A Estimation Model of Missing Value Based on Wireless Sensor Network

Abstract

Wireless sensor network node affected by its own resources and environmental factors, exists the problem of the missing acquisition data. In view of the missing of WSN data, this paper uses ARIMA model and wavelet decomposition to forecasting the missing value, obtain the estimation of missing data in the time series. The estimation model based on time series named TS model. The adjacent node of missing data acquisition nodes as a reference, through the advantage of neural network model in non-linear relationship, sc model is proposed. In this paper, from TC model and ST model, the time series are combined with the space correlation, STC model is proposed. Experiment shows STC model has a good effect on the estimation of missing data.