Human daily activity recognition in robot-assisted living using multi-sensor fusion

In this paper, we propose a human daily activity recognition method by fusing the data from two wearable inertial sensors attached on one foot and the waist of the subject, respectively. We developed a multi-sensor fusion scheme for activity recognition. First, data from these two sensors are fused for coarse-grained classification in order to determine the type of the activity: zero displacement activity, transitional activity, and strong displacement activity. Second, a fine-grained classification module based on heuristic discrimination or hidden Markov models (HMMs) is applied to further distinguish the activities. We conducted experiments using a prototype wearable sensor system and the obtained results prove the effectiveness and accuracy of our algorithm.