A grey system based missing sensor data estimation algorithm

Nowadays, the Internet of Things (IoT) are attracting more and more researchers´ attention. To develop a practical and popular IoT system, there are still a lot of key issues to be solved. One of them is that how we can estimate the missing sensor data, which may be caused by bad weather conditions or failure of sensors. To solve this problem, in this paper, a new sensor data estimation algorithm (GSEA) which is based on grey system model GM (1, 1) is proposed. Firstly, according to the correlations between the target sensor which has missed values and its neighbor sensors, we serialized these neighbor sensors by the method that the one with a higher correlation value will be arranged nearer to the target sensor. We calculate these correlation coefficients mostly based on the distance between the failure sensor and its neighbors. Next, we construct a grey system model GM (1, 1) based algorithm to estimate the missing sensor data. At last, the experiments prove that our approach is accuracy and effective.