Nowadays, wearable devices embedded with huge sensors can distinguish human behaviors outside the laboratory setting through health and activity monitoring. We focus on Human Activity Recognition (HAR) with wearable devices, which is one of the most eminent areas because of a great potentiality to be extended in diverse domains, such as ambient assisted living and human-computer interaction. Most of previous works in HAR only exploited the sensing input from inertial sensors, such as accelerometer, gyroscope, and magnetometer, to understand human activity. Not many works adopt the physiological sensors for the HAR system even if they are already included in many wearable devices. Physiological sensors, such as the Photoplethysmogram (PPG) and Electrodermal activity (EDA), can define the physiological responses of an individual towards a stimuli event. We believe that the fusion of ineritial and physiological sensors can enhance the stability and robustness of the HAR system.
