Abstract
Over the last two decades, technology has evolved fast causing mass evolution in every aspect of life, among which health care domain and human-machine interaction. However, the benefits of the technology progress were limited for its elderly users, since the traditional human-machine interfaces were always a barrier between them and any new devices. Compared to those interfaces, multimodal interfaces offer a more natural way of interaction with the machines, coping with some inherent difficulties of elderly people. The paper presents a system dedicated to assist people with special needs, such as elderly people or people recovering from specific diseases, to perform rehabilitation exercises by using a multimodal interface. The system presents to the user a set of required exercises prescribed by the doctor or by the caregiver, and tracks the user exercise evolution, informs the user about his/her performances and mistakes, changes exercise levels depending on the users' physical condition and results. Furthermore, the system gives advice to its user to optimize the results of the exercise. To be easily accessible by elderly people, the interface is a key-factor of the system. The supervision of exercises is achieved by means of a Kinect camera, successive images of user's exercises being analyzed automatically. The feed-back to the user is given either by messages displayed on the screen or through voice output. The user is represented on the screen by an avatar, which is mirroring the user's movements during the exercise. The user can interact with the system by giving simple voice commands or by using the touch facilities of the screen. In this way, it allows the user to interact with the system without interrupting his exercise and in the most natural possible way. The interface can be adapted to user preferences by configuring the type and color of fonts, the arrangement of icons on the screen and the aspect of the display. After repeated interactions with the system, the icons corresponding to the most used commands given, either by touch or by voice, will become bigger on the screen or better placed. We claim that our proposed approach will help the elderly to perform their rehabilitation exercise in their preferred medium with a grade of independence and to improve the efficiency of the exercise and well-being. |