Distance Estimation-Based PSO Between Patient with Alzheimer’s Disease and Beacon Node in Wireless Sensor Networks

Abstract

In recent years, research in wireless sensor networks and their application in health care and environmental monitoring have attracted significant interest. In such applications, the accuracy of the distance estimation between a patient and a beacon node is crucial for determining patient location. In this study, the distance between the mobile node (carried by the Alzheimer’s patient) and the beacon node was measured using the received signal strength indicator (RSSI) with ZigBee technology in indoor environments. The distance estimation was determined by two path loss models: a log-normal shadowing model (LNSM) and a derived model using a polynomial function (the POLYN function) obtained with the MATLAB curve fitting tool. Next, particle swarm optimization (PSO) was merged with the polynomial function (called the PSO–POLYN function) to obtain the optimal coefficient values for the POLYN function. The resulting path loss model can improve the distance error between a patient with Alzheimer’s disease and a beacon node. The results revealed that the merging of the PSO–POLYN model enhanced the mean absolute error (MAE) by 20% relative to the LNSM, where the MAE for distance was 1.6 m for the PSO–POLYN model and 2 m for the LNSM. In addition, after applying PSO, the correlation coefficient (R2) of the regression line between RSSI and the estimated distance improved to 0.99, while that obtained through the LNSM was 0.94. The presented method based on PSO–POLYN outperformed models in the literature, both in terms of MAE and correlation coefficient in indoor environments.