(G-260) EyeSync: Electrooculography (EOG) Eye Tracking Assistance Technology

Introduction:
People with physical disabilities often face barriers using cell phones. According to the World Health Organization, over one billion people globally have disabilities, many with physical impairments that make traditional cell phones challenging to use. These challenges can lead to feelings of isolation and exclusion resulting from a struggle to stay connected and access information.
The United Nations has set several goals related to improving the lives of people with disabilities, including the Sustainable Development Goals (SDGs). SDG 10 seeks to reduce inequalities and promote the social, economic, and political inclusion of all individuals, including those with disabilities. Furthermore, SDG 4 aims to ensure inclusive and equitable quality education and promote lifelong learning opportunities for all, which can be facilitated by assistive technology such as electrooculography (EOG) controlled devices.
Our project, EyeSync, is an affordable assistive technology that improves touchscreen accessibility for those with motor impairments. Unlike costly camera-based solutions, EyeSync uses electrooculography (EOG) for eye tracking, allowing users to operate a numerical keypad via eye movements. The system uses a five-electrode setup to capture eye movements, transmit signals via Bluetooth to a user-friendly app, and translate the signals into meaningful actions. Current experiments show a 70% success rate of correct translations of eye movements into app response. Future goals include refining performance and expanding applications, like developing a cursor system and an alphanumeric keypad.


Materials and Methods:
Electrodes are positioned on the subject, with positive and negative electrodes for horizontal and vertical eye movement connecting to the non-inverting and inverting nodes respectively of an AD620 operational amplifier, creating a band-pass filter. This filters out both low and high-frequency noise from the outgoing EOG signal.
The filtered signal is then collected by an Arduino Uno microcontroller. The vertical and horizontal filtered outputs are sent to separate analog pins on the Arduino and assigned distinct variables. The collected EOG signal represents a relative voltage change based on the user's eye movements.
By establishing specific voltage thresholds for each direction (up, down, left, or right), the Arduino code detects when these thresholds are reached and transmits the corresponding movement type via Bluetooth to the EyeSync app.
The EyeSync app features two main pages displaying a numerical keypad for user interaction. Each cluster of numbers on the keypad corresponds to a specific eye direction. After initially selecting a cluster by looking in the corresponding direction, a secondary screen appears. The user then focuses their gaze on the desired number on the secondary screen to make a selection. Afterward, they are returned to the main keypad to make further selections. This process can be repeated for as many numbers as needed.
Finally, a 3D enclosure was designed on a computer-aided design (CAD) software to accommodate all circuit elements, ensuring safety and ease of use of the device.


Results, Conclusions, and Discussions:
In testing EyeSync, a voltage change corresponding to eye movements was identified as integral for successful operation. This change was made discernible via the Arduino microprocessor's coded 3-second delay, ensuring that re-centering of users' gaze did not trigger unwanted system responses. Through rigorous trials, EyeSync achieved 72% detection accuracy for left and right eye movements, 78% for upward eye movements, and 44% for downward movements. These results, as confirmed by a one-sample proportion test, underscore EyeSync's promising potential as an assistive communication tool for individuals with physical disabilities—with tweaking to the downward movement function.
In EyeSync's development, several challenges surfaced, including loose electrode attachments, unstable wiring, and involuntary eye movements which introduced errors. Addressing these, we integrated a band-pass filter to reduce DC and low-frequency noise, and a 3.3 V DC offset to capture a broader range of EOG signal. We also devised a novel power solution using two portable chargers, enhancing the device's portability.
In terms of further improvements, we plan to integrate a headband with metal electrodes, known for superior signal conductivity. This change aims to improve signal strength and detection accuracy, especially for downward movements. Another envisioned development is the inclusion of a cursor system, broadening EyeSync's functionality to control a wider range of digital interfaces.
PCB refinement is another key area of focus. A more efficient, compact, and reliable PCB design is expected to boost performance, reduce size and power consumption, and potentially extend battery life and portability. A long-term goal is to integrate EyeSync with smart home systems, offering users control over their environment through eye movements.
Our commitment is to continually improve EyeSync. Through enhancing user comfort with adjustable headband and ergonomic electrodes, and incorporating machine learning for pattern recognition, we aim to create a personalized and enhanced user experience. EyeSync is not merely a project; it's our contribution to creating an affordable, user-friendly platform that democratizes digital accessibility.


Acknowledgements: We would like to thank our lab professor and coordinator, Dr. Olukemi Akintewe, for providing the lab space and sharing her invaluable knowledge, which was instrumental in the successful completion of this project. We would also like to thank Bryce Fuller for his exceptional assistance in circuit design and troubleshooting throughout the project. Additionally, we are thankful to Randall Buck for his contribution in designing the enclosure for the circuit in SolidWorks.

References:
[1] Q. Abbas, “Intro to xcode cloud - 2022,” YouTube, 27-Nov-2022. [Online]. Available: https://youtu.be/bNvUPb6K5jI. [Accessed: 23-Apr-2023].
[2] S. Abdullah, A. Ahamparam, E. Dulzo, and S. Ismail, “Electrooculography Controlled Cursor,” University of Michigan Deep Blue Libraries, Apr-2021. [Online]. Available: https://deepblue.lib.umich.edu/bitstream/handle/2027.42/169565/EOGCapstoneFinalReport_Abdullah_Samina.pdf?sequence=1. [Accessed: 23-Apr-2023].
[3] S. Yathunanthan, L. U. Chandrasena, A. Umakanthan, V. Vasuki, and S. R. Munasinghe, “Controlling a wheelchair by use of EOG Signal,” 2008 4th International Conference on Information and Automation for Sustainability, 2008.
[4] T. A. Team, “Liquid Crystal Displays (LCD) with Arduino,” Arduino Documentation. [Online]. Available: https://docs.arduino.cc/learn/electronics/lcd-displays. [Accessed: 04-May-2023].
[5] “Experiment HM-3: Electroculogram Activity (EOG) - iworx.com.” [Online]. Available: https://www.iworx.com/documents/LabExercises/EOG.pdf. [Accessed: 04-May-2023].
[6] “Experiment HM-3: Electroculogram Activity (EOG) - iworx.com.” [Online]. Available: https://www.iworx.com/documents/LabExercises/EOG.pdf. [Accessed: 04-May-2023].
[7] M. M. Alam, M. M. S. Raihan, M. R. Chowdhury, and A. B. Shams, "High Precision Eye Tracking Based on Electrooculography (EOG) Signal Using Artificial Neural Network (ANN) for Smart Technology Application," 2021 24th International Conference on Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 2021, pp. 1-6, doi: 10.1109/ICCIT54785.2021.9689821.
[8] Y. Jia and C. W. Tyler, “Measurement of saccadic eye movements by electrooculography for simultaneous EEG Recording,” Behavior Research Methods, vol. 51, no. 5, pp. 2139–2151, 2019.