Target Localization In Wireless Sensor Networks Base On EM Algorithm

As a new technology to abstain data information, WSNs has many advantages inscientific research, space exploration and many other applications than traditionalinstruments. Among all those applications, target localization is particular important.in this paper, the EM algorithm-BEM iterative algorithm based on binary data (i.e., asingle byte of data between0and1where0represents the absence of target and1represents the existing of target) has been investigated and improved.In wireless sensor networks, target localization algorithms based on binary decisionsalways present inaccurate performances in practical applications since the information ofreal data implied in binary decisions which generated by sensors is limited. Furthermore,in order to ensure the accuracy of localization, it is inevitable to use global searchalgorithms during the process of iterations, which greatly increases the time complexityof the algorithm, undermining the requirements of real-time localization in practice.Considering the disadvantages of BEM algorithm mentioned above, this paperinvestigate a new algorithm called LEM iterative algorithm base on BEM algorithm. Thepresenting algorithm has reduced the time complexities theoretically. And with thedescription of the simulation by Matlab7.0,LEM algorithm presents a more accuracyperformance than the existing BEM algorithm. Furthermore, LEM algorithm is morestable than the BEM algorithm in high noise environment.Finally, we give a summary for this article and make a generalization for the workdirection of future.