Page 171 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

proper predictions. Thus, the total training time is an hour, broken down into three 20-minute
sessions.

Use Case Status: In development Process: 4.1-Healthcare

Partners: None

2�2 Benefits of Use case

Project Outcomes: The wheelchair uses EEG to improves the patient's mobility, leading to
imporved health and quality of life.

The technology mobilizes the paralyzed, helping them have more of a social life by eliminating
reliance on others.
The brain-computed wheelchair is a groundbreaking technology, which combines assistive
technology and neuroengineering.

2�3 Future Work

For future work, we plan on expanding the functionality of the EEG wheelchair to meet more
demands. Some of the future implementations for our device:

• Safety features: More safety features, like proximity alert or emergency stop, will be added
to the functionalities.
• Integrating EOG: We plan on integrating EOG technology as well to track eye movement
as well to help them in many more aspects than just locomotion.
• Linking up with IoT: While our device currently serves its basic purpose to assist people
with mobility, upon integration with IoT, we will be able to present a device with unlimited
capabilities, from cooking/cleaning/household work to covering all the tasks that need to
be done for these people.
• Thermal Throttling: A thermal throttling system can be implemented to ensure that
the system adapts to temperature changes. The system can be added using a simple
temperature sensor added to the L298N, interfacing it with the Arduino, with a code to
monitor the temperature levels. Measures to reduce the heating or to augment system
activity can be taken, e.g., a cooling fan, to ensure the smooth operation and durability
of the machine.

3 Use Case Requirements

REQ1: It is critical that the system accurately acquire raw brainwave (EEG) signals from the
user's motor cortex, as per the international 10-20 system, and efficiently preprocesses these
signals to remove noise and artefacts for reliable interpretation.

REQ2: It is critical that the system employ a trained machine learning model to classify
preprocessed EEG signals into discrete wheelchair commands (e.g., forward, backward, turn
left, turn right, stop) with a high degree of accuracy and minimal latency.

REQ3: It is critical that the system translate generated commands into real-time wheelchair
movement, ensuring smooth, responsive, and predictable control to enable the user to navigate
their environment effectively.

135