A new algorithm has been invented by scientists at the University of Sussex that have the ability to teach a smartphone how to not only detect your every move but record it too.  Some smartphones on the markets can already recognize some human activities, particularly sports such as running or yoga, but these still have to be ‘hard coded’ into the device to begin with.

Hristijan Gjoreski / University of Sussex

With the new algorithm, activities are detected in real-time as they happen, and not just during exercising. The new method can detect movements like cutting vegetable or brushing teeth and will even record how long you’re sitting or lying down for. Dr. Hristijan Gjoreski of the University of Sussex said, “Current activity-recognition systems usually fail because they are limited to recognizing a predefined set of activities, whereas of course human activities are not limited and change with time.”

This machine-learning approach is a great way to record human activity in real time and far outperforms any similar approach out there currently. Where traditional models tend to clump together bursts of activities to estimate what the person’s been doing, they don’t tend to account for any breaks, pauses or interruptions. So, if the person were to go for a walk but have two interrupted short stops, it would be calculated as three separate walks.  That’s no longer an issue with the new algorithm.

Dr. Daniel Roggen is head of the Sensor Research Technology Group at the University of Sussex and he says, “Future smartwatches will be able to better analyze and understand our activities by automatically discovering when we engage in some new type of activity.  As well as fitness and lifestyle trackers, this can be used in healthcare scenarios and in fields such as consumer behavior research.”

Research Article Published @EurekAlert

More News to Read