Monitoring walking and cycling of middle-aged to older community dwellers using wireless wearable accelerometers

Changes in gait parameters have been shown to be an important indicator of several age-related cognitive and physical declines of older adults. In this paper we propose a method to monitor and analyze walking and cycling activities based on a triaxial accelerometer worn on one ankle. We use an algorithm that can (1) distinguish between static and dynamic functional activities, (2) detect walking and cycling events, (3) identify gait parameters, including step frequency, number of steps, number of walking periods, and total walking duration per day, and (4) evaluate cycling parameters, including cycling frequency, number of cycling periods, and total cycling duration. Our algorithm is evaluated against the triaxial accelerometer data obtained from a group of 297 middle-aged to older adults wearing an activity monitor on the right ankle for approximately one week while performing unconstrained daily activities in the home and community setting. The correlation coefficients between each of detected gait and cycling parameters on two weekdays are all statistically significant, ranging from 0.668 to 0.873. These results demonstrate good test-retest reliability of our method in monitoring walking and cycling activities and analyzing gait and cycling parameters. This algorithm is efficient and causal in time and thus implementable for real-time monitoring and feedback.