Abstract:

The purpose of this paper is to present the design and development of an application as a serious game that aims to help elderly people to learn how to perform physical exercises in order to maintain a healthy lifestyle in their homes - the higher the age is, the lower the availability for physical exercises become.

Nowadays, most physical exercises programs for elderly are effectuated under medical surveillance in a specialized hospital facility. However, a personalized program has as main advantage a higher flexibility, as it offers the patient the possibility of adjusting his training schedule according to his preferences and daily activities. In this sense, there is need for a sustainable platform designed specifically for performing physical exercises at home that would promote increasing mobility, especially for elderly people. An efficient method to accomplish this purpose is through the development of a cheap, accessible and easy to use a serious game, based on the motion capture technology.

By using the Kinect motion capture device, our application analyses the user’s movements and compares them to those of a trainer character, by taking into account the performance and efficiency parameters. The exercises that need to be performed by the patient are automatically selected according to his profile, medical condition and acquired progress, based on a complex algorithm. In addition, the application generates reports and statistics concerning the user’s health progress and training schedule, constantly informing the doctor about the patient’s evolution. All these information can be viewed through a user personal account. The intensity level of the exercises are adapted in accordance with the real-time performances of individuals. The system’s architecture is composed of six interrelated modules that ensure the application’s reliability: motion capture, movement analysis and feedback, activity control, animation control, user interface and progress management.

A comprehensive example of use case is the following: the user introduces the login credentials and the application saves them automatically at request. When the motion capture sensor is connected, the activity controller loads the exercises program, informing the user both auditory and visually about the start of the training. The trainer character starts performing the exercises at a certain speed and the patient is required to do the same exercises simultaneously. Based on the recording captured by the Kinect sensor, the feedback analysis module compares the two movements, returning the speed difference and the movement replication accuracy. The user is therefore informed about the results and they are sent to the progress management module to be stored. If the exercise has not been performed correctly, the system will automatically choose another type of exercise to be performed at a lower speed.

Our system has been tested with several users who acquired a certain degree of progress. Therefore, we conclude that the presented solution can be successfully used for stimulating the physical activity and improving the mobility of elderly people, representing a great contribution to the field. Future developments will include a latest generation motion capture sensor and providing options, such as the possibility of performing exercises with a remote partner via a network or adapting the intensity level of the exercises based on the medical health parameters of the user.