Fitness Activity Recognition on Smartphones Using Doppler Measurements

• DOI: 10.3390/informatics5020024
• Publication Date: 2018-05-04T07:08:21Z
• Authors: Biying Fu; Florian Kirchbuchner; Arjan Kuijper; Andreas Braun; Dinesh Vaithyalin...
• Keywords: Smartwatch

Quantified Self has seen an increased interest in recent years, with devices including smartwatches, smartphones, or other wearables that allow you to monitor your fitness level. This is often combined mobile apps use gamification aspects motivate the user perform activities, increase amount of sports exercise. Thus far, most applications rely on accelerometers gyroscopes are integrated into devices. They have be worn body track activities. In this work, we investigated a speaker and microphone smartphone exercises performed close it. We active sonar Doppler signal analysis ultrasound spectrum not perceivable by humans. wanted measure weight bicycles, toe touches, squats, as these consist challenging radial movements towards measuring device. tested several classification methods, ranging from support vector machines convolutional neural networks. achieved accuracy 88% for 97% toe-touches 91% squats our test set.