Wearable sensors based human intention recognition in smart assisted living systems

Human-robot interaction (HRI) is an important topic in robotics, especially in assistive robotics. Here we propose a smart assisted living (SAIL) system to help elderly people, patients, and the disabled. In this paper, we address the human intention recognition problem and design a hidden Markov models (HMM) based online recognition algorithm to classify hand gestures. The data is collected by a single inertial sensor worn on a finger of the subject. We implemented a dynamic duration segmentation method based on the FFT and investigated the training method related to the recognition decisions and accuracy. Several hand movements are performed to represent different commands. The obtained results prove the effectiveness of our method.