加速度波形を用いた歩行の識別

概要

論文の詳細を見る
We have attempted to classify the acceleration signal in walking both level and up and downstairs using wavelet analysis. The acceleration signal close to the body's center of gravity was measured while the subjects walked in a corridor and up and down a stairway in their own shoes. The data for four steps were analyzed and Daubecies 3 wavelet transform was applied to sequential data. The discriminate variables were the waveformsre lated to leve1-4and-5. The sum of square values at each step was compared at level-4 and -5. Downward walking could be discriminated from other types of walking showing the largest value for the level -5 calculation. Then level walking was compared with upstairs walking for level -4 calculation. It was possible to discriminate the continuous dynamic responses of walking by wavelet transform.
ライフサポート学会の論文