Self-driving cars: I want to believe

Departments - Editor’s page

Subscribe
July 13, 2017
Robert Schoenberger Editor || rschoenberger@gie.net

If you have any connection to motor vehicle technology, you’ve likely heard these questions: Do you think we’ll actually have self-driving cars available in our lifetime? Will delivery drivers be out of work if trucking companies automate driving? Do you think the technology is really going to work?

Autonomous driving has caught a lot of attention as companies invest millions into developing sensors, controllers, and algorithms to take over driving functions. Tech companies are buying autonomous-driving startups in hopes of participating in the next big thing in transportation technology.

But that last question I keep getting at parties is the most important – will the technology really work?

Here are two things I know about that question – automakers, commercial truck companies, tier suppliers, and technology companies are investing heavily to make the answer yes. And dumping billions of dollars into a technology does not guarantee that it will ever be viable.

Look at hydrogen fuel cells. Several companies have test fleets on the road, you can lease fuel-cell-powered Hondas in California, and automotive experts are saying the technology is about 10 years away from commercial viability. With some minor tweaks, that previous sentence would have been true 30 years ago. A 2002 article in Wired magazine talks about General Motors’ billion-dollar bet on hydrogen. Congress passed legislation in 1995 to push commercialization of the technology. Earlier this year, Toyota, Daimler, and BMW joined a consortium pledging to invest $10 billion during the next five years to make hydrogen fuel cells viable.

Despite these massive investments, hydrogen power remains 10 years away, and the finish line doesn’t appear much closer today than it did in the 1970s when R&D efforts began.

Autonomous mobility has vastly different technical challenges, so it might be easier to accomplish than hydrogen power. Automating the physical aspects of driving – steering, braking, acceleration – is simple because so many of those functions are computer controlled. The challenge is developing sensors that can gather data on what’s going on in the driving environment and computer systems to interpret that data and translate it into driving instructions.

Lots of money is flowing into the sensor/algorithm side of the problem, and technology is developing quickly. The goods-hauling industry thinks the technology could alleviate a shortage of drivers, and safety regulators like the idea of getting stupid human decision-making out of the driving process. Multiple states are in a weird sort of arms race to rush autonomous vehicles onto public roads for testing.

But as hydrogen power continues to show, billions of dollars of R&D investments do not automatically create solutions. Sensor systems may never become sophisticated enough to identify and categorize every potential driving hazard. Algorithms may never develop decision-making processes that will satisfy regulators. The driving public may never fully embrace the idea of fully automated roads.

Will we actually have self-driving cars in our lifetime? I’ll hedge and say sure, they’ll be here in 10 years, powered by hydrogen fuel cells.