Node localization in underwater sensor networks (UWSN)

This dissertation focuses on node localization in underwater wireless sensor networks (UWSNs) where anchor nodes have knowledge of their own locations and communicate with sensor nodes in acoustic or magnetic induction (MI) means. The sensor nodes utilize the communication signals and the locations of anchor nodes to locate themselves and propagate their locations through the network. For UWSN using MI communications, this dissertation proposes two localization methods: rotation matrix (RM)-based method and the distance-based method. Both methods require only two anchor nodes with arbitrarily oriented tri-directional coils to locate one sensor node in the 3-D space, thus having advantages in a sparse network. Simulation studies show that the RM-based method achieves high localization accuracy, while the distance-based method exhibits less computational complexity. For UWSN using acoustic communications, this dissertation proposes a novel multi-hop node localization method in the 2-D and 3-D spaces, respectively. The proposed method estimates Euclidean distances to anchor nodes via multi-hop propagations with the help of angle of arrival (AoA) measurements. Simulation results show that the proposed method achieves better localization accuracy than existing multi-hop methods, with high localization coverage. This dissertation also investigates the hardware implementation of acoustic transmitter and receiver, and conducted field experiments with the hardware to estimate ToA using single pseudo-noise (PN) and dual PN(DPN) sequences. Both simulation and field test results show that the DPN sequences outperform the single PNs in severely dispersive channels and when the carrier frequency offset (CFO) is high --Abstract, page iv