Gait Identification Using Cumulants of Accelerometer Data

This paper describes gait identification using cumulants of accelerometer data. Accelerometer data of three different walking speeds for each subject (normal, slow and fast) was acquired by a cell phone placed on the person's hip. Data analysis was based on gait cycles that were detected first. Cumulants of order from I to 4 with lags from 0 to 10 for second, third and fourth order cumulants were calculated from the cycles and used as feature vectors for classification which was accomplished by support vector machines (SVM). Six healthy young subjects participated in the experiment. According to their gait classification the average recognition rate was 93.1%. A similarity measure for discerning different walking types of the same subject was also introduced using principal component analysis (PCA).