Not to belittle handoff problems, but this seems of a piece with the general consensus that driverless cars are much harder than we thought just a few years ago ("Who's in the driver's seat?" July). What gives?
Two years ago, I attended the Intelligent Transportation Society annual conference in Montreal. I asked this question in Q&A sessions and of individual autonomous vehicle experts: "Can you imagine a future in which a large percentage of the total vehicle fleet is autonomous, but not connected?"
The answers were unequivocal: Vehicle-to-vehicle technology so simplified the challenge of full autonomy that V2V should be thought of as a prerequisite to higher levels of autonomy.
At the time, only Tesla appeared to be working towards "fake human" driving, by which I mean, relying on cameras/visible spectrum and using artificial intelligence to anticipate the behavior of other actors on the road.
Other OEMs seemed to envision a different road to full autonomy (not easy, but easier). They envisioned that increasing penetration of vehicles operating at increasing levels of autonomy would come in an increasingly connected environment. And that by the time full autonomy was at all common, most nearby vehicles would be communicating intent and nearly all, at the very least, sharing location/vector data.
Today, most OEMs seem resigned to Tesla's approach — albeit with more diverse sensors — in which vehicles operate on their own, without direct input from surrounding vehicles. Refocusing on vehicle-to-everything connection might speed progress to higher-level autonomy and improve safety in the interim.
MARK GARDINER, CEO/Communications strategist, HAVstory.com午夜剧院