Postural Transition Detection using a Wireless Sensor Activity Monitoring System
Advisor Information
Jong-Hoon Youn
Location
Milo Bail Student Center Ballroom
Presentation Type
Poster
Start Date
8-3-2013 9:00 AM
End Date
8-3-2013 12:00 PM
Abstract
Mobility health is an important aspect of the overall health status of a person. Many tests exist that determine the mobility health of a subject, but there are several issues associated with these tests. As such, a great deal of work has been done to develop a mobility classification system that not only consolidates these tests, but also does not have any of the associated issues. Even so, many of these systems in development are complicated, and lack the ability to track important mobility related measurements. In particular, many systems in development disregard postural transitions and postural transition time as mobility measures. The goal was to not only remove the human error associated with observational mobility health tests, but to make the system as simple, energy-efficient, and inexpensive as possible. In addition, we wanted this system to be able to detect with accuracy of over 90% six mobility states and postural transition frequency and duration. The goal to keep the system simple and inexpensive was accomplished by using a single wireless waist-mounted triaxial accelerometer. The goal to keep the system energy-efficient was accomplished by processing data on the sensor. This decreased the amount of information to be transmitted to the base station, thereby significantly decreasing energy consumption. The goal to detect six mobility states in addition to postural transition frequency and duration was accomplished by the activity classification algorithm used. The original accuracy goal was surpassed, and the system is able to detect mobility states with up to 98% accuracy.
Postural Transition Detection using a Wireless Sensor Activity Monitoring System
Milo Bail Student Center Ballroom
Mobility health is an important aspect of the overall health status of a person. Many tests exist that determine the mobility health of a subject, but there are several issues associated with these tests. As such, a great deal of work has been done to develop a mobility classification system that not only consolidates these tests, but also does not have any of the associated issues. Even so, many of these systems in development are complicated, and lack the ability to track important mobility related measurements. In particular, many systems in development disregard postural transitions and postural transition time as mobility measures. The goal was to not only remove the human error associated with observational mobility health tests, but to make the system as simple, energy-efficient, and inexpensive as possible. In addition, we wanted this system to be able to detect with accuracy of over 90% six mobility states and postural transition frequency and duration. The goal to keep the system simple and inexpensive was accomplished by using a single wireless waist-mounted triaxial accelerometer. The goal to keep the system energy-efficient was accomplished by processing data on the sensor. This decreased the amount of information to be transmitted to the base station, thereby significantly decreasing energy consumption. The goal to detect six mobility states in addition to postural transition frequency and duration was accomplished by the activity classification algorithm used. The original accuracy goal was surpassed, and the system is able to detect mobility states with up to 98% accuracy.