The Research Of Clustering Of Wireless Sensor Networks And Its Implementation On Tinyos Platform

Wireless sensor network (WSN) consists of a large number of low-cost, low-power, multifunctional sensor nodes that are small in size and communicate untethered in short distance. These tiny sensor nodes, which are always equipped with sensing, data processing and communicating capabilities, operate cooperatively in ad hoc manner to sense the information in coveraged area, process the sensed data and then transmit the result to the viewer. Wireless sensor network is widely used in applications such as target tracking, battlefield surveillance, health and environment monitoring. Clustering Algorithm and node architecture design which are two of the key techniques in wireless sensor networks have great impact on the performance of wireless sensor networks.First, we summarize wireless sensor networks, including the architecture, features and the key techniques. We also give an overview of several representative clustering algorithms recently proposed and their weakness.Then, to address the unbalanced energy consumption problem of existing clustering algorithms, we propose a clustering algorithm called LBEED (Load-Balanced Energy Efficient Distributed Clustering). We make the network have minimum energy consumption through modeling the energy consumption and optimizing the parameters. Furthermore, power control is used to make the network form almost equal size clusters which consequently balances the load of every node. MATLAB simulation reveals that the proposed algorithm performs better than HEED and LEACH in load balance aspect.Software and Hardware Platform design for sensor nodes is also an important aspect of wireless sensor networks. After a detailed description of the Hardware Platform which is based CC2430 Chip and the TinyOS operating system, HEED and LBEED are implemented on the Platform. We implement X-MAC and introduce an ultra-low power operating mode as well. Based on above design, a small-scale network is set up to test the performance of the clustering algorithms and X-MAC, the results prove that LBEED makes the network form equal-size clusters; besides X-MAC makes nodes have low duty cycle and consequently prolongs the network lifetime.