Affectiva’s biometrics identifies driver behavior like drowsiness, anger

Updated Dec 28, 2018

Roughly 30 years ago I watched in amazement as a pickup with a boat in tow slowly drifted off U.S. 98 near Mexico Beach, Fla. and crashed into the woods. Thankfully, there were no major injuries, but the driver was left with the lifelong embarrassment of having fallen asleep behind the wheel. Some folks aren’t nearly as lucky.

While today’s lane keeping systems definitely help there’s also the growing field of biometrics which can identify important clues even before a driver loses control. Enter Affectiva. The company’s Human Perception AI detects behavior like a driver’s head dropping, yawning, irregular eye movement (i.e., texting and driving), anger and joy. Dangerous behavior can be quickly addressed to avoid disaster.

Following Affectiva’s Automotive AI presentation at the LA Auto Show last month, we followed up with them to learn how their technology might fit within the commercial vehicle space.

The Massachusetts-based company is packed with plenty of M.I.T. grads and others eager to make life safer for all of us on the road. Special thanks to Gabi Zijderveld, chief marketing officer of Affectiva, for taking the time to answer questions.

HWT: How can Affectiva products help make driving safer among Class 1 – Class 8 commercial vehicles?

Zijderveld: Commercial van, truck and bus drivers spend countless hours on the road, driving up to hundreds of miles at a time. With this in mind, it’s no surprise that many drivers experience drowsiness and fatigue while behind the wheel. In fact, a study conducted by the Federal Motor Carrier Safety Administration found that fatigue was a critical factor in 13 percent of all accidents involving large, commercial vehicles. At Affectiva, we are developing technology that can help reduce that statistic.

Affectiva is the pioneer of Human Perception AI: software that can detect nuanced human emotions, complex cognitive states, behaviors, activities, interactions and objects people use. Built on deep learning, computer vision, speech science and massive amounts of real-world data, Affectiva’s technology can be used in fleets to detect signs of potentially dangerous driving behavior, including drowsiness and distraction, and alert the driver to curb the behavior before it’s too late. For example, if Affectiva’s AI detected that a driver was drowsy, the vehicle could prompt the driver to pull over and rest at the nearest truck stop. Alternatively, if it detected the driver was distracted, it could send an audio or visual alert to remind him or her to keep their eyes on the road.

HWT: Have any commercial truck and/or van fleets used Affectiva biometrics tech yet? If so, please list these products and how they were used.

Zijderveld: Affectiva is currently working with fleet management solutions provider GreenRoad Technologies to create safer driving in fleets. Specifically, GreenRoad is using Affectiva’s AI to enable real-time driver impairment monitoring. With the help of AI, GreenRoad is able to detect signs of drowsiness and distraction by measuring key indicators such as yawning, blink rate, eye closure and head pose. And in the event dangerous driving behavior is detected, the system can alert the driver in real-time to prevent accidents. Affectiva Automotive AI is also able to help with change management and coaching to encourage better and safer driving habits and improve overall fleet safety.

HWT: How can fleet managers effectively utilize data gathered from Affectiva products?

Zijderveld: The data gathered from Affectiva’s AI can be used for ongoing coaching, and to improve the safety of the fleet as a whole. For example, at GreenRoad, the company records incidents related to dangerous driving behavior. These video clips are then uploaded to GreenRoad Central, GreenRoad’s cloud-based management portal, for review by fleet managers and drivers. From there, GreenRoad’s change management methodology allows for a continuous self-correction cycle across all safety-related incidents, improving the overall safety of the fleet over time.

HWT: How has driver feedback been thus far with Affectiva biometrics tech? (For instance, any concerns expressed over constant monitoring of facial and vocal expressions?)

Zijderveld: Many commercial drivers are already accustomed to some form of performance monitoring as a necessary aspect of their role, and all drivers are explicitly informed of such systems in their vehicles. But it’s also important to call out that Affectiva’s AI does not actually store data on an individual’s emotions and cognitive states. Rather, fleet management systems like GreenRoad’s, only record data related to specific events such as harsh braking, hard cornering, rapid acceleration or deceleration. In the event that GreenRoad is storing more information than merely tracking these driving events, it’s typically only snippets of video that are kept by the fleet management system. These snippets can actually help drivers defend themselves when their driving is called into question, or can be used to coach drivers to improve their technique. Aside from these instances, the footage is tossed out and not reviewed at all.

HWT: Any auto segments where you think Affectiva products will prove most beneficial?

Zijderveld: In addition to fleets, Human Perception AI will enable OEMs and Tier 1 suppliers to build cars that address critical safety concerns for light consumer vehicles. According to recent studies, one out of every ten crashes is caused by drowsy driving and there are 1,000 injuries daily in the U.S. involving distracted drivers. Already, Human Perception AI is powering advanced safety features to combat these issues, and also meet the requirements of industry regulations such as the European New Car Assessment Programme (Euro NCAP). Affectiva’s technology, and the safety systems it powers, will continue to scale with evolving safety standards and future of mobility needs.

Affectiva’s Human Perception AI also has implications for the ridesharing sector – especially as self-driving vehicles become mainstream. As companies like Uber and Lyft deploy fleets of autonomous vehicles, it will be of the utmost importance to monitor what’s going on inside the vehicle, to ensure that riders feel comfortable and safe. For example, if a vehicle’s in-cabin AI sensed that riders were scared or nauseated, the vehicle could adapt its speed or driving behavior accordingly.

Beyond safety, Human Perception AI will redefine the future of the occupant experience in vehicles. Human Perception AI can detect occupants’ emotions and cognitive states, as well as how many passengers are in the vehicle, who they are, how they are interacting with each other and the in-cabin environment, and more. In turn, all players in the automotive space – including commercial fleets, ridesharing companies, and auto manufacturers – will be able to design intelligent vehicles that can adapt the transportation experience according to individual’s preferences and needs for a particular ride.