Research On Distributed Detection In Wireless Sensor Networks And Ambient Backscattering Communication

Wireless sensor network(WSN) has been widely used in applications such as remote environmental monitoring and target tracking, as the sensor nodes become smaller, cheaper and intelligent. Due to the constrained communication resources in WSN, it could be prohibitive for a node to send original observations to the FC.On the other hand, ambient backscattering device can communicate with other devices battery-freely, which has gained world-wide attention. One key problem to solve here is for the receiver to decide whether the transmitter has backscattered the ambient signal or not.In our research, we study traditional distributed detection algorithm in WSN and ambient backscattering communication. A 1-bit quantization based GLRT detection scheme in WSN, and two distributed detection schemes in ambient backscattering communication, one based on amplify-and-forward relay and the other based on local likelihood ratio test, are proposed.For WSN, we propose a scheme which quantizes original observation into a binary data at the sensor nodes to meet its stringent power and bandwidth constraint. Each sensor quantizes its observation into a binary data which is sent to the FC to make a global decision. We come up with the GLRT detector design and the optimal quantizer design. The asymptotic performance of the detector is analyzed for both cases where the data are sent via perfect and imperfect channels. Theoretical analysis indicates that to meet the same detection performance, the one-bit GLRT detector requires ?2 times the number of sensors used by the clairvoyant GLRT detector. Simulation results are carried out to corroborate theoretical analysis and to illustrate the performance of the proposed scheme.For ambient backscattering communication, we transform the problem of deciding whether the transmitter has backscattered the ambient signal or not into a binary hypothesis test problem. An amplify-and-forward relay based distributed detection scheme and a local LRT based distributed detection scheme are proposed, and the performance of the detectors is analyzed. Simulation results corroborate our theoretical analysis, and indicate that by increasing the number of the receivers, detection performance can be improved.