Recognition of Arm & Elbow Exercises using Smartphone's Accelerometer

Abstract

Human activity recognition using Smartphone’s sensors is a growing area now a day. This study is concerned with health monitoring and typically recognized arm and elbow exercise activities with the help of Smartphone’s accelerometer. The recognized arm and elbow exercises are: Bicep Curl, Active Pronator, Active Supinator, Assisted Biceps, Isometric Biceps and Isometric Triceps. The data were collected by placing Smartphone at two positions, i.e. “at wrist” and “in hand”, using supervised approach. Twenty (20) volunteers (ten male and ten female) were engaged for the experiment. Each participant performed these activities approximately 20 minutes and total dataset includes around 400 minutes time. Various algorithms based on literature were used for the recognition of defined activities. Results show that Smartphone’s accelerometer can be used for the recognition of arm & elbow exercises, which can further be extended for the application of stroke and injured patients.