Have you ever wondered how many people are not free to move by their own depend on someone? Have you ever wondered how much frustration these people have because of sitting on a wheelchair?
We aim to provide these people a simple solution that allows them to move freely without needeng another person or even without needing their arms.
The idea is pretty simple: it consists of developing a cost-effective wheelchair controlled by the brain.
As you may know, our neurons in the brain work by firing small electrical currents, allowing smart communication between themselves. Since we are able to capture these small currents (known as EEG), it is possible to record a huge amount of information of the brain.
To do so, we have designed and 3D printed an EEG cap that allows to place a couple of electrodes and record these signals.
Once all the electrodes are placed in the regions corresponding to the motor cortex, we need to convert the voltage difference measured to a digital signal that a computer can process.
Most of the devices used to acquire these signals and send them to a computer are developed by enterprises with commercial purposes. This does not only makes the price increase, but also reduces the compatibility with other softwares, because only a few of them are open-source.
Since one of the main goals is to develop a cost-effective open source system, we got rid of commercial EEG devices and we are developing our own acquisition system.
In order to provide a simple but replicable solution, we decided to build an Arduino-Based EEG system, which provides two main qualities to our project: on one hand Arduino is the most used and spread electronic board, and on the other hand it is relatively easy to code (which enables other people to build their own systems based on ours).
Once we have acquired EEG signals we need to process them in order to extract useful information corresponding to the desired direction of movement. Several nonlinear signal processing algorithms can be used so as to extract the most useful information. Their complexity and performance is different one from the other, which provides a high degree of freedom to chose the best balance between accuracy and optimization.
One of the most challenging aspects of the whole framework is to process acquired data in real-time and continue acquiring new signals simultaneously. To do so we are implementing specific and reliable algorithms combined with a pre-defined model obtained with ground-truth labeled data. We use advanced classification (machine learning) algorithms learned from the best teachers in the world.
We'll be able to read patients' thoughts and control a motor with this information. Join our ambitious journey!
This project involves not only a single person but a team of biomedical engineering undergraduates with strong background and broad knowledge about these combined fields.
Risks and challenges
The main threat regarding the final execution of the project remains on the challenge of being relly useful for disabled patients. In fact, the complexity of the technology is not extremely complicated, which lowers considerably the risk of not achieving our goal. We can't find any threats considering rewards, since everything in this project is cheap and open-source!