Driving Toward Vehicle Automation
“Many of the technologies we are using for automated driving are actually active safety systems,” says Andrew Whydell, providing an explanation as to why automated driving is at the cusp of realization. As Global Director, Vehicle Systems Product Planning, ZF Group TRW Active and Passive Safety Division (zf.com), Whydell has seen the proliferation of safety systems throughout the vehicle, as well as calls from regulators and safety organizations to make some of the necessary technologies standard equipment. As they have matured and tied in to other units in order to increase system capability, they have been joined by new features that make autonomous control of a vehicle possible.
“Stability control is already mandated here in the U.S.,” says Whydell. “Electric power steering is coming to help CAFE compliance, and Automatic Emergency Braking [AEB] has been added to NCAP [New Car Assessment Program]; you will have to have some form of AEB in order to get a five-star rating. So the foundations for automated driving are coming into the vehicle anyway, with the key parts—sensors, actuators, cameras, and radar—for things like AEB providing the information necessary to support automation.”
Which is not to say that arriving at this point has been easy. AEB development has been driven by pressure build-rate levels far faster than anything seen in a braking system previously. Electronic stability control systems require 50 bar (725 psi) of pressure in 200-250 milliseconds, and the unit is decelerating just one wheel. AEB is different in that it is decelerating all four wheels simultaneously, while applying 100 bar (1,450 psi) of pressure in much less time, about 150 milliseconds.
“There are two parts to any of these new technologies,” says Whydell. “First, what does the customer think it’s going to do? Second, what does the system actually deliver?” The confusion over Tesla’s Autopilot system capabilities neatly illustrates this situation; the name implying an automation capability beyond that supplied by the hardware and software. According to Whydell, “The consumer is confused about AEB. Does it mitigate the crash, as is the case with some systems, or prevent the crash, as in others?”
Mix in the fact that the Insurance Institute for Highway Safety (IIHS) will be testing and rating these systems, giving a higher score to those that are more aggressive, and you have a potential minefield for automakers and consumers alike. Less aggressive systems may have fewer false alarms, while more aggressive ones have more. Automakers will have an incentive to not stray too close to either extreme, but consumer expectations will have to be managed at both the OEM and retail level through a thorough communication of the system’s real world capabilities and a suitable man-machine interface created. That’s a problem facing driver assistance systems (DAS) of all types.
Whydell recounts the results of a Dutch government study of DAS technologies that showed consumers were fond of automatic cruise control, feeling it was calming and relaxing. Lane departure warning was another thing, however. Stress levels increased in those vehicles that beeped or buzzed or vibrated the driver’s seat when the car looked like it would be straying out of its lane. Yet cars fitted with a torque overlay system that helped steer the car back toward the lane centerline were much more well liked.
“The problem isn’t the technology,” says Whydell, “but the interface with the technology. It’s the difference between having someone ask if they can help carry your packages, and someone taking them without asking. One is much more pleasant.” And more discrete. With a guidance-style interface, no one else in the vehicle needs to know what warnings are being given.
It will take time to work through these challenges, but vehicle automation is coming both more and less quickly than expected. Though the industry is pushing the idea of fully automated driverless pods whisking passengers from Point A to Point B in the near future, the reality is somewhat different. Moore’s Law has created affordable processing power that will allow automated mass market vehicles. Mandates and rating incentives are pushing adoption of new active safety and fuel economy technologies. Suppliers are extending the capabilities of low-cost sensing technologies to bring these capabilities to more vehicles. With each automaker looking to steal a march on the competition by making these technologies available across the board (Toyota has announced that, by the 2018 model year, AEB will be standard in almost every car it sells in the U.S.), even the lowliest vehicle could be capable of traffic jam assistance or a level of automated driving. “This doesn’t mean, however, that we’re going to be traveling in driverless modules in the foreseeable future,” claims Whydell.
“Even if the industry introduced totally autonomous pods tomorrow, it would take 25 years to turn over the passenger car fleet in the U.S., and 36 years to do the same with the light truck fleet,” says Whydell. That means it would take until 2042 at the earliest to get all of the old cars off the road. In the interim, Whydell says, each iteration of DAS technology will be better than the one it replaces, and drivers will come to appreciate the ability to let the car do most of the work in traffic jams and other stressful situations—and turn it off when they wish to take control. However, the driving force behind vehicle automation will continue to be reducing traffic deaths. Says Whydell: “The best way to protect drivers and passengers is to prevent the crash in the first place.” Thirty percent of the accidents in Europe are single-vehicle crashes where the vehicle runs off the road and rolls over, etc. In the U.S., however, these accidents account for more than 60 percent of the total due, in large part, to the fact that more people live in rural areas and half the population drives light trucks that are dynamically less stable than a passenger car. “If we can prevent the vehicle from drifting off the asphalt and onto a soft surface,” says Whydell, “we have a really good chance of addressing those numbers.”