Smart wireless health care system using graph LSTM pollution prediction and dragonfly node localization

Wireless sensing networks (WSNs) have been applied on various research applications such as monitoring health of humans, targets tracking, natural resources investigation, air quality prediction, water pollution prediction and radiation pollution. The challenge on predicting these applications still exists. Suitable monitoring systems are necessary, to maintain the healthy society with sustainable growth. With the advancement of Internet of Things and modern sensors, the environmental monitoring systems have become smart monitoring system. These wireless sensors are scattered around the environmental locations and places. The localization of sensor place -ment at the correct place will reduce the redundancy of the sensing environment and cost of the equipment. More nodes are placed at the area that has more pollutant. Accurate node sensor placing on the needed area will reduce the cost of sensors and increase the prediction accuracy. This helps to keep our health safe by selecting less polluted environment. Hence, this article focuses on introducing the deep learning algorithm called Graph Long Short-Term Memory (GLSTM) neural network to predict the air quality characteristics. Next, the evolutionary algorithm called Dragon fly optimizer has been used to localize the node based on the prediction. Deep evolu-tionary based algorithms will improve the air pollutant prediction and node localization sensor cost.

Automated summary

Wireless sensing networks play an important role in environmental monitoring and prediction. This article introduces a deep learning algorithm and an evolutionary algorithm to improve air quality prediction and node localization accuracy.