Study On Three-dimensional Node Location Methods For UWB Wireless Sensor Network In Dense Multipath Environment

By the use of unique advantages of UWB signals, wireless sensor networks(WSN) node localization technology is used in measuring seismic signal, soil testing,military reconnaissance and medical monitoring. It is quietly changing the way oflife.In this paper, based on the time of arrival (TOA) estimation for UWB wirelesssensor network node localization problem in dense multipath environment. Theenvironment is considered more obstacles and the propagation path is affected by thewind, rain, snow and other factor of natural environments. The important step of nodepositioning technology is how to measure the distance between the target signal nodeand known node. We consult the detection of CFAR (constant false alarm rate)threshold in radar to find out the TOA estimation value of the first path exceeding thethreshold in dense multipath environments. Then we use the localization algorithmfor WSN node. Under the background noise in the TOA estimate, the traditionalthreshold TOA estimation methods are generally in the Gaussian distribution, themajority of methods are the energy-based detection. Although these methods maydetect TOA value to some extent, since the detection threshold is fixed, the TOAestimate will lead to large errors at low SNR. In WSN positioning, most of thetraditional algorithms located in two-dimensional, their positioning accuracy islimited.Firstly, this paper studied three threshold TOA estimation algorithms based onenergy detection under the traditional Gaussian background noise and analyzedalgorithm performance in simulation. Aim at the deficiency of traditional algorithms,this paper improved TOA estimation method, doing innovative work as follows:Due to the traditional background noise distribution is single; noisemathematical distribution mode cannot be expressed precisely in dense multipathenvironment. We proposed three-dimensional WSN node positioning method basedon ML-CFAR under Weibull distribution background noise. This method is suitablefor dense multipath and the background noise environment obeys Weibull distribution.The simulation results show the ML-CFAR (maximum likelihood-constant false alarm rate) algorithm can estimate the first path TOA value exceeding the thresholdvalue. However, based on ML (maximum likelihood estimation) algorithm mustestimate the average level of background noise, but when the number of sample issmall, the TOA estimation have large errors.Aim at the defect of the traditional energy detection algorithm is fixed threshold,the three-dimensional WSN node localization method is further proposed based onCA-CFAR (cell averaging-constant false alarm rate) under the Weibull distributionbackground noise. The algorithm extracts the part units as the noise reference units,and adding scale factor T, the threshold can adaptive background noise leveldynamically. The simulation results reveal that the error of the algorithm to estimatethe TOA is smaller.Since the traditional WSN node localization algorithms can only be positioned ina two-dimensional plane, and cannot be used in three-dimensional space, we proposeto extend the traditional Taylor series algorithm to three-dimensional space, thenlocate WSN nodes. Under the circumstances that the position of the anchor nodes arereasonable, the number of the anchor nodes is suitable, the algorithm has a goodeffect on location, and the positioning error is small.