e-Shoes: Smart shoes for unobtrusive human activity recognition

摘要: Many approaches to human activity recognition such as wearable based or computer vision are obtrusive in the sense that they prevent users from performing activities a natural way, might raise privacy invasion concerns. This paper presents e-Shoes — smart shoes for unobtrusive recognition. E-Shoes instrumented with tiny wireless accelerometers embedded inside insole of shoes. The sensors seamless making system suitable recognizing everyday activities. To analyze sensor signals, we propose convolution neutral networks (CNN) model automatically learns features sensing data and makes predictions about We verify effectiveness approach real dataset covers seven daily achieved 93% accuracy average, which is very promising, while being energy efficient easy use.