Brain Machine Interface Software Application for Data Collection, Thought Analysis, and Robotic Arm Control

The overall purpose of the ongoing Brain Machine Interface (BMI) project is to develop an electroencephalography (EEG) interface and a robotics control application which will further enable people with disabilities to achieve autonomy. The project consists of developing, building, and testing an end-to-end system to translate raw EEG data into actionable information. This can be used to control a robotic arm and for other research purposes. A BMI is a system that collects the brain’s electromagnetic signals by utilizing sensors, extracts meaningful signals from the data, classifies thoughts, and ultimately uses thoughts as an input to a computer system. The computer system then has the ability to control hardware and software, which for this project is a robotic arm. The team improved the robotic arm user interface, developed a graphical user interface (GUI) for thought recognition, and explored future research paths by partnering with local experts. To improve the usability of the robotic arm user interface, the team developed software that allows easier performance of useful activities, such as using a pen to play tic-tac-toe, playing piano, and picking up objects. The Insight headset by Emotiv was used by the team for data collection. The headset can stream real time EEG data and control signals, however the Emotiv software solution for data collection is closed and proprietary. To use the Vision Lab’s noise reduction and muscle signal removal algorithms, the team created a GUI to train the thought classification system and collect and process the data. EEG phoneme detection is a future research path that allows for thought to speech translation. The team investigated EEG phoneme detection by implementing algorithms which can identify phoneme sounds from audio recordings. Using these working algorithms, further research will implement phoneme detection using only EEG signals with no audio.https://ecommons.udayton.edu/stander_posters/2325/thumbnail.jp