Range-based Localization Algorithm In Wireless Sensor Networks And Its Applications

Localization is one of the core issues in the researches of wireless sensornetworks (WSNs), and it holds significant position in many applications ofWSNs. Among the localization algorithms, range-based algorithms attract theattentions of researchers and engineers because of its advantages on precision.Under the scenario of large sports stadium, based on its existing needs onlocalization, on security concerns, this paper discusses and studies thefollowing two in need localization issues in details: Localization of strongimpacts on steel beam and that of patrol officers and robots within thestadium region, and further proposes corresponding range-based localizationalgorithm as solutions.This paper constructs two data collecting hardware platforms:1.Platform to collect audio and vibration signals from surrounding. This iscomposed of MSP430f5438a sensor nodes of Texas Instrument andaccelerometers from Silicon Design, and it is capable of time-stamping thecollected data.2. Platform to collect inter-motes ranging measurements. Thisis made of NanoPAN5375RF motes from Nanotron, and it is capable of ranging in the sense of Time-of-Flight (ToF).For localization of strong impacts on steel beam, based on characteristicsof real signals, this paper proposes a Time-Difference-of-Arrival (TDoA)localization algorithm to estimate the location of vibration sources in thesense of data fusion. This paper firstly presents a GeneralizedCross-Correlation (GCC) based TDoA estimation method for audio andvibration signals based on their characteristics. Then utilizing the obtainedestimates, with the idea of sparse signal recovery, this paper gives the TDoAlocalization algorithm based on propagation modeling. At last, adopting theabove algorithms, a fused localization strategy is described with theconsideration of accuracy and property of vibration sources on steel beam.For localization of patrol officers and robots within the stadium regionbased on network model with ToF ranging, this paper presents AWM-SMCalgorithm (Adaptive MDS-dependent Weighting based Sequential MonteCarlo) to achieve a fast and accurate localization of one or more mobiletargets. AWM-SMC applies modified SMC framework and localize mobiletargets utilizing adaptive adjustment and Multidimensional Scaling (MDS).To verify its performance, simulation and field experiments are conducted atlast.