Research And Implementation Of The Counting Step Algorithm And The Calculating Stride Length Based On Acceleration

With development of society, more and more people pay attention to their health, the walk and run is a convenient and effective way to exercise, so the demand is growing for the accurate pedometer. The walking and running information of people is obtained by the accelerometer sensor. The obtained acceleration combining with the advisable algorithm can calculate the steps and distances. In this context, this thesis after researching the pre-existing counting step algorithm puts forward new counting step algorithm based on it and the calculating stride length algorithm. Now the pre-existing counting step algorithm uses the acceleration of having apparent cyclical changes to accurately count steps, such as the acceleration of waist in walking and running. The thesis puts forward the new counting step algorithm on the basis of a thorough research into the counting step algorithm of using the waist acceleration. The new counting step algorithm improves the method of calculating threshold and adds the function of searching step and identifies the walking state, so it can use the unstable arm acceleration to accurately count steps.This thesis constructs a dynamic threshold formula of new counting step algorithm by analyzing the relation of acceleration and threshold in walking and running. The thesis selects different people for walking test under controlled conditions to obtain a large number of acceleration and records the real number of steps. Then according to the acceleration and real number of steps, the thesis deduces the ideal threshold corresponding to different accelerations. Then the thesis uses the Mathematical statistics to analyze the relation of acceleration and threshold, search their inherent law and establish the regression equation between them, this is the dynamic threshold formula.The calculating stride length is combined with the counting step algorithm in the thesis, the stride length is got by calculating the forward direction acceleration of double integrals in the time domain. And the thesis uses a correction factor to get rid of interference components in the acceleration measured by sensors, so getting the accurate stride length.The thesis uses the large number of data to test accuracy of the counting step algorithm and calculating stride length algorithm. Those data are tested by the different gait conners walking and running with different pace. And one of those the accuracy of the counting step algorithm presented by this thesis using the acceleration of arm to count steps is above 97%, meanwhile, the accuracy of calculating stride length using the acceleration of arm is above 90%. The results show that the counting step algorithm and calculating stride length can accurately calculate the number of steps and distance of the body’s normal gait.