Greg Norman, from the University of California San Diego talked about using smart phones for monitoring, categorizing and recording physical activity. Sedentary lifestyle is leading to obesity.
Scientists have developed and used a number of commercial devices to track physical activity but these are single purpose devices that track physical activity but do not give any feedback to the user. The vision his team had was using a smart phone to detect and track physical activity and also to communicate and motivate the user to increase their physical activity. They really wanted a one device solution.
The first attempt at this used an iPod Touch. The lack of a cell phone plan made the device particularly economical. They created a program to grab gyroscope and accelerometer data from the iPod and then use that information to categorize activity in to one of 9 categories:
- slow walking
- normal walking
- brisk walking
- normal upstairs
- normal down stairs
- brisk upstairs
- brisk downstairs
Overall they were able to accurately deduce the activity with 90% accuracy (kNN classification).
They had difficulty with the iPod: when the application was not front and center, the resolution of the data dropped. The authors then restarted using the Android platform. They had even better results on Android, though they attributed some of that to reducing the number of categories from 9 to 6.
Joel Topf, MD
eAJKD advisory board member
Check out all of eAJKD’s coverage of Medicine 2.0 here.