Advanced sensor technologies and machine learning are increasingly improving the ability to “see" and understand the world around a vehicle. But what about the situation inside the car? With driver distraction known to be a major cause of road accidents, there is substantial potential for improving safety by incorporating driver monitoring into Advanced Driver Assistance Systems (ADAS).
According to the World Health Organization, traffic accidents are among the top 10 causes of death in the world. In the USA alone, NHTSA reports that there were 3,166 “distraction-affected" road fatalities in 2017. In Europe, safety ratings agency Euro NCAP plans to introduce driver monitoring into its criteria for primary safety assessment starting in 2020.
“Accident figures show that even a small lapse in attentiveness can have tragic consequences. HARMAN's Driver Monitoring system can play a key role in helping drivers stay focused at the wheel."
“Accident figures show that even a small lapse in attentiveness can have tragic consequences," says Bernhard Pirkl, Vice President of ADAS/AD at HARMAN. “As part of our extensive ADAS capabilities, HARMAN's Driver Monitoring system can play a key role, not only in helping drivers stay focused at the wheel, but also in mitigating the consequences of distraction. The system uses sensors to capture data, enabling it to assess the state of the driver's momentary attentiveness. We can also adapt the warning and intervention levels to 'very capable and highly attentive, focused driver' – which helps with user acceptance of these advanced features."
Known as first-order biometrics, the most important of these include direction of gaze, head position and pupil diameter. Second-order biometrics, such as heart rate and breathing measured by auditory sensors, provide indicators of emotional activity and cognitive load.
But what does this mean in real terms?
“If it detects driver distraction, fatigue or emotional overload, the system can instigate any number of safety measures ranging from warning signals or prompts for the driver to take a break to automatic adjustment of cabin temperature, music and more," says Pirkl.
Using sensor fusion, the system cross references this data with other inputs from the ADAS spectrum, including the vehicle's motion sensors, cameras, radar and LiDAR to determine the need for physical intervention with braking or steering maneuvers. In simple terms, if the driver's gaze is diverted and heart rate elevated, and the vehicle is moving erratically, he/she could be arguing or distressed. “This is not a good state of mind for being in control of a vehicle," adds Pirkl.
Occupant monitoring offers benefits in terms of determining the other dynamics taking place in the car that could impact driver performance. For instance, the system detects when the driver is emotionally stressed or distracted because of passenger activity in the vehicle. It can then suggest a relaxing music playlist or activate soothing ambient lighting to stabilize the situation.
Looking to the future, systems such as HARMAN's Occupant Monitoring are a crucial aspect in the development of autonomous driving. Level 3 autonomous cars, also known as conditionally autonomous, are able to manage most driving tasks themselves. However, if certain conditions are not met, they require the human driver to take over. “Maintaining awareness of driver and cabin status is a critical element of the handover process and will be key to the success of semi-autonomous vehicles in future," says Pirkl.