Design of Smart Vest to Monitor Physical Activities of Children

Monitoring the physical activities of children is a vital task for parents/caretakers. The use of information and communication technologies (ICT) jointly with artificial intelligence and smart devices can help the parents to monitor their children. Thus the regular physical activities of children can be monitored with high precision and less difficulty. In this paper, a smart vest is designed for monitoring physical activities of children which will be able to let the caretakers/parents to remotely monitor the children activities with the inertial sensor embedded in vest. We have proposed a sparse based classification algorithm for activity classification. The wearable vest is built around the LilyPad Arduino as a microcontroller with an inertial accelerometer sensor. The transmission of the data is via Bluetooth module. The performance of proposed system is compared with other standard classifiers like k-NN, SVM and decision tree. The overall accuracy obtained is 95.32%.

I. Introduction

Human activity recognition (HAR) is an emerging area of research in the field of pervasive computing. Human activity recognition is broadly classified into two types namely vision based and wearable sensor based. Vision based [1] involves recognition of activities from videos captured by cameras under well controlled laboratory settings. This method does not produce reliable result due to clutter occlusion etc [2]. Privacy issue is another drawback in vision based method. Wearable sensor based method provides robust action recognition in both indoor and outdoor environment.