Human Activity Recognition with Smartphone Inertial Sensors Using Bidir-LSTM Networks

Owing to its extensive applications in solving real-life, human-centric problems, human activity recognition (HAR) has become an important research area in pervasive computing. Most Human activity recognition researches are based on multi-sensor based approaches which use three or more sensors attached on the different parts of human body. However, many sensors were redundant for specific motions. The more sensors being used, the less convenience the users have. The current generation of portable mobile devices incorporates various types of sensors that open up new areas for the analysis of human behavior. Smartphone can be used for the activity recognition of rehabilitation, aged care and so on. In this paper, we propose a bidirectional LSTM structure for human activity recognition using time series, collected from a smartphone (with the accelerometer and gyroscope embedded on the phone) worn on the waist of human, the best recognition accuracy of which can reach 93.79%.