Hybrid Brain-Computer Interfaces Could Improve Road Safety by Reading Drivers' Minds — Literally
Designed to reduce accidents as an additional input to an intelligent driver assistance system (IDAS), this hBCI approach shows promise.
A team of engineers at the Beijing Institute of Technology and the Beijing Machine and Equipment Institute has come up with a means of figuring out whether a driver means to trigger hard or soft braking — using a hybrid brain-computer interface (hBCI).
"The work is valuable for developing human-centric intelligent assistant driving systems to improve driving safety and driving comfort," claim co-authors Longxi Lui, assistant professor, and Jiawei Ju, research assistant, of the team's work, "and promote the application of BCIs."
Designed as a means of reducing road-traffic accidents, which are responsible for a claimed 1.35 million deaths and up to 50 million injuries each year, the system is designed to provide additional data to an intelligent driver assistance system (IDAS) — specifically, determining when the driver intends to brake, and whether they intend for hard braking or soft braking, and triggering the action earlier.
To do so, the team turned to reading the driver's mind — literally — using a hybrid brain-computer interface (hBCI), which uses an electroencephalograph (EEG) sensor to read brain waves along with an electromyograph (EMG) sensor to monitor muscle activity. A range of capture models and sensor-fusion approaches were tested on human subjects driving a simulated vehicle, and the results proved promising: An approach dubbed "hBCI-SE1," a sequential approach with spectral features and a one-vs-rest classification strategy that prioritizes EEG signals over EMG, hit an impressive 93.37 percent average accuracy.
There's work still to be done before the system reaches a commercial IDAS implementation, however — not least of which is that subjects were required to wear an EEG headset, EMG sensors around their bodies, and even clip sensors to their ears in a process the team admits is "inconvenient for subjects."
The team's work has been published under open-access terms in the journal Cyborg and Bionic Systems.