Name: Clement Hiscock
Degree: Biomedical Engineering
Project Title: Development of resources for new Scenario
Supervisors: Eve Hatten
My project focused on developing a system for mouse control using electrooculography (EOG) and testing its validity as an undergraduate scenario. The project really appealed to me as it provided an opportunity to further my research, technical and communication skills.
My role in the project was to try and make a range of EOG mouse systems and collect data on their strengths and weaknesses. I started the project by researching EOG; To very briefly summarise EOG, it works by using electrodes to detect the deflection of the eye. The electrodes detect the deflection of the retina cornea dipole and this is then mapped to deviation of the eyes.
In order to make a functional mouse system we decided that it would need to be able to move the mouse up, down, left, right and to be able to click. With these requirements in mind I created four unique systems.
The first system used a signal conditioning circuit which involved amplification and some noise removal. This first system had the DC component of the signal left in; This meant three channels could provide the five degrees of freedom as one channel could distinguish between opposite movements.
The second system used a signal conditioning circuit which had a high pass filter to remove the DC component of the signal. This provided stability to the signal.
The final two systems I created aimed to overcome the largest problems with the first two. The first system had very clear distinction between opposite movements on the same channel but there were big problems with baseline drift. The second system was very stable but it proved very complicated to garner five degrees of freedom from only three channels as the difference in opposite movements was indistinguishable.
To solve the problem in the first system I introduced a peak detection; This algorithm meant that instead of checking for a voltage threshold it averaged the last few values and then checked for outlier values that signalled eye movement . This meant the code effectively ‘followed’ the signal removing the baseline drift issue
The solution to the issues in the second system was much more simple. I added two further channels so that each movement of the mouse was controlled by a separate channel. This removed the need to distinguish between opposite movements on the same channel.
Eve was a great supervisor; She really gave me the freedom to explore all of the ideas I had without restriction but was also very present when I had issues that I was struggling with. She also provided invaluable insight into the considerations that go into designing projects for students.
I would highly recommend other students to apply for this studentship. It was a great project that developed my research, communication and technical skills. My only advice to other students doing the project would be to keep your approach broad; It is very easy to get locked into your first approach but try and avoid this and keep developing your whole range of ideas!
