Open access repository

Home Open access repository

In 2014, we launched our open-access repository which offers full text access to conference proceedings from many of our events including the INC and HAISA series. These papers are free to access and distribute (subject to citing the source).

» Openaccess proceedings » Eleventh International Network Conference (INC 2016)

Eleventh International Network Conference (INC 2016)

Eleventh International Network Conference (INC 2016)
Frankfurt, Germany, July 19-21, 2016
ISBN: 978-1-84102-410-3

Title: Using Machine Learning Techniques for User Specific Activity Recognition
Author(s): Jaideep Chawla, Matthias Wagner
Reference: pp25-29
Keywords: human activity recognition, body area networks, inertial sensors, wearables, Machine learning, classification
Abstract: This paper explores the possibility of using wireless
sensor networks and a machine learning based approach to
classify activities performed by the wearer. The network consists
of inertial sensors connected to a wrist worn module which sends
data to the smartphone using Bluetooth Low Energy(LE) and
the inertial sensors present in the smartphone themselves form
a node of the network. The motivation behind this structure is
to develop a Human Activity Recognition(H.A.R.) system which
causes minimal to no inconvenience to the user. Movement data
(acceleration and orientation) was collected from different users
for activities ranging from jogging and running to fitness based
activities like push-ups, squats and features were extracted from
the raw data. The extensive list of features was reduced using
a correlation based subset evaluation method. The performances
of four different classification algorithms namely the K Nearest
Neighbor, Support Vector Machine, Artificial Neural Network
and Classification and Regression Tree(CART) was evaluated
for classification accuracy. Classification was performed using the
features extracted in the previous step. Classification accuracy in
excess of 90 percent was obtained which points to the possibility
using such a system which trains itself specific to the movement of
a user for Real Time Human Activity Recognition while causing
minimal inconvenience to the user.
Download count: 768

How to get this paper:

Download a free PDF copy of this paperBuy this book at

PDF copy of this paper is free to download. You may distribute this copy providing you cite this page as the source.