Clock Synchronization, Positioning And Data Security Storage Algorithm Of Wireless Networked Instrument

Wireless networked instrument can be flexibly used in various occasions, andeasy-to-build distributed test and measurement systems, remote monitoring andcontrol can be achieved as well as the software and hardware resource sharing.Wireless networked instrument is more suitable than wired instrument to work in anoutdoor or field test and measurement system. Without cable connection, themechanism of synchronization trigger can only be accomplished through an unifiedclock. Positioning and data security of wireless networked instrument are importantfor distributed system’s well running.In this thesis, the characteristics of wireless channel are analyzed and channelmodels are determined according to the working environment of instrument, thestudies on this basis as follow:1) The instrument’s internal clock need to provide accurate time to performinstrument synchronous trigger and data fusion in distributed systems. The crystalconstituting the clock will affect the clock stability and accuracy due to environmentalfactors and inherent characteristics. In this thesis, the clock with the best performanceis master clock, while the other clocks are slave clock. The slave clock can calculateand correct its own clock parameters by exchanging the synchronization message withmaster clock. Using Precision clock synchronization protocol for networkedmeasurement and control systems (PTP) to calculate the clock offset in wirelessnetwork is inaccurate because of the asymmetry transmission. Based on the PTPalgorithm, we proposed a clock synchronization algorithm based on square-rootKalman filter (SR-KF). State equation of Kalman filter can be established accordingto the clock model and the exchanged synchronization message between master clockand slave clock, and then use SR-KF algorithm to obtain optimal estimations of theclock offset and skew, the accuracy can be stabilized after several iterations. In thesame simulation environment, the SR-KF clock offset estimation precision is muchhigher than PTP.2) The position of instrument is important for remote monitoring and data fusion.Positioning with non-line-of sight (NLOS) TOA can lead to large position errors.There are broadly three ways to cope with the NLOS condition. The first way islocalize with processed TOA, the difficulties are in obtaining an accurate model ofNLOS. The second way localizes with all TOA measurements, but provides weightingto minimize the effects of the NLOS contributions. The problem is that NLOS errors are always present though reduced. The third way is to localize with the LOS-Anchoruse maximum likelihood detection. This method requires at least three LOS-Anchorsto localize (in Two-dimensional environment). In the thesis, many experiments showthat the position estimations with LOS-Anchor are concentrated in a small area. Withmore NLOS-Anchors, the position estimations are more dispersed. An elasticenclosing circle based NLOS-Anchor discrimination algorithm is proposed in thethesis to identify and remove NLOS-Anchors. Localizing with the rest of anchors,then using residual weighting algorithm to the result of estimation, a position estimateof unknown node can be obtained. Compared to the maximum likelihood method andthe weighted residual method, positioning accuracy of the proposed algorithm hasimproved significantly. The algorithm is suitable for both two-dimensional andthree-dimensional environment.3) Measured data stored in wireless networked instruments may be destructed orloss till they are collected. The redundant storage based on replication will cause a lotof storage overhead and communication overhead. The erasure codes based algorithmdistributes and store data with lower security, SST algorithm is proposed according tosecret sharing and Tornado code to encrypt data and distribute in adjacent instruments.Statistical calculations show that, when instrument nodes are failure so that data areunavailable, SST can accurately recover original data with high probability. Comparedwith the storage strategy based on RS code, the restoration of data needs a little higherthe number of nodes, but it can improve the speed of data distribution and storage andreconstruction when the storage data quantity is higher.