Performance Analysis of Distributed Wireless Sensor Networks with Data Fusion

Distributed signal detection and fusion in wireless sensor networks (WSNs) have emerged in a variety of applications such as battlefield monitoring, surveillance, weather forecasting, and intruder detection systems. However, performance analysis of such systems has not been carefully considered in the literatures. This paper considers the performance analysis of distributed WSNs with noisy communication links between the distributed receivers (or sensors) and the access point (AP) (or central receiver). We considers the three main data fusion schemes; hard decision (0 or 1), soft decision (between 0 and 1), and observations fusion. The performance analysis is considered in case of two different scenarios of non-coherent multiple-receiver communication systems, wherein each individual receiver provides the AP with a hard decision, or a soft decision, or an observation. The performance loss between the observation fusion, soft decision fusion, and hard decisions fusion is investigated. The needed number of levels, in case of soft decision fusion approach, to achieve a performance comparable to the observation fusion approach, is also addressed.