Target Tracking In Wireless Sensor Networks Based On Quantized Data

Wireless sensor networks(WSN) as an emerging technology, which in the field ofenvironmental monitoring, telemedicine, battlefield surveillance and other filed have a widerange of applications. Target tracking is an important application of wireless sensor network,because transmission bandwidth and energy of each node is limited, in order to extend the lifeof the network, researching target tracking method in wireless sensor networks which cansave energy and bandwidth is of great importance.In view of the wireless sensor network node energy and bandwidth constraints, this papermainly studies target tracking method based on quantized data, and focus on the analysis ofthe particle filter method based on quantized data and the method based on data censoring andquantized innovations for target tracking. For the particle filter method based on quantizeddata mainly introduces the auxiliary particle filter and the cost reference particle filter, the twomethods both can achieve the target tracking in nonlinear system, the difference is that theauxiliary particle filter need to know noise distribution, and the cost reference particle filterdoes not need to know that. The method based on data censoring and quantized innovationsfor target tracking is to use the quantized innovations and the information conveyed bymissing data due to data censoring to estimate the state of the target. Becausethe data information content of all sensor nodes is not the same, this paper introduces the ideaof data censoring, firstly we should set the appropriate threshold, according to the informationcontent of the observation, if the observation is greater than the threshold, then its innovationwill be quantized and after that will be sent to the fusion center, otherwise they will bemissing, because the missing data also conveys useful information, the accuracy of estimationcan be improved by exploiting the information contained in the missing data.This paper analyzes simulation experiments respectively about target tracking usingparticle filter method based on quantized data, and the method based on data censoring andquantized innovations. The simulation results show the two methods can reduce themagnitude of the communication in the network, saves the energy of the nodes and bandwidth,and both of them also have higher tracking precision. Data censoring firstly take a data selection, the retained data which has a greater amount of information will be quantized andwill be sent to the fusion center using particle filter to estimate the target state, which reducesthe amount of communication.