| In recent years, utilization of big data in wireless sensor network (WSN) is increasingly common in daily life, business operation, meteorological observation, military and other important fields. Clustering analysis is one of the important ways to meet the demand of big data classification.We established a system to complete data compression, transmission, storage and clustering, based on the above requirements of wireless sensor network data acquisition and classification. This system solves the problems such as the serious energy consumption of current data transmission, the inefficient of storage systems when it comes to sensory data and the method which clustering operations based on cannot handle the discrete data efficiently, and achieved the following achievements:1. Sensory data is a kind of data with the property of large amount, frequent updating, and the large resource consumption of its requesting and sending. In view of the above properties of sensory data and the situation that traditional database management system or file systems cannot satisfy the current situation of the corresponding data storage requirements, we join together and compress the sensory data in the node, and refactor them in the base station, thus optimize the energy consumption of sensory data acquisition and transmission. We build a distributed storage system, and design the mechanism of data read and storage architecture.2. Given the large size, discrete features of sensory data, low efficiency of the traditional clustering analysis method when dealing with sensory data, we established the suffix tree construction algorithm of discrete big data, based on which we operate the suffix tree clustering operation for classification.By comparison with traditional protocol, Oracle database system and ERA, we verified the reliability and effectiveness of each corresponding part of this system, and ensured the efficient implementation of the classification of data storage and operation. |
PDF Full Text Request