| dc.description.abstract | Taiwan is located in a seismically active zone and faces the problem of a large number of earthquakes. Many disastrous earthquakes usually cause the loss of lives and properties over the years. Therefore, how to design a series of preventative measures and reduce the risk of damages caused by earthquakes is an important issue. So far we still do not have a better way to get earthquake warning immediately. As a result, the Earthquake Early Warning System (EEWS) would be a useful tool and has become the urgent need of human beings.
Nowadays countries neighboring the Ring of Fire put a lot of efforts and resources on investigating the EEWS. Among them, Japan, The United States of America, and Taiwan have the most abundant achievements. Constructing the EEWS has some common problems. Firstly, the need of the special equipment for seismic wave sensing costs highly. Secondly, the number of earthquake detection stations cannot increase arbitrarily due to the lack of funds. Thirdly, people are not familiar with the EEWS. Until now, people get the earthquake news from television, radio stations or social APPs. The EEWS does not fully come into our lives.
In this thesis, we propose a new algorithm and architecture of the EEWS named as the EQ-system. The hardware part of EQ-system is a G-sensor for detection. The software part of EQ-system is a library of earthquake early warning called LibEQ. LibEQ combines several theories such as Support Vector Machine, Fuzzy Inference and so on. We can build a biggest earthquake detection network with low cost. People can get earthquake alarm by APP. Consequently, for earthquake early warning, EQ-system can produce the best possible results. | en_US |