Prediction Revision Strategies for Data Transmission in Wireless Sensor Networks

In this paper, a novel strategy for data transmission that is based on Prediction Revision Dynamic Adjustment data gathering algorithm (PRDA) is proposed in WSNs. The key idea of the PRDA is to separate the data prediction and model computing, and the autoregressive process model is employed for prediction revision algorithm. The model computing of PRDA is conducted by sink node firstly according to sampling data sequence from sensor nodes, then, sink node sends the parameters of model to sensor node. Each sensor node predicts the values of the data with parameters, and then, determines whether the current sampling data are sent out or not according to the comparison results of predicting data and sampling data. The dynamic adjustment mechanism of model computing is used to fit the variety of sampling data. Simulation results show that PRDA is able to reduce the amount of data transmission and lead to more significantly energy saving than the traditional approach.