the ability of fuzzy sets to deal with imprecision and uncertainty, a multi-sensors data fusion technique was developed in this study by using fuzzy clustering and predictive tools. The data were first clustered based on their similarity using an entropy-based fuzzy C-means clustering technique and the obtained clusters were utilized to develop a fuzzy reasoning-based predictive tool. The novelty of this study lies with the application of a clustering algorithm, which can ensure both compactness and distinctness of the developed clusters, and development of a reasoning tool utilizing the information of obtained clusters. Two types of multi-sensors data were used to test the performance of the proposed algorithm. Results were compared with those available in the literature, and the developed technique was found to perform better than the previous approaches on both the data sets. The better performance of the proposed algorithm could be due to its in-depth search of the data set through similarity-based fuzzy clustering followed by the development of fuzzy reasoning tool utilizing the information of obtained clusters.