The new Drowsiness Detection feature from Samsara uses comprehensive AI models trained on its large-scale data set to detect signs of drowsiness.
The system triggers real-time cabin audio alerts for drivers and notifies managers via text or email to triage fatigue-related events. These insights can be viewed as reports within the Samsara platform, enabling managers to decode patterns of fatigue across a fleet, focus on driver coaching and ultimately improving safety and efficiency.
The National Safety Council reports that drivers are three times more likely to be in a crash if they are fatigued, and according to the AAA Foundation for Traffic Safety, more than 17% of all fatal crashes involve a drowsy driver. The commercial trucking industry in particular is prone to long hours and unpredictable road conditions, leading to drowsiness. Although AI advancements and machine learning have made proactive alerts possible, “drowsiness remains an incredibly nuanced behavior to train AI models for detection”, said Samsara.
“It’s hard to detect when someone is truly drowsy. It’s more than a single behavior, like yawning or having your eyes closed. Drowsiness can be less common than other risky driving behaviors, so accurate detection is only as good as the data that feeds and trains AI models,” said Evan Welbourne, VP of AI and data at Samsara.
To ensure accuracy, Samsara’s drowsiness detection system is trained to consider several behaviors that indicate fatigue, in alignment with leading and clinically validated standards for defining drowsiness. These behaviors include head nodding, slouching, prolonged eye closure, yawning, rubbing eyes, and more.
Yawning alone is often not a sufficient detector of drowsiness. Samsara first announced Drowsiness Detection at its annual Beyond Conference in June, which hosted more than 2,000 physical operations leaders from the industry.
Samsara’s petabyte-scale data set collects more than 10 trillion data points each year and is used to train AI models that automate workflows, accelerate time to value and provide personalized, actionable insights for customers.
