Three-dimensional Localization Methods For UWB Wireless Sensor Networks Based On Super-resolution TOA Estimation

Wireless sensor networks (WSN) is a multi-hop ad-hoc network systemcomposed of a large number of sensor nodes with multiple functions of informationcollection, data processing and wireless communications randomly distributed in aspecific area. It has become one of the three major high-tech industries in the futureowing to its unique potential and superiorities. In most applications, the informationcollected by the node would be meaningless without the node locations as theprerequisite and the basis. Since the energy and the processing power carried bysensor nodes are limited, the research of node positioning technology is of greatsignificance.So far most localization algorithms for WSN are two-dimensional and cannot beapplied to three-dimensional environments. This thesis focuses on three-dimensionallocalization methods for UWB wireless sensor networks based on TOA estimation.Under the principle of low cost, low complexity and low power consumption of nodeequipment, considering multipath effect, none-line-of-sight communication andcomplex noise, this thesis proposes super-resolution TOA estimation methods basedon Unitary Matrix Pencil algorithm (UMP), the Amplitude and Phase Estimation(APES) method and Cell Average-Constant False Alarm Rate (CA-CFAR) algorithm,respectively and combines them with multilateral range-based localization algorithmsto realize precise three-dimensional localization of unknown nodes in wireless sensornetworks. First the significance of the subject is described. Besides, commoncategorization and the corresponding research status are introduced. Then threelocalization methods based on super-resolution TOA estimation for UWB WSN arefurther proposed. Simulation analyzes the effect of SNR, number of anchor nodes anddistribution on their localization performance. This thesis is supported by JilinNatural Science Foundation “3D node localization using threshold-based TOAestimation for UWB wireless sensor networks in dense multipath environment”(201215014). The main innovative work is as follows:1、 Localization methods based on MP algorithm and UMP algorithmrespectively are proposed. MP algorithm and UMP algorithms are used in super-resolution TOA estimation and distance measurement, and combined withthree-dimensional Taylor algorithm and three-dimensional Chan algorithm tocalculate positions of unknown nodes. Both of these two algorithms can suppressnon-Gaussian noise and only needs single sampling.UMP algorithm is improved fromMP algorithm, which has smaller computational amount and higher accuracy on thebasis of retaining the advantages of MP algorithm.2、Localization method based on APES technique is proposed. TOA and signalamplitude are estimated by APES technique with super-resolution and then used inthree-dimensional position calculation. Through estimation of signal amplitude,APES algorithm can directly pick up the direct path in Line-of-Sight situation andovercome multipath effect.3、Localization method based on CA-CFAR algorithm is proposed. CA-CFARalgorithm is originally used in radar system and here it is adapted to estimate TOA forunknown nodes. CA-CFAR algorithm is a threshold-based method and can determinethe direct path in dense multipath environment to suppress the effect ofNone-Line-of-Sight. Besides, it is robust and has small computation amount, which issuitable for battery-limited systems. Simulation proves the validity.