Research Of Human Fatigue Detection Based On Gait Acceleration

Many factors can lead to fatigue, such as increasing working strength, longer working hours, heavier mental pressure, etc. When tired, human is likely to work inefficiently, which causes safety accidents and healthy problems. Thus, the real-time fatigue detecting of human body attracts more and more research interests recently, especially on the areas of sports training, body building and medical rehabilitation, etc. In this thesis, gait acceleration data are collected and analyzed, and the studies of fatigue gait authentication are organized in two aspects as follows.This research starts of observing the waveform figure of the human body’s anterior-posterior gait acceleration signal. It is obviously different between pretest and posttest. According to the change of the signal in time, we use the correlation coefficient method to get the threshold value for detecting fatigue. The detection accuracy rate is reached up to93.06%, with9testing samples. This method applies to test the same person, and it needs normal gait acceleration signal to comparative judgment with measured data.In order to mining more characteristic parameters of the gait acceleration signal to detect fatigue, the further research carries out by establishing a mathematical model for the gait acceleration signals, and furthermore, we design a corresponding algorithm. The gait characteristic parameters, such as gait cycle, cadence, the RMS of acceleration, autocorrelation coefficients, peak-to-peak values, FFT are determined. We calculate all the mean and standard deviations of the gait parameters of pretest-posttest separately, which are used for paired t-test, and then analyzes the changes in the scopes of frequency and time domain. The results show that, under the condition of fatigue, signals stability from each trial of the human body become worse, and the differences between amplitude and frequency of the vertical direction are very obvious. These gait characteristic parameters can be used to detect fatigue in different aspects and levels in the future.The purpose of this research is to show that the gait acceleration characteristics can be used to detect fatigue feasibly. The first aspect of this research gives a specific fatigue detection method, and the second aspect shows an overview of the possibility with more gait characteristic parameters for detecting fatigue. More practical applications of fatigue detection will be researched in the future.