Intel's Mobileye unveils first 'production-grade fully electric self-driving vehicle,' partners with Sixt for Munich launch Intel's Mobileye arm has announced a partnership with car rental giant Sixt to launch a fully autonomous robotaxi in Munich, showcasing the tech at the IAA Mobility conference this week – before admitting the platform won't open to the public until regulatory approvals are granted. The brightly coloured vehicle showcased by …
Old process nodes do have their advantages as they are less susseptible to upsets due to things like radiation.
Whilst the probability of any particular device being hit by such an upset is statistically very low, it becomes a certainty when you have large fleets of vehicles (i.e. millions) in operation at any one time.
Regulatory approval or not, over recent months I have come to the conclusion that prior to such vehicles being publicly launched, the CEO of whatever company it is, should be strapped, comfortably, with adequate padding etc., to the front of said vehicle and stay in place for at least three months of everyday use, allowing for adequate sleep/food/toilet breaks etc. This way potential buyers will, I feel, be able to pass sound judgment on the validity of the claims made regarding the safety of such contraptions! I know I may sound like the nay sayers who railed against Stevenson's Rocket, but I feel very strongly that the proof really is in the eating of this particular pudding!
Oh damn and double blast! It has been pointed out to me by a fellow researcher at the establishment where I pass my time, that in all likelihood the outcome of the scenario that I outlined above, would result in the electronic-type computer "brains" controlling the vehicle, rapidly coming to the conclusion, through means of machine learning and algorithms etc., that having a squidgy meat-bag of a human draped over the front of the car in no way impedes forward progress and normal driving! Thus, potentially, foreshadowing the rise of the machines etc.! Upon reflection then, perhaps it would be more prudent to have said CEO merely as a front-seat passenger for the duration of the demonstration.
I would have thought that the ultimate goal of these robotaxi projects is to get to the point where hardly anyone owns a car and instead everyone hails a low emission robotaxi. Which means far fewer cars will be needed so that 5x increase in semiconductors may be true per car but overall the number of semiconductors used in the automotive industry will be similar or fewer than now.
Trains already have this problem. That's why train journeys are so much cheaper out of peak times as it encourages people who don't need to travel at peak times to wait. Taxis like Uber do this too, but you still have a driver who wants paying as well as the vehicle costs to cover. With a driverless taxi all it needs to do is beat running, depreciation and finance costs by a small margin to be profitable. A private vehicle still has those three costs to cover but the depreciation and finance costs are likely to be spread over many more journeys.
'"This vehicle has two layers of sensors working independently from each other at the perception level,"''
Don't all sensors work at "the perception level" in some way or other?
"each layer performs independently, they do not rely on each other in order to get a redundant system"
Does this mean the result is (or is not) a redundant system?
Parts of the announcement appear suspiciously like (in the words of George Orwell) doubleplusgood duckspeak.
Apropos of which - what happens when you have a large number of cars in close proximity, all spitting out Lidar pulses? I've not worked with the technology, so I have no idea, but I can't help wondering about sensors receiving data from the wrong laser... is there some sort of negotiation going on to say 'my turn', 'your turn', 'no, no, after you' between vehicles, or is it down to collision detection of the signal (or something more esoteric)?
Enquiring minds want to know!
Sixt are famous for being bloody awful with insurance claims, e.g double claims for scratches.
How is this going to affect their bottom line? Are they going to be responsible for all their own driving!
Or will their insurance dept come up with a few more pages of small print and still try to charge you when the car decides to drive into a lampost itself?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
