Mobile Node Localization Using Fusion Prediction-Based Interacting Multiple Model in Cricket Sensor Network

This paper addresses an approach to estimating the location of a mobile node based on the range measurements of Cricket sensor network (CSN), where the coordinates of the mobile node are calculated via the method of trilateration. There are, in general, two kinds of obstacles to be tackled and overcome in CSN: One is noisy distance measurements, and the other is the low data rates of Cricket sensors. To overcome these problems, we propose a fusion prediction-based interacting multiple model (FPB-IMM) algorithm. The FPB-IMM algorithm utilizes multiple position measurements produced by trilateration and a self-tuning algorithm; it takes advantage of these multiple measurements to minimize the effect of noisy measurements and the low data rates by modifying a cycle of IMM with fusion prediction. The experimental results demonstrate that the proposed algorithm outperforms existing algorithms such as the Kalman filter and the conventional IMM.

I. Introduction

Localization plays an important role in the application of robotics, signal processing, sensor network, and so forth. It helps robots understand where they are located at each particular time interval and helps robots plan their path and various tasks, including target localization and tracking in image processing and biomedical engineering. Up to now, various methods for indoor and outdoor localization techniques have been proposed [1]–[10].