Deep Learning Approach of Raw Human Activity Data
Hamdi Amroun, M'Hamed (Hamy) Temkit, Mehdi Ammi · 2018
This chapter aims to study human activity recognition, in an uncontrolled environment, using an iPhone, an Apple Watch and an Apple TV remote. The iPhone and the Apple Watch contain an accelerometer, a gyroscope and a microphone. The Apple TV remote contains an accelerometer and a gyroscope. After extracting sensor data from these three devices, they are then given as input to a deep neural network (DNN) without prior and specific preprocessing of these data, that is, by taking raw data alone. The iPhone is then put in the pants pocket of participants, while the Apple TV remote is held in the hand. The chapter shows 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.