With the recent development of body sensor technology,
portable device has been widely used for monitoring purpose.
Emerging device is needed to provide a solution of monitoring
gait pattern beyond lab environment to enable pervasive
sensing. Data analysis using accelerometers has risen to be
a suitable tool for pervasive application of motion analysis.
The demand for extracting meaningful gait features from
accelerometer data is a key technique for such purpose and
has grown rapidly recently. E-AR sensor is a small device
can be worn on ears with 3D accerlometer sensor which
give measurement of acceleration in medio-lateral, anteriorposterior,
and vertical direction. In this paper, we propose
a feature extraction method for ear-worn sensors to obtain
meaningful gait parameters for quantification of gait patterns.