By Lytx Chairman and CEO Brandon Nixon
The recent headline-making collisions and heart-breaking loss of life related to Tesla Motor’s Autopilot have sparked a long-overdue debate about the co-existence between humans and automated driving technology. The Tesla team is quick to point out that the technology is meant to augment – not replace – the driver, but it’s not that simple.
We’ve been reviewing and tracking driver behavior, and improving driver performance for nearly 20 years and over more than 50 billion miles. Through the depth and breadth of that data and the associated driving events captured on video, we’ve become very familiar with the role human nature plays in driving. Even experienced professional drivers can become lax and vulnerable to unsafe driving habits over time, often with serious consequences, particularly with in-cab distractions. Our data shows that driver inattentiveness – a hazardous by-product of some automated driver-assist technologies -- is found to be a factor in 52 percent of collisions.
We also track which tools developed to date can either overcome drivers’ weaknesses or effectively augment their strengths. And even then, new technologies, new distractions and ever changing environmental conditions may shift that equation, creating a dynamic that can only be anchored by the sole constant: the driver.
It’s our contention that the driver is the most sophisticated technology in the vehicle. Our brains can process so much information about a sudden situation that it manifests as prescience or intuition. And even with all of that processing power, we still make bad choices. It’s not arrogance or ignorance. It’s largely a matter of insufficient training.
Before drivers can understand the implications and responsibilities that accompany the use of any automated driving feature, they need training on the role—and limits -- of that automated technology. Training is an ongoing task when technology continues to evolve and the environment in which it operates is dynamic.
People – drivers – constantly overestimate the abilities of automated technology. Technology does not think. It’s meant to operate under routine conditions, leaving humans to jump in when conditions change. And technology can fail. How many times have you had to reboot your computer?
It’s popular to compare technology running autonomous vehicles to aircraft autopilot technology. Tesla has even captured that in its sub-brand. But the cockpit is not vacant when a plane is on autopilot. In fact, both pilot and co-pilot are engaged, and constantly monitoring for exceptions in the environment, instrumentation, and and so on. They’ve received intense training to remain vigilant in the unlikely event that environmental changes occur or an equipment emergency arises – or both. To them, they’re still flying – it’s just a different way to fly the plane. Because autopilot doesn’t think, the pilots must do the thinking as well as reacting.
Driving, however, is a much more complex task because at any given moment, a vehicle is surrounded by unpredictable factors and the consequences of inappropriately reacting can be swift and severe, as we’ve sadly learned recently. Everything that makes autopilot work in planes is absent in vehicles, and in particular, the training critical to understanding its limitations.
We have video evidence that without proper training, drivers have been inclined to rely on automated technology, like auditory alerts signaling lane departure, rather than staying alert, engaged, and in control of their vehicle. Throw in a distraction like a ringing cell phone, and the act of driving can be completely derailed. Any symbiotic relationship between the driver and the machine can be easily defeated by happenings outside the vehicle, whether another car, an errant pedestrian, or, an animal entering the roadway.
Recently, researchers from Virginia Tech have advocated standards for the human-machine interface. We’d advocate going one step further and strongly recommend the systematic adoption of technology-specific driver training and ongoing coaching, particularly in commercial vehicles deploying any form of automated technology, to ensure that driver behavior remains appropriate to the limits of the technology, and that safety remains the overarching objective of any human-machine endeavor.