Automotive News reporter Hans Greimel inspects prototype millimeter radar sensor in Lexus GS (Photo: Bertel Schmitt)
When it comes to autonomous driving technology, “legacy automakers” are putting their money where Tesla’s mouth is. Today, Toyota agreed to take control of Fujitsu Ten, one of the world’s leading developers of millimeter wave radar, a key technology for dependable autonomous drive.
Once the deal is consummated by the end of the current fiscal on March 31, 2017, Toyota, together with its group company DENSO, will own 86% of Fujitsu Ten, with electronics powerhouse Fujitsu holding the rest. DENSO is one of the world’s largest OEM parts suppliers, similar in size to Bosch, Continental, or Magna.
For the past few years, DENSO and Fujitsu Ten jointly developed a revolutionary 3D millimeter radar unit that will give cars the visual acuity that is required to make autonomous driving a reality. Current technology mainly uses old style radar, and cameras, and it provides rudimentary vision that would make humans flunk any eye test at the DMV. Experimental autonomous cars rely on radar of the rotating coffee-can type, but that would be utterly impractical for regular use. Without that, radar projects a narrow beam, leading to a severe case of tunnel vision. Then there are wide angle cameras. They are shortsighted.
What the car needs to really make it see is long-range sensors with high depth perception, sensors that are not confused by sun, rain, or big white trucks. The sensors must be unobtrusive, and affordable enough so that four of them can be fitted into the car. There are no such sensors on the market, yet, but they are coming.
Who’s afraid of the big white truck? (Picture: Fujitsu Ten)
I viewed prototypes of such a sensor two years ago. Then last year, I trusted my life to an experimental Lexus GS that, fitted with the sensors, threaded its way through thick Tokyo traffic without the driver touching the wheel.Last year, Ken Koibuchi, General Manager of Intelligent Vehicle Development at Toyota, told me that the tested tech won’t be on the market before 2020, because that’s how long it takes a responsible company to test a quantum-leap like this, and that’s how long it will take to produce the still experimental sensor arrays in the required form factor and cost. This is where Fujitsu Ten comes in.
Toyota already has shown its commitment to autonomous technology by investing $1 billion into an R&D enterprise in the U.S. focused on artificial intelligence, and led by Gill Pratt, who it poached from DARPA. The Fujitsu Ten acquisition helps to apply the developed science to real life.
Once again, it is shown that the true innovators in the autonomous drive field are the established automakers. They have the money, and the required scale of millions of cars needed to make advanced technology affordable for all. Everything else is, as Sascha Lobo, tech commentator extraordinaire at Germany’s Spiegel Magazin said so aptly a few days ago, “just stock-exchange-contaminated technology folklore, ritualized suppositions by people who believe that technology will change the world. The truth is that technology hardly changes the world, it’s more the way people interact with technology.”
For that, the tech must be practical, affordable, reliable, and in the hands of millions of people. An automaker with 10 million cars a year can bring that change much faster than one with 50,000.
No comments:
Post a Comment