What it does
Wake Up collects EEG signals emitted from the driver’s brain in real time to monitor the level of alertness of the driver. Once the the signals showed patterns of drowsiness, the device will emit warning sound to the driver via bone conduction headphones.
Your inspiration
Our team shares a common passion for improving car safety. Through a preliminary survey, we discovered that many long-time drivers, such as bus or courier drivers, tend to underestimate the dangers of driving while fatigued. This motivated us to explore innovative solutions. With some team members having a background in biology, we were inspired to use biosignals as a way to detect and respond to driver drowsiness in real time.
How it works
The device consists of three main components: an EEG electrode to detect brain activity, a microcontroller for signal processing and bone conduction earphones for output. The EEG electrode picks up raw brain signals, which are first sent to the Ganglion board for basic signal processing and filtering. These signals are then passed to a Raspberry Pi, which runs a custom program designed to recognise patterns linked to driver fatigue. If a fatigue pattern is detected, the device activates the bone conduction earphones to play an alert sound, which is deigned to be enough to awaken the driver gently without causing a sudden shock or panic.
Design process
We began by exploring how to detect drowsiness through blink rate, initially prototyping an EMG-based circuit. However, we faced signal inconsistencies and comfort issues due to the placement of electrodes near the eyes. In parallel, we purchased the OpenBCI Ganglion EEG kit and discovered it could reliably detect blinks through the FP1 and FP2 electrodes. This allowed us to shift fully to EEG-based detection, improving both accuracy and comfort. We iteratively tested electrode placement, developed blink detection algorithms using threshold methods, and refined them to reach 80% accuracy. We also began analysing alpha wave intensity to assess fatigue, and later combined both features into a multiparametric drowsiness detection model coded in Python. For hardware, we first attempted to 3D-print a headset but found it uncomfortable for real driving use. We switched to embedding our EEG system into a cap with an adjustable band, significantly improving wearability. Through constant prototyping and testing, we developed a compact, non-intrusive system that detects drowsiness in real time—both reliably and comfortably.
How it is different
Our wearable device addresses key limitations in current drowsiness detection systems. Most existing products rely on dashboard-mounted cameras to monitor eyelid closure, which often fail in low-light conditions or when drivers wear sunglasses. This makes them unreliable, especially for night-shift drivers. In contrast, our device uses EEG biosignals, allowing for direct, real-time monitoring of brain activity, including subtle signs of pre-drowsiness. Unlike loud commercial alarms that disturb passengers, our design uses bone conduction earphones to deliver a discreet alert only to the driver. The sound and vibration are carefully tuned for effectiveness and comfort, making it ideal for use in shared or ride-hailing settings.
Future plans
We plan to develop a companion app that helps drivers track their driving patterns and gain insights into their drowsiness tendencies, encouraging lifestyle changes to prevent fatigue. Additionally, we aim to integrate machine learning to improve the accuracy of drowsiness detection and even predict fatigue before it occurs.
Share this page on