The air pollution monitoring by sequential detection of transient changes

According to the World Health Organization, the air pollution is a serious problem that threatens the human health and the environmental balance. The goal of this paper is to study the on-line detection of the air pollution augmentation by monitoring the level of the micro particles concentrations in the outdoor air. The proposed statistical approach is tested by using the real PM10 pollution data, which are provided by the air monitoring stations located in Strasbourg. First, an autoregressive model with exogenous variables is designed to describe the PM10 concentrations of several stations. Next, a special sequential change detection test is used to detect a transient change in the PM10 concentrations. Two versions of this test are proposed: the centralized Finite Moving Average (FMA) test and the decentralized FMA test. The criterion is based on the minimization of the missed detection probability provided that the false alarm rate is upper bounded. The theoretical analysis and the comparison of the proposed approaches on real data are discussed.