Selective Predictors of Environmental Parameters in Wireless Sensor Networks

This paper investigates multiple alternatives to predict environmental parameters (temperature and relative humidity) inside an intelligent container with the aim of the supervision of cool chains. A wireless sensor network will help us to measure those parameters in some distributed points of the closed space. To achieve the aim of fault detection and also total energy saving in the sensor network, there will be several possibilities. This research deals with prediction of future values of the environment in a few specific sensor nodes by using some of active sensor nodes. It inspects several requirements of the predictors to pick out the most applicable identification styles among ARX, ARMAX, OE, BJ and SS. It employs some key-sensor nodes (KSNs) as predictors of the parameters in a few desired sensor nodes (DSNs). The DSNs either have already been turned to sleeping mode to reduce batteryconsumption or deactivated due to energy depletion. Keywords— Prediction, model, temperature, relative humidity, air flow