An accurate and adaptive pedometer integrated in mobile health application

An interesting and valuable application in sensor network technology is Personalised Rehabilitation. In particular, through the combination of a smart phone and wearable sensors, patients can follow and have feedback on their exercise programmes, while enhancing the medical monitoring. In this paper, we design a system to measure the patient's activity by estimating his walking habits. The system, ready to be integrated in a mobile health application, consists of an inertial sensor with a tri-axial orthogonal accelerometer attached to patient's foot, while the sensor is connected to a smart phone for data processing. Additionally, we propose an algorithm for step detection and gait state estimation. The challenge is to provide reliable and accurate detection when pacers are in different gait states. Also, the algorithm will reduce error drifts at start moment when sensor signals are very unstable. Extensive experiment results will demonstrate that the algorithm provides good performance in terms of accuracy, adaptability and memory use.