A Fusion Algorithm for Target Detection in Distributed Sensor Networks

Abstract

In this article, we propose a target detection method in wireless sensor networks based on distributed data fusion. Firstly, we introduce a tree topology. It is different from the conventional tree topology, the sensors in our topology are assigned with weights which are proportional to the received Signal to Noise Ratio (SNR), and are arranged orderly. So the Fusion Center (FC) achieves the highest SNR. Secondly, we propose a fusion algorithm which takes the channel noise into consideration on the base of this topology. The sensor makes a decision relying to its two children nodes and its own observation. At last, we prove that the probability of detection is optimized (maximized) for a given pre specified small probability of false alarm. Our simulation results show the effectiveness of our algorithm by comparing with other detection approaches and analyzing the performance of our algorithm.

A novel design of algorithm and framework for decentralized CFAR signal detection without prior information

“…As [8] combines the detection results of each detector, and optimize the weight coefficients for any given knowledge to improve the detection performance. With the development of communication ability, a means of double threshold detection arises for distribution detection [9][10][11]; it must be known that the signal noise rate (SNR) or the detection performance of every detector for reaching the optimal performance [12][13][14]. If making fusion information directly without the prior knowledge (SNR), the result must lead to a decrease performance for the entire system.…”