A key point in debates revolving around self-driving cars is if the technology should be introduced piecemeal or brought to market only when all the elements are in place for fully autonomous vehicles. What does that mean exactly?
Nearly every new, light vehicle, car and truck, today has some form of automated technology either standard or at least available as an option. Blind spot monitoring and automatic cruise control are probably currently the two most democratized automated technologies. That is, they are the most available on the widest range of vehicles. But, there are literally dozens of others that have been developed in the race to self-driving cars.
Stringing together some of these automated systems, developers have created what the government refers to as Level 2 autonomy. Nissan’s ProPILOT Assist, found on its new Rogue and Leaf, is a Level 2 system that assists the driver, but requires the driver to keep hands on the wheel, maintaining ultimate control. Usually, as with ProPILOT, fail-safes are built into the system to remind the driver to maintain control; going as far as bringing the vehicle to a stop if the driver doesn’t respond.
The real issues begin, however, with Level 3 and Level 4 automation, allowing control of the vehicle to pass back and forth between the automated system and the driver. Level 3 is conditional assistance in which the automated system monitors surroundings, controlling steering, acceleration and braking in some situations; but the human driver must be prepared to retake control at a second’s notice if the car requests it. Level 4 is high automation in which the automated system performs all functions without human assistance in specific conditions, but not all. A human driver is still necessary to regain control in areas where the system can’t reliably function.
Where all of this is heading is toward Level 5, which is full automation or fully autonomous vehicles. Level 5 won’t require a steering wheel, pedals or even a human to be present in the vehicle.
In the ramp up to fully autonomous vehicles, not only do the hand offs between human and technology in Level 3 and Level 4 vehicles have huge potential safety problems, such as the human driver becoming too complacent, distracted or relaxed to retake control in the tick of a clock. Moreover, though, is the question: When is an automated system really ready for prime time? Is it better to introduce a new automated system into a vehicle when it works relatively effectively or wait until it’s perfected. Is there a safety benefit in doing it sooner rather than later, even if, in its current form, it might only be moderately safer than a human?
The Enemy of Good
Published in book form, the research company Rand Corporation released findings of its study to determine how safe a highly automated vehicle (HAV) should be before it’s put in the hands of a consumer. Entitled “The Enemy of Good: Estimating the Cost of Waiting for Nearly Perfect Automated Vehicles,” the research looked at the difference in fatalities over time when HAVs that are just 10 percent safer than human-driven vehicles are put on the roads, as opposed to waiting for HAVs with significantly improved safety performance of 75- to 90-percent better than humans.
The Key Findings
Rand assumed, reaching the 75- to 90-percent safety improvements over humans will take years. Saving lives, beginning now, with vehicles that are 10-percent safer, which will become progressively safer in the future, will cumulatively save hundreds of thousands of lives over the years that otherwise will be lost in the wait for nearly perfect automated systems.
What it means to you: We are going to need to wait a long time for totally safe self-driving cars. We tend to think of automated technology as fail proof. It isn’t and it won’t be for years to come. As automated cars do become measurably (10 percent or better) safer than human-driven ones, it makes sense to embrace the technology. At least Rand believes so.