| CPC A61B 5/1116 (2013.01) [A61B 5/1126 (2013.01); A61B 5/6828 (2013.01); A61B 5/7225 (2013.01); A61B 5/725 (2013.01); A61B 5/7257 (2013.01); A61B 5/742 (2013.01); A61B 2562/0219 (2013.01)] | 18 Claims |
|
1. A method for analyzing a posture of a user, the method being implemented using a system that includes a processor, and a sensor unit which is coupled to the processor and which includes an inertial measurement set and an electrical signal sensor set that are to be worn by the user, the inertial measurement set including a three-axis accelerator worn on a leg of the user, the method comprising:
while the user is doing a physical activity, continuously receiving, by the processor, a plurality of original sensor datasets from the sensor unit, each of the original sensor datasets being associated with a specific time instance and including data generated by the inertial measurement set and the electrical signal sensor set;
a) determining, by the processor, a plurality of extreme time instances at which the inertial measurement set was at one of a relative highest location and a relative lowest location;
b) establishing, by the processor, a number of activity periods based on the plurality of extreme time instances; and
c) for each of the activity periods, generating, by the processor, an analysis result with respect to a number of original sensor datasets received within the activity period;
wherein a number N of original sensor datasets are received from the sensor unit at a number N of corresponding successive time instances, each of the original sensor datasets including a front-rear acceleration associated with a front-rear direction, a left-right acceleration associated with a left-right direction, and a top-bottom acceleration associated with a top-bottom direction;
wherein step a) includes
a-1) using a band-pass filter to filter the original sensor datasets individually, so as to obtain a number N of filtered sensor datasets,
a-2) for each of the filtered sensor datasets, calculating a root mean square (RMS) value of the front-rear acceleration and the left-right acceleration,
a-3) using a low-pass filter to filter the RMS values calculated for the filtered sensor datasets so as to obtain a number N of filtered RMS values, and constructing a waveform using the filtered RMS values and the time instances, and
a-4) determining the plurality of extreme time instances by one of
identifying a plurality of crests on the waveform, and determining a plurality of time instances that correspond with the crests as a plurality of top time instances at which the inertial measurement set was at the relative highest location, the plurality of top time instances serving as the extreme time instances, and
identifying a plurality of troughs on the waveform and determining a plurality of time instances that correspond with the troughs as a plurality of bottom time instances at which the inertial measurement set was at the relative lowest location, the plurality of bottom time instances serving as the extreme time instances.
|