Making One Size Fit All – Accounting for Human Variability in Wearable Monitoring
Abstract:
Sport and fitness wearables are leading the popular adoption of wearable tech. For a sport wearable monitor to be useful, the metrics it provides must be reliable and accurate. But people differ in how they execute movements, they have erratic and unpredictable behaviours, and sometimes even when the wearable gets it right, it still gets it wrong.
In this talk, Megan will examine how human variability impacts accuracy, and describe solutions to mitigate the impact. Examples will touch on real-life scenarios encountered during the development of the TritonWear swim monitoring system.
Bio:
Megan Holtzman leads the Digital Signal Processing team at TritonWear Inc., where she drives development of the wearable algorithm, and is the resident Metrics Maven.
Her interest in athletic motion monitoring stems both from an academic and athletic standpoint. Receiving her Ph.D. in Electrical and Computer Engineering at Carleton University in 2014, Megan’s focus was on analytics and signal processing from multisensor systems. During her Ph.D., she published or co-authored over 20 papers on activity recognition and movement monitoring from ubiquitous sensors, and won several prestigious awards.
Megan has a keen personal interest in improving sport performance through data analytics. She represented Team Canada at the 2017 ITU World Triathlon Age Group Championships in Rotterdam, and made the podium at the 2006, 2008, and 2014 International Dragon Boat Federation World Club Crew Championships.
