Algorithms for gait measurement with 3-axes accelerometer/gyro in mobile devices

An apparatus for characterizing movements of a subject using a gait pattern, comprising:

• a mobile device having a tri-axial accelerometer located approximately at a center of gravity of said subject and configured to measure using the tri-axial accelerometer at a predetermined data sampling rate, acceleration data in x, y, and z directions of the center of gravity of said subject;
• and
• one or more processors configured to:
• receive the measured acceleration data;
• identify a gait cycle from the measured acceleration data, which comprises:
• identifying a time of a heel strike which corresponds to a maximum downward velocity of the center of gravity by performing a integration on the vertical-direction acceleration data, and
• identifying a time of a toe-off which corresponds to a minimum height of the center of gravity by performing a double integration on the vertical-direction acceleration data;
• wherein the gait cycle comprising at least one stride or at least two steps of said subject, and
• wherein said at least one stride or at least two steps comprise at least one heel strike followed by at least one stance phase and at least one swing phase;
• time-mark, based on the measured acceleration data, two or more key events in the gait cycle, the key events comprising the at least one heel strike and at least one mid-single leg support;
• convert the measured acceleration data to a time dependent power spectrum, the conversion comprising computing a Dynamic Instability Index (DII) at each time-marked event in the gait cycle and constructing tine variations of DII to generate the time dependent power spectrum;
• extract from the measured acceleration data a phase relationship between accelerations along different orientations at each time-marked event in the gait cycle;
• and
• generate a graphical representation on the phase relationships between accelerations along different orientations;
• wherein the gait pattern of said subject comprises the time dependent power spectrum and the graphical representation of the phase relationships between accelerations along different orientations.