LAS VEGAS -- Intel and its Mobileye subsidiary have devoted considerable resources to developing technologies to support fully self-driving vehicles. But they're no longer waiting for an autonomous era to arrive before deploying them.
Call it trickle-down technology.
Amid a laundry list of deals and partnerships announced this week at CES, the two companies said they're taking two of the linchpin technologies they've created for autonomous vehicles and implementing them in advanced driver-assist features to get them on the road as soon as possible.
The companies will use their high-definition, crowdsourced maps developed for fully autonomous vehicles and use them in driver-assist systems developed in conjunction with Volkswagen. The maps will augment lane-keeping features, helping systems maintain a steady lane even if markings are obscured.
Separately, Intel and Mobileye say global supplier Valeo will integrate their Responsibility-Sensitive Safety driving policy into driver-assist systems. Essentially, the policy takes human ideas of safe driving and codifies them in mathematical formulas that set operational boundaries for following distances and speeds.
The policy was developed to set rules for self-driving systems. But Intel says it thinks it can prevent crashes with humans still behind the wheel.
"What we've realized, I guess, in the course of developing technology for fully automated driving, while we wait for regulators and the legal system to sort out, we've thought, 'Let's deploy some of these as soon as we can in an ADAS context,' " said Jack Weast, vice president of autonomous vehicle standards at Mobileye and senior principal engineer at Intel. "If we can save lives now by trickling down technologies and making them active while a human is still driving, then we have a moral imperative to do so."
In driving-assist applications, Responsibility-Sensitive Safety may help alert human drivers when they're tailgating or approaching a curve at too high a rate of speed. Another technology, Road Experience Management, could provide similar checks, ensuring lanes are maintained when sensors aboard the vehicle cannot determine the proper path ahead.
Responsibility-Sensitive Safety
Mobileye has touted Responsibility-Sensitive Safety as a potential industry standard for determining safe rules of the road for autonomous vehicles. No such standards exist, and regulators have broadly been intrigued with the idea of a standard set of driving policies — if not this one, then something similar and agreed upon by others.
For Intel and Mobileye, the Valeo integration is a signal the industry is receptive to that idea. Valeo is the second announced partner on Responsibility-Sensitive Safety. Last summer, Baidu said it would integrate it on its platforms for Level 4 autonomous systems.
In a sense, the spread of these technologies into driver-assist systems is similar to an idea that Toyota has endorsed — that the safest car on the road might not be one with a fully autonomous system nor a human driver, but a merger of the two. Toyota has developed a Guardian system that runs in the background while a human handles driving. It essentially lies quiet except when it detects an imminent hazard and intervenes to thwart a crash.
Among the differences in what Intel and Mobileye are proposing with Responsibility-Sensitive Safety and Road Experience Management active in driver-assist systems, it seems, is they would be more active in preventing hazardous conditions from developing, perhaps nudging a car back toward the center of a lane, and maintaining comfortable riding conditions as part of their normal operations.
"If you look at automated emergency braking, it's a valuable feature, to be sure, but it doesn't kick in until the last second before a collision, and it provides full braking force," Weast said. "That's great for avoiding an accident, but it's not a good experience. In a fully automated vehicle, we wouldn't do that. So RSS can maintain a safe distance that provides a reasonable amount of braking distance. In doing so, it can provide a human-driven car the same safety guarantees.
"Just like electronic traction control kicks in and you don't even know it ... a safe-distance formula kicks in and you don't even know it," he said. "But it's keeping you safe."
Full autonomy
Despite the focus on these trickle-down technologies, Intel and Mobileye are hardly ignoring full autonomy.
As part of their spate of CES announcements, the companies said the Beijing Public Transport Corp. will incorporate Mobileye's fully self-driving system for Level 4 operations on public transportation vehicles in China. The system incorporates cameras, radar and lidar sensors. The vehicles are expected to reach public roads in 2022.
This marks the second partnership involving Mobileye's self-driving system. In October, Mobileye formed a joint venture with Volkswagen to deploy an autonomous ride-hailing service in Israel. Testing for that service is expected to begin early this year, with commercial service slated to begin in 2022.
There's more: a partnership with British mapping company Ordnance Survey that shows how data from Mobileye's mapping information might be monetized in ways unrelated to vehicle operations, and an update on Intel's partnership with Warner Bros. that reimagines interior cockpit designs.
Combined, the Intel and Mobileye announcements at CES show progress that's both incremental and tangible.
"We're making forward progress," said Weast, "not vaporware announcements, but actual, real things that help move the ball forward on a couple of different fronts."
