A multi-sensor multi-rate algorithm for motor rehabilitation with Augmented Reality devices

Using Augmented Reality (AR) could offer stimuli to rehabilitation from neuro-motor disorders, since the patient can be aided in a better known reality than Virtual Reality (VR). The main goal for an AR system is to achieve a good quality of tracking the real object to align with virtual contents. Often a single sensor could not provide enough information to that end due to a low updating rate; therefore joining an other high updating rate sensor could be indispensable, but how to combine data from different sensors especially when they work all at different rates? In this paper an approach based on recursive parameter estimation, focusing on multi-rate tracking in AR devices is suggested. The system of this study has a multi-sensor configuration provided by a webcam and a 3-axis MEMS inertial sensor, both working at different sample rates. Also an augmented reality application, designed for the neuro-motor rehabilitation of the upper limb, has been developed and shown.