2 Deep Learning Approach of Raw Human Activity Data

This chapter proposes to study the recognition of certain daily physical activities by using a network of smart objects. The approach consists of the classification of certain human activities: walking, standing, sitting and lying down. The study exploits a network of common connected objects: a smartwatch, connected remote control and a smartphone, worn by participants during an uncontrolled experiment. The sensor data of the three devices were classified by a deep neural network (DNN) algorithm without prior data preprocessing. We show that DNN provides better results than decision tree (DT) and support vector machine (SVM) algorithms. The results also show that some participants’ activities were classified with an accuracy of more than 98%, on average.

2.1. Introduction

Nowadays, connected objects have become indispensable devices in daily life. The areas of application are broad as they cover the domain of security, home automation and health.

With the development of the Internet of Things (IoT) in daily life, new uses have emerged, paving the way to many perspectives in the field of human activity recognition in terms of use and application.

Many studies have been carried out on the recognition of human activity using standard and portable connected objects [MIL 06]. These methods provide easy-to-use, flexible, and above all lightweight devices and platforms for effective human activity control on a daily basis, and include ...

Get Challenges of the Internet of Things now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.