Localization Using Multisensor Fusion of IMM Fixed Lag Smoother in a Cricket Sensor Network

Abstract

We propose an approach for estimating the location of target moving over a time delayed sensor field via a distributed interacting multiple model fixed-lag smoother in conjunction with multiple pseudo measurements obtained using a trilateration positioning algorithm. Proposed pseudo measurements that indirectly measure the target position are defined using the results of trilateration positioning method, in a Cricket sensor network, as single nodes only can measure the distance to the target, the nodes need to form groups of three to be able to perform trilateration. Also, since the trilateration method ignores distance-measurement errors, adequate error covariance information is estimated for pseudo measurement model using a self-tuning method. Each distinct group makes local pseudo measurements with a time delay and fused using matrix weight fusion. Based on these algorithms, localization performance enhancement of a moving target is achieved.