Noncoherent fusion detection in wireless sensor networks

Abstract

The main motivation of this thesis is to design low-complexity high efficiency noncoherent fusion rules for the parallel triple-layer wireless sensor networks (WSNs) based on frequency-hopping Mary frequency shift keying (FH/MFSK) techniques, which are hence referred to as the FH/MFSK WSNs. The FH/MFSKWSNs may be employed to monitor single or multiple source events (SEs)with each SE having multiple states. In the FH/MFSKWSNs, local decisions made by local sensor nodes (LSNs) are transmitted to a fusion center (FC) with the aid of FH/MFSK techniques. At the FC, various noncoherent fusion rules may be suggested for final detection (classification) of the SEs’ states. Specifically, in the context of the FH/MFSK WSNs monitoring single M-ary SE, three noncoherent fusion rules are considered for fusion detection, which include the benchmark equal gain combining (EGC), and the proposed erasure-supported EGC (ES-EGC) as well as the optimum posterior fusion rules. Our studies demonstrate that the ES-EGC fusion rule may significantly outperform the EGC fusion rule, in the cases when the LSNs’ detection is unreliable and when the channel signal-to-noise ratio (SNR) is relative high. For the FH/MFSKWSNs monitoring multiple SEs, six noncoherent fusion rules are investigated, which include the EGC, ES-EGC, EGC assisted N-order IIC (EGC-NIIC), ES-EGC assisted N-order IIC (ES-EGC-NIIC), EGC assisted r-order IIC (EGC-rIIC) and the ES-EGC assisted r-order IIC (ES-EGC-rIIC). The complexity, characteristics as well as detection performance of these fusion rules are investigated. Our studies show that the ES-EGC related fusion rules are highly efficient fusion rules, which have similar complexity as the corresponding EGC related fusion rules, but usually achieve better detection performance than the EGC related fusion rules. Although the ES-EGC is a single-user fusion rule, it is however capable of mitigating the multiple event interference (MEI) generated by multiple SEs. Furthermore, in some of the considered fusion rules, the embedded parameters may be optimized for the FH/MFSK WSNs to achieve the best detection performance.
As soft-sensing is often more reliable than hard-sensing, in this thesis, the FH/MFSK WSNs with the LSNs using soft-sensing are investigated associated with the EGC and ES-EGC fusion rules. Our studies reveal that the ES-EGC becomes highly efficient, when the sensing at LSNs is not very reliable. Furthermore, as one of the applications, our FH/MFSK WSN is applied for cognitive spectrum sensing of a primary radio (PR) system constituted by the interleaved frequency-division multiple access (IFDMA) scheme, which supports multiple uplink users. Associated with our cognitive spectrum sensing system, three types of energy detection based sensing schemes are addressed, and four synchronization scenarios are considered to embrace the synchronization between the received PR IFDMA signals and the sampling operations at cognitive spectrum sensing nodes (CRSNs). The performance of the FH/MFSK WSN assisted spectrum sensing system with EGC or ES-EGC fusion rule is investigated. Our studies show that the proposed spectrum sensing system constitutes one highly reliable spectrum sensing scheme, which is capable of exploiting the space diversity provided by CRSNs and the frequency diversity provided by the IFDMA systems.
Finally, the thesis summarises our discoveries and provides discussion on the possible future research issues.

More information

Identifiers

ORCID for Lie-Liang Yang: orcid.org/0000-0002-2032-9327

Catalogue record

Export record