The toaster-shaped shuttle slammed into a concrete barrier. Shards of glass sprayed through the air. Hub motors on the rear wheels detached from the underlying vehicle structure.
For Local Motors, this constituted a good start.
Where none existed before, the mobility company — perhaps most well-known for 3D-printing vehicles — had developed testing protocols to vet the crashworthiness of its purpose-built Olli autonomous shuttle. That was a year and a half ago.
The company took lessons from that initial round of testing and iterated on its printing methods, its frontal-crash structures and made design changes to the vehicle itself. In January, Local Motors ran a second gauntlet of crash tests more narrowly focused on components that showed substantial gains in how crash energy is dissipated in a vehicle without traditional crumple zones.
Among the findings: Local Motors says its 3D-printed frontal crush structure performed at an equivalent level to the crush structure found in a Mercedes-Benz E-Class sedan. Such results, reached with testing conducted by an independent engineering and testing firm, might not only shed light on the safety potential of 3D-printed parts but also showcase the speed at which design changes can be iterated and implemented when relying on additive manufacturing.
“At first, I thought, ‘God, what are these guys thinking,’ ” Bob De Kruyff, a longtime General Motors engineer who now serves as vice president of engineering at Local Motors, tells Automotive News. “I just wasn’t sure how you’d do that. But frankly, I’ve become a believer in 3D printing. I was a bit skeptical at first, and now I see the advantages of it, and what it’s done for the vehicle.”
The crash tests that have cemented his conversion are believed to be the first of their kind for autonomous shuttles.
Crash tests may be commonplace for conventional vehicles, but for Local Motors and its low-speed, bespoke autonomous vehicle, the testing regimen represented an unusual project for which there was no blueprint.
No government or third-party safety standards exist for these vehicles or others like them in the fledgling self-driving shuttle industry. Nor are there any requirements that make developing such tests necessary.
Federal regulations permit four-wheeled vehicles that travel 25 mph or slower to be classified as “low-speed vehicles,” and these vehicles do not need to meet the same safety standards traditional passenger vehicles must follow.
The government instituted this classification so small golf carts and similar vehicles could be used in controlled, low-speed areas, such as retirement communities. Many purveyors of self-driving technology have developed business plans that address those geofenced areas. They ensure their vehicles travel at those speeds, in part, so they can remain within the bounds of low-speed vehicle rules and avoid more stringent safety standards.
While the Olli speeds stay under that threshold, Local Motors CEO Jay Rogers took the opposite approach when it came to crash testing.
“No one is asking us to do this, but if we weren’t, we would be hiding behind a law,” he said. “That’s not the right way to be building public trust. If we weren’t doing this, it would be woefully inadequate.”
Though self-driving technology offers the specter of a future without crashes, that future remains a long way away. Even when self-driving vehicles are ready for everyday consumers to purchase, it will be decades before they’re in the hands of most motorists.
In the interim, computer-controlled vehicles will share the road with human-driven ones. Avoiding collisions will be the aspiration; mitigating crash forces will remain an important engineering consideration.
For Local Motors, which first caught attention for its automotive ambitions with 3D-printing at the 2015 North American International Auto Show, that comes with added complexity. Most companies are upfitting self-driving systems onto traditional vehicles. Local Motors has not only designed the Olli from scratch, but it used 3D-printed components to produce the vehicles.
For outsiders, the most jarring aspect of Local Motors’ initial test was the hail of shattered glass. Those reactions surprised Rogers and De Kruyff, who were more interested in evaluating the vehicle structure.
“Obviously when you are watching that, there’s a lot of glass, and that’s pretty spectacular,” De Kruyff said. “But the actual impact loads and free forces were not that bad. We were encouraged. We learned quite a bit about bonding the upper and lower structures, and the material properties. So we decided to keep going.”
Switching to tempered glass for future tests was the easy part. Evaluating how crash forces spread through the Olli and developing new countermeasures for absorbing that energy proved challenging.
The Olli’s rectangular shape leaves no room for the typical crumple zones found on passenger cars. So De Kruyff and his engineering team developed a 3D-printed structure made from a proprietary blend of carbon fiber and other materials, iterated on that and took the structure to independent testing lab Calspan in upstate New York early this year.
This time, the company added a twist. Instead of running the tests at 25 mph as it did during initial tests, it upped the speed to 35 mph.
Recordings of the ensuing sled test show the crush structure splinter into smithereens in a scene every bit as dramatic as shattering glass from its first test. This time, it was the desired effect. The structure dissipated energy in a way that represented a 253 percent improvement from the tests 18 months before.
In some respects, Rogers’ interest in better understanding the crashworthiness of the Olli was stirred by a bottle of dish soap.
He knew someone who worked in product development at consumer-goods conglomerate Procter & Gamble, a friend who enlightened him on the entanglements involved in testing and validating the plastic bottle which holds Joy soap.
A hard-plastic cap. A softer, squeezable bottle. Varying amounts of liquid inside. Sounds simple, right? Procter & Gamble used computer modeling to validate all the different scenarios in which the Joy bottle could be twisted and turned at every conceivable weight.
“Modeling for them involves a supercomputer, and that’s in a two-part, maybe three-part system,” Rogers said. “So you can imagine why automotive companies have a difficult time validating in just pure software.”
In his mind, it underscored the limitations of computer modeling for ground vehicles made from thousands of parts that interact at different speeds with different weights and different fuel levels and a limitless number of obstacles with their own weights, speeds and trajectories.
Vehicles, he said, need real-life crash tests. For Local Motors, its 3D-printing may hold an unparalleled advantage.
“If something doesn’t work, we go back and print a different version. It’s that simple,” Rogers said. “If you are a traditional manufacturer, that’s new tooling costs for something that may not work. That’s really expensive. Really expensive.”
For a small company such as Local Motors, crash testing still isn’t cheap. But it may help the company accelerate toward a milepost it views as something of an inevitability — increasing its operating speeds and traveling public roads at 35 mph.
“Many of these vehicles, including our competitors, run at 15 miles an hour, and other people on the road, frankly, get pissed off that these vehicles are slow,” De Kruyff said. “So the writing is on the wall that we’re going to have to get up to 30 or 35 miles an hour at some point.”
Short of adhering to the same safety requirements passenger vehicles must comply with, Local Motors hopes regulators may consider a set of standards for creating an entirely new vehicle category, akin to the European Union’s M2 bus classification, that establishes something of a middle ground between the golf carts and conventional cars defined in federal motor vehicle safety standards.
“We do think there is a category there,” De Kruyff said. “And we want to find that, because that’s what keeps it attractive. It’s nimble and allows us to operate on battery power in a relatively light vehicle. If we had to design to FMVSS, it wouldn’t be an Olli anymore. It’d be an E-350 van.”
Those are practical reasons Local Motors does crash tests today. But the overriding factor behind the testing is a philosophical one.
“We want to know how good we can be,” Rogers said.