Rough Set Based Generalized Fuzzy C-Means clustering Algorithm with Spatial Constraints for Distributed WSN Data Stream

Abstract

In practical applications, wireless sensor network (WSN) generate massive data streams with the spatial and sensor measurements information, and moreover, energy source of sensors is usually limited. Therefore, minimizing sensors energy expenditure and consequently extending the network lifetime is the major challenge in WSN. This paper presents an efficient distributed clustering algorithm based on FCM by incorporating the neighborhood sensor spatial information into the FCM algorithm (FCMS) to meet sensor data correlation increasing with decreasing spatial separation. FCMS can overcome the disadvantages of the known fuzzy c-means algorithms and at the same time enhances the clustering performance. The major characteristic of FCMS is the use of a fuzzy local ((both spatial and sensing measurements) similarity measure, aiming to partitions the sensor data into a set of spatial regions with similar sensing measurements. FCMS is initialized by Subtractive Clustering algorithm, in which the number of cluster and the cluster centers is taken to the FCMS Algorithm. The distributed FCMS (DFCMS) forms clusters of the sensor nodes sensing similar values and transmits features from the created local clusters as opposed to raw data per sensor node as in central clustering algorithm. Thus, DFCMS can significantly reduce the number of transmissions, which results in energy savings. Simulations reveal that DFCMS algorithm is effective and efficient with significantly less energy than that required by central FCM algorithm.