Autonomous driving has a software problem and human driving a distraction problem. Which of these can be solved first is the trillion dollar question.
Following the Uber and Tesla accidents last month, thousands of articles and blogs have been written comparing the record of AI-driven vehicles with human drivers. Let us be clear, human drivers are involved in too many accidents, but they also cover many miles.
However the commentary about the Uber and Tesla accidents overlooks a crucial comparison: Human drivers have short attention spans, slow reaction times and good situational awareness; AI computers have infinite attention spans, fast reaction times and poor situational awareness. It doesn’t matter how many tens of thousands of dollars of sensors and GPUs are fitted to an autonomous vehicle if the AI software cannot understand the context of what it sees and react accordingly.
The Uber crash footage presents this issue both graphically and tragically. Autonomous driving thus has a software problem and human driving a distraction problem–which of these can be solved first is the trillion dollar question.
Humans are not best suited to the repetitive and typically uneventful nature of driving. Through a combination of factors such as boredom, distraction and fatigue, some drivers wind up passing countless miles on the freeway looking at their cellphone or entertainment system–or driving while drowsy–all the while supposedly in control of two tons of metal traveling at sixty miles or more an hour.
Here’s another statistic: 90 percent of all light vehicles in use on our roads and highways have no automated or assisted driving features at all–otherwise known as SAE Level 0. So rather than jumping straight to AI, Semicast advocates making human drivers into better drivers, for example with the mandatory installation of autonomous emergency braking and steering (AEB/AES) and camera-based driver monitoring systems from companies such as FotoNation, Seeing Machines and Smart Eye.
Aftermarket solutions from suppliers such as EDGE3 Technologies and Guardian also exist for this technology to be installed in trucks and buses. These changes alone would be a great start to reducing fatalities and making our roads and highways safer. Better still, these technologies are cheap enough for immediate mass market adoption, are well proven and have no liability issues.