The topic of this thesis is the implementation of a ’Motion-based distance education and monitoring system’, which is able to record a series of movements and by accomplishing these recorded moves you can verify your performance without having an expert beside you. I have seen many opportunities in creating this system because it is a frequently reappearing problem that the trainer is not able to show up at the training. That is why I have chosen this topic. A system like this can be used widely, let us just think of the education of different sports, practising physiotherapy at home or dance lessons. The trainer just records the movements at his own home (or office), the recorded exercise is immediately downloadable for the trainees and they can practise for the next lesson.
To implement a system like the one mentioned above I needed a motion sensor that follows and recognizes our movements, then transmits the data to the appropriate program. Then the program can compute the difference between the user’s movements and an other predefined series of movements. By displaying the result in a form which is readable to the user we get a feedback of our performance. From the several options of motion sensors I have chosen a widely available but functionally still excellent one.
In the late autumn of 2010 Microsoft released a motion-based controller (henceforth also mentioned as Kinect), which had a significant effect on the console market, because this was the first time that the audiance could afford a brand new controller, which is basically different from all other controllers available. Though Nintendo also released its motion-based controller (Nintendo Wii), Microsoft’s product is unique considering that there is no need to hold anything in your hands, your body itself is sufficient for controlling the applications. As a matter of fact you are the controller.
The sensor was originally released to be compatible with Xbox, but later new drivers were created in order to make it compatible with PC also. In my thesis I used the OpenNi drivers released by PrimeSense to connect the Kinect with PC.
During the implementation I was keeping an eye on the application to be user friendly, so a non-professional person could also enjoy every function of the program. By the end of the semester I managed to accomplish all the specifications. I created a cohesive system that is able to connect different profiles for different users, register and handle these profiles, and modify them if needed.
Users connecting their Kinect to their PC and launching the application are able to create their own exercises which are stored in a database. These exercises are accessible to other users all over the world since after the creation they are immediately uploaded to the web-server. Everyone is able to download these exercises and for better education, it is allowed to upload your own videos. After accomplishing the downloaded exercises, stored results can be viewed on the webpage.
The modern controlling system and sharing results on the web are strong foundations for the program to become a successful product in the future. The application created under this thesis is expandable by defining new movement types (three types have been defined so far: dance, physiotherapy and martial arts). Since the goal of the thesis was not the fancy view but to demonstrate the working of the system, more users could be attracted by adding a nice graphical surface.