Research On Joint Time Synchronization And Localization Algorithms In Sensor Networks

Time synchronization and localization are two key technologies in the sensor networks.Traditionally, the links between time synchronization and localization in sensor networks havebeen ignored. As a result, these two problems have been investigated separately for a longtime. However, in view of the signal processing, the sensor node’s time synchronization andthe source localization have very close relationships and share many aspects in common. Inview of this, from the perspective of signal processing, this paper intensively andsystematically analyzes some key technologies of joint synchronization and localization underthe condition of anchors with different parameter random errors. This study improves andenriches the theoretical framework of time synchronization and localization techniques insensor networks, and strongly supports the time synchronization and localization engineeringin sensor networks.Firstly, according to the time and space characteristics of the nodes in sensor networks,three major factors that breaking the performance of joint synchronization and localization insensor networks have been given, namely time of arrival (TOA) noise, anchor clock errorsand anchor position errors. For the presence of TOA measurements, this paper established thejoint synchronization and localization model in presence of different anchors parameters, andthen analyses the mathematical solving methods and evaluation methods for estimationperformance.Secondly, considering these factors, the precision problems of joint synchronization andlocalization in the present of known TOA noise, anchor position errors or anchor clock offseterrors has been deeply and systematically analyzed, respectively. We derive the Cramér-Raolower bound (CRLB) of the source position and the source clock offset in correspondingscenarios, and then quantifies the degree of estimation performance degradation due to anchorposition errors and anchor clock errors. A mean square error (MSE) analysis is conducted andit shows that ignoring the errors in anchor positions and anchor clock offsets would lead toestimation accuracy worse than the corresponding CRLB.Finally, this paper proceeds to develop a new joint time synchronization and sourcelocalization algorithm, and then extends it to corresponding scenarios of sensor nodesparameter errors. The new proposed algorithm based on the two stage weighted linear leastsquare method, and has the advantage of being a closed-form solution. Theoreticalperformance analysis proves that the proposed algorithms can attain the corresponding CRLBaccuracy under some mild noise condition. Computer simulations are used to corroborate thetheoretical derivations and demonstrate the good performance of newly developed algorithms.