Integrated sensing from multiple wearable devices for activity recognition and dead reckoning

Date created: 
2016-03-09
Identifier: 
etd9464
Keywords: 
Inertial sensors
Wearable devices
Sensor fusion
Human activity recognition (HAR)
Pedestrian dead reckoning (PDR)
Abstract: 

Wearable devices are increasingly prevalent in our everyday lives. This thesis examines the potential of combining multiple wearable devices worn on different body locations for fitness activity recognition and inertial dead-reckoning. First, a novel method is presented to classify fitness activities using head-worn sensors, with comparisons to other common worn locations on the body. Using multiclass Support Vector Machine (SVM) on head-worn sensors, high degree of accuracy was obtained for classifying standing, walking, running, ascending/descending stairs and cycling. Next, a complete inertial dead-reckoning system for walking and running using smartwatch and smartglasses is proposed. Head-turn motion can derail the position propagation on a head-worn dead-reckoning system. Using the relative angle rate-of-change between arm swing direction and head yaw, head-turn motion can be detected. The experimental results show that using the proposed head-turn detection algorithm, head-worn dead-reckoning performance can be greatly improved.

Document type: 
Thesis
Rights: 
This thesis may be printed or downloaded for non-commercial research and scholarly purposes. Copyright remains with the author.
File(s): 
Senior supervisor: 
Edward Park
Department: 
Applied Sciences: School of Mechatronic Systems Engineering
Thesis type: 
(Thesis) M.A.Sc.
Statistics: