Compressed Sensing Based Abnormal Events Detection In Wireless Communication Net-works

With the rapid development of information technology, people's demand for information also meets a huge explosion. In wireless communication networks, abnormal events are common together with the transmission of information. We need to get the information of nodes in the wireless communication networks when detecting the abnormal events in them. Traditional signal processing is based on Shannon/Nyquist sampling theorem, while the huge amount of data may cause a waste of resources, and moreover, the transmission and storage of huge data is also a very complex work.Compressed sensing (CS) is a novel theory which combines the sampling and compressing procedures together. By exploiting the sparsity characteristics, the original signal can be sampled at a rate much lower than the Nyquist rate and still be recovered exactly. The application of CS to wireless communication networks can not only improve the development of information technology, but also give advancement to the theory of compressed sensing itself.This paper first describes the fundamental theory of CS, particularly focusing on the following three aspects:the sparsifying of original signal, the designing of measurement matrix and the reconstruction method. After that this thesis introduces the application of CS to wireless communication networks and some improved theories of CS, and also the basic features and weak points of current research on abnormal events detection, which sets up a primal foundation for future research.In this paper, the main topic is to design an algorithm of abnormal events detection based on CS in the scenario of wireless communication networks. The formal research on this topic usually depends on prior information, while iterative scheme of weighted basis pursuit can efficiently solve this problem, and also control the influence of noise to ensure the accuracy of detection. Considering the asymmetry of anomaly in the networks, this paper proposes a distributed detection method. By the distributed scheme, it can achieve more proper utilization of energy and save much more transmission energy, and does not affect the detecting ability at the same time.In the end, this thesis points up the later work of this topic, as well as some other research topics on the application of CS to wireless communication networks in the future.