Target Location Estimation Algorithm Based Wireless Sensor Networks

With the significant advances in networking, wireless communications, andmicroprocessors, the topic of wireless sensor networks (WSNs) has become afast-growing research area. Especially the target location that based the WSNs has agreat potential for both military and civil application. A sensor node has small volume,low price, self-organization and good concealment, at the same time it also can bewireless communication and random deployment, therefore it very suitable for targetlocation and tracking. Considering the practical situation, this paper completed the morereasonable description of the research content and the system, so we can further studythe target localization algorithm based the intensity of energy attenuation. This paperpresents the estimation algorithm which combining particle filter algorithm and MLalgorithm, finally realize the multiple parameters target localization based WSNs. Themain contributions can be summarized as follows:①Based on the basic theories of target location technology, the intensity ofenergy attenuation model and the characteristics of the WSNs, this paper preliminaryanalysis the target location and the WSNs, reasonably assume that the basic parameterfor the system model, and completed target location of the abstract model based themaximum likelihood estimation. To analyze the estimation performance, the lowerbounds of the performance as CRLB based on the priori distribution of the unknownparameters is derived, and three heuristic weighted average methods are introduced forcomparing. Simulation results show that this estimator is much more accurate than theheuristic weighted average methods, and it can reach the CRLB even with a relativelysmall amount of data.②The improved particle filter algorithm is presented for the random distributionof target location and the energy intensity attenuation unknown situation duringtransmission. During the research, this paper collected the data which represents thedifferent environment path attenuation impact on target positioning accuracy, andreasonably describe the object and the system, assume the signal attenuation follows theHATA model, in order to make full use of energy and limited bandwidth of WSNs, thereceived data on each sensor node are quantized firstly and the discrete values aretransferred to the fusion center. Finally combining with the particle filter algorithm andthe maximum likelihood algorithm, this paper implements the target location estimate based on the multiple parameter.③Thresholds for the energy-based target localization are designed using aheuristic method. This design method is suitable for uniformly distributed sensors andnormally distributed targets. Furthermore, to save sensor energy, a sensor selectionmethod is presented. The sensor selection method can substantially reduce energyconsumption for our target localization method, and spends less calculation time.，simulation results show that it has a good performance of estimation with a few nodes.④Byzantine identification scheme and a dynamic non-identical quantized designscheme are presented. The two design schemes not only enhances the systemperformance by reducing the estimation error of the target location but also makes itrobust to the presence of Byzantines by making them ineffective when combined withour Byzantine identification scheme.