An Improvement in Maximum Likelihood Location Estimation Algorithm in Sensor Networks

Abstract

This paper reviews on one of the localization algorithm works based on Maximum Likelihood Estimation Method and tries to improvement the performance of this algorithm using a new method. this localization algorithm is a model-based localization algorithm which could be used to estimate location using RSS when a statistical model is available.

An indoor localization system considering channel interference and the reliability of the RSSI measurement to enhance location accuracy

“…Pre-processing that we propose considers each channel of IEEE 802.15.4 to find an optimal channel. Many localization researches use maximum likelihood estimation (MLE) method [7] ~ [9]. In statistics, MLE is a method of estimating the parameters of a statistical model.…”

“…The MLE-based localization algorithm uses a probability distribution to improve the location accuracy. Researchers in [7] propose a new MLE-based localization algorithm to improve performance. This algorithm includes four steps to estimate location, and doesn't require to any extra hardware and it has the less complexity and results the less RMSE (Root Mean Square Error) than the original MLE.…”

Rethinking of the Uncertainty: A Fault-Tolerant Target-Tracking Strategy Based on Unreliable Sensing in Wireless Sensor Networks

“…Therefore, this kind of methods seem hard to be implemented. Another kind of prevalent tracking strategy is model-based, which is achieved by successively estimating the localization [2], velocity [4][5] and trace [14] of the target with target movement modeling, estimation [15] and filtering [16] [17] (e.g., Kalman filter [18], Particle filters [19], Bayesian networks [20], Variational filter [21]). Most of the model-based methods use time-correlated measurements for localization, i.e., sensors may use the measurements of previous target positions to infer current target location.…”