Research Of A Max-expect Service Based Cluster-Heads Algorithm In WSN

Due to recent advances in micro-electro-mechanical systems(MEMS), wireless sensor networks (WSN), which are mostly used to collect data, have received a lot of attentions both in military and civil fields.Generally, such systems consist of mass tiny nodes with sensing, data processing and wireless communication modules on them.These nodes collaborate with each others in a self-organized way, to send the sensed information in the vicinity back to the sink node through the network effectively.Since this kind of system is able to cover large sensing fields while guaranteeing high quality, it is widely used in military sensing, security,environmental monitoring, traffic surveillance, medical treatment, building and structures monitoring and even anti-terrorism.With resources and energy constraints, research of a clustering algorithm available for data fusion has become a key task in wireless sensor networks. Recent studies indicate that some clustering algorithms have been proposed.LEACH algorithm is one of classical wireless communication algorithms for WSN.It is a self-organized, adaptive, energy-efficient, clustering network algorithm, Which accounts for properties of WSN and prolongs the life-time of WSN efficiently.But the mechanism of stochastic Cluster-Heads(CHs) selection in LEACH algorithm isn't optimal,which can't guarantee the proper position of CHs and doesn't account for energy of CHs left. So it would lead to the energy of CHs exhausted and reduce the life-time of network wholly. On the other hand, single-hop communication is taken when CHs send data to the base station, which aggravates the energy consummation of the CHs.In this paper, after analyzing advantages and disadvantages of LEACH algorithm in forming clusters and considering there may be CHs nodes' service failure due to limited nodes resources and wireless link failure, a max-expect service clustering algorithm is put forward on the assumption that CHs service failure rate is p. As simulation results show, max-expect service demands are affected by many factors, such as service distance, service failure rate and number of cluster heads. In a word, The max-expect service clustering algorithm performs more efficiently than traditional clustering algorithm, namely LEACH.