Real-time safety system to assist drowsy drivers

Machine learning and biotechnology offer unprecedented opportunities to improve road safety and vehicle performance.

Driver fatigue and distraction accounted for 28% of fatal car accidents in Australia last year. Public education campaigns are currently the only method of addressing this problem, however a new study taking place at UTS Tech Lab aims to change this. Researchers are applying their knowledge of human-machine interaction to develop a real-time safety system that will alert drivers to the early warning signs of fatigue, stress and distraction.

A $1.5 million 6-degrees-of-freedom immersive driving simulator – the first of its kind in Australia – is located in the Motion Platform and Mixed Reality Lab at UTS Tech Lab. Combined with physiological measurements, the equipment allows the team to assess driver cognition and predict their performance when faced with stress, fatigue or distraction in a controlled environment.

Wearable sensors are used to collect biodata from the driver. Every millisecond, brain dynamics, eye movement, heart rate, face and hand position are measured, providing a real-time assessment of driving behaviour. Machine learning techniques are applied to uncover correlations in the data and produce an algorithm that indicates when a driver’s concentration is impaired.

This algorithm will form the basis of a real-time safety system that aims to support drivers based on their cognitive state. If the system detects that a driver is tired, disoriented, motion sick or stressed, it will trigger an audio warning or, in the case of self-driving vehicles, may even switch into autonomous mode.

Applications of the system go beyond next generation driving, with potential benefits to OHS in construction, freight, logistics and manufacturing.