Design of a Smart Alert System Based on Electroencephalography Signal Analysis

Abstract The rapid evolution of technology provides us with diverse opportunities to enhance our lives and well-being, addressing essential aspects such as socialization and health analysis. Expanding on this potential, utilizing brain-computer interface (BCI) would allow us to explore and improve aspects like attention deficits. Distractions present persistent challenges to sustained focus across various aspects of life, potentially resulting in compromised academic performance or risks to road safety. This shows how life-changing it would be to design an alert system that boosts efficiency and safety in these areas by targeting to minimize attention losses. By analyzing electroencephalography (EEG) signals associated with concentration levels, the proposed system aims to deliver timely alerts, prompting users to refocus when attention falls below a predefined level. Consequently, avoiding prolonged distractions and encouraging a greater self-awareness of the issue. This research aims to create a comprehensive warning system by combining EEG technology with deep learning techniques. The initial phase involves non-invasive data collection using a 16-channel EEG cap, complemented by Fast Fourier Transform (FFT) analysis to extract features linked to active and passive tasks. During this phase, adhering to the guidelines of the Office for Human Research Protections (OHRP) and the relevant university department is essential to maintain ethical standards and safeguard participant confidentiality and privacy. The collected data will then be used to write a Python code that employs deep learning to identify parameters indicative of various attention levels. The software will utilize this data to set an attention range and send alerts to an external device, notifying when the user has lost focus. Additionally, the system will exhibit intelligent recognition of recurrent short concentration periods, suggesting breaks to prevent mental fatigue. As the project advances, there is potential to enhance the system's capabilities by exploring signal classifications, particularly emphasizing evoked signals associated with external stimuli.