Vibration Signal Analysis And Detection Method Based On Smartphone Motion Sensor Sampling

The Earthquake Early Warning System(EEWS)can quickly release early-warning information to people who lived in the risk area before the arrival of the temblor,prolongs the emergency reaction time,and effectively reduces the number of casualties and property losses.So far,earthquakes are detected by traditional seismographs such as seismomenters in China.Although they can sample high-fidelity seismic signals,it is difficult to widely deploy due to limitations such as the high price,large power consumption,bulky,and difficuly in installation.In recent years,motion sensors such as accelerometers and gyroscopes equipped smartphones which are widely adopted by humans,can sample seismic waves as the similar way to those of seismometers,potentially become the the devices for monitoring earthquakes.In order to solve the problem that the ordinary smart phone motion sensor may have large noise and poor precision in the detection of abnormal vibration events.This paper attempts to implement,test and improve the classical P-wave phase pick-up algorithms on ordinary smart phones,simulates sampling a large number of abnormal vibration events,analyzes the signal characteristics of the sensor sampling,and proposes a useful abnormal vibration event detection method.Secondly,in order to simulate the intelligence perception characteristics of massive smartphone users,this paper proposes a corresponding data fusion method and uses multiple smart phones for verification,which effectively improves the accuracy of abnormal vibration event detection and provides a reference for the application of ordinary smart phones in earthquake early warning systems.The main work of this thesis is as follows:Firstly,analyze the waveform characteristics of abnormal vibration event data collected by motion sensors of ordinary smartphones,and establish seismic phase models that describe non-customized sensor sampling under dynamic environment and noise conditions.Then,adopt and improve three typical seismic signal monitoring methods were used and improved to analyze and detect nine types of abnormal vibration events and find the most effective vibration signal monitorig methods.The results show that the STA/LTA algorithm of picking up results is lagging behind the true arrival time,but it is close to the true arrival time,and the error is less than 0.1s.The AIC criteria algorithm of picking up results is in advance of the actual arrival time,and it is diffetent from the real one,which is not apply to the detection of abnormal vibration signals based on sensors of ordinary smartphones.The STA/LTA method combined with the AIC criterion method has a detection error of less than 0.05 s,and has the best picking-up effect.It can be used to detect abnormal vibration events based on sensors of ordinary smartphones.Secondly,the data fusion method of abnormal sensor event monitoring by multiple smartphone sensors is analyzed.Using ordinary smartphones from different manufacturers to collect abnormal vibration signals from different scenes,and the method of combination of STA/LTA+AIC criterion picking up P waves,then generate local decisions,and then the local decision is transmitted to the base station for data fusion of the collected results of multiple smart phones.Pick up the results for data fusion.The experimental results show that data fusion greatly improves the accuracy of picking up,reduces the false alarm rate,and effectively improves the system's perceived quality of abnormal vibration events.