EEG potentials predict upcoming emergency brakings during simulated driving.

Area of Science:

• Neuroscience
• Automotive Safety
• Human-Computer Interaction

Background:

• Current emergency braking systems rely on external sensors and pedal input, which can be too late for crash prevention.
• A delay in detecting the driver's intent to brake can lead to increased braking distances and potential accidents.

Purpose of the Study:

• To investigate the potential of using electroencephalography (EEG) and electromyography (EMG) to predict a driver's intention to perform emergency braking.
• To compare the predictive accuracy and timing of EEG and EMG-based detection with traditional pedal dynamics.
• To assess the impact of earlier detection on reducing vehicle braking distance.

Main Methods:

• A driving simulator study was conducted to record driver responses.
• Muscle activation (EMG) and cerebral activity (EEG) were measured alongside pedal dynamics.
• A simulated emergency braking assistance system integrated EEG and EMG data for early detection.

Main Results:

• Driver's intention to brake was accurately predicted using both EEG and EMG prior to behavioral response.
• EEG and EMG provided similar predictive accuracy, with EEG being faster than EMG, and EMG faster than pedal dynamics.
• The simulated system using EEG and EMG detected emergency braking 130 ms earlier than a pedal-response-only system.
• This earlier detection reduced braking distance by 3.66 meters at 100 km/h.

Conclusions:

• Neuroergonomic signals (EEG and EMG) offer a promising method for early detection of emergency braking intention.
• Integrating EEG and EMG into driving assistance systems can significantly reduce braking distances, enhancing road safety.
• The characteristic spatio-temporal superposition of EEG components in critical situations is key to the observed prediction performance.