Waymo fleshes out self-driving car tech with hardware that goes soft at first sign of trouble Waymo has been granted a patent to deck out its self-driving cars in a material that becomes less rigid when its sensors detect a high chance of collision. Autonomous cars have unresolved safety issues. Although these machines might not suffer from the dangers of fatigue or road rage, they still aren’t completely safe. …
So they want to mess with the aerodynamics and the handling of the car at close to the point of a collision. Exactly when you most need to have a solid, repeatable, and understandable response from the car.
Yep that's a good idea. Totally.
Also, dear Lord the cost of repairs to your car must be atrocious after this. No more just changing the panels and the curmple zones, you may as well now buy a whole new car.
Finally, I thought a patent had to have a specific implementation not a "it will do this, via Magic and I dont know cables or something". Oh wait, this is filed in America. Sorry forget that, stupid patents are the norm in the states...
The cost of repairs is irrelevant if the purpose is to prevent injury to humans. They are even more expensive to repair.
Makes you wonder though, why they can't just use a permanently soft or resilient material anyway - an inch-thick layer of expanded foam over the whole body with a thin skin for colour and aerodynamics. Combine a seriously impact-absorbing outer shell with a seriously rigid cabin shell and you have the making of something useful.
Actually, with sufficient development it's a bloody clever idea, at low speed impacts you want a relatively soft deformation structure, but with high speed impacts you want a deformation structure that is much more rigid (to absorb the energy).
The problem is making it act reliably and consistently. Especially in the high speed impact scenario, I.e. You don't want your tech becoming the weak link in the chain compromising the safety capabilities of the vehicle.
Great park it in your garage and it senses "objects" within crash distance on all four sides and rapidly then flops to the ground as a jelly like mess requiring the purchase of a new car. This is either the height of stupidity or a genius ploy to boost profits by getting customers to need a new car every X days.
...only work if you have solid structure reacting the tension loads in compression. File this one next to the patent for the British Rail nuclear powered flying saucer.
Or in other words, somewhere under the piano wires holding the shape of the car, there's a nice solid safety cell holding them tight which will protect the vehicle occupants. I for one don't want to be run
over by a high speed egg slicer that's falling to pieces.
It would be cheaper and easier to use high density foam or honeycomb with a rigid(ish) plastic coating to absorb an impact, attached to a space-frame underneath. (Think of a weaker version of a formula 1 car chassis, supported by a stiffer, possibly similar internal structure.) This approach would also behave much more predictably in an impact, because if you've got enough tension in a cable to hold the external shape of a car against aerodynamic loads at motorway speeds, you've got quite a lot of stored energy *in* the tension cables...
While these floppy cable panels are seemingly obviously dumb, anyone reading these comments should realise a lot of people here are, at best, doing educated guesses. Myself included.
Foam inside panels seems so much better, but what do I know. I'm no... car physicist. My guess is the panel doesn't go completely floppy, it just softens to the point where you could push it like the flimsy plastic / thin metal the outside will be. With the cables, hopefully there will be enough of them to stop the cheesewire effect if something really does snap them or push them or something.
Needs more testing, is my verdict. Which I guess is what they're doing. Makes this whole comment seem sort of pointless, doesn't it?
"The force of the vehicle's impact is a primary factor in the amount of damage that is caused by the vehicle."
This was identified as a possible factor after the engineering review group tested a revised prototype with the hardened steel spikes removed from the front, and confirmed in subsequent testing after the next prototype revision lost the front-mounted impact-triggered directional shaped charge warheads.
"Coating the bumper in a sticky layer will supposedly glue the colliding object – such as a pedestrian – to the front of the car during the impact like flypaper, to protect them from 'secondary impacts.'”
I'm waiting for something like GTA Lite where people try to hack the driving AI in order to collect an entire bumper-full of ex-pedestrians!
What seems to be needed, is the opposite of Silly Putty. So the car body would remain solid when bumped, but flow when hit at high velocity.
It'll be interesting to see more innovative vehicle designs, that are not based on accommodating a human driver. Backwards-facing seats, cameras/screens instead of fragile windows etc. I know what you're all thinking, yep, me too: SPV from Captain Scarlet.
I drive a big V8 Durango. This time of year I have not filled the tank for over a month since I ride my bicycle. When I do need to pick up something big/heavy then I drive the Durango. Last time I was out in it I was turning at a two lane wide left turn across oncoming traffic (stopped at the red light for them). I was on the inside of those two lanes. Another person in a very small car turned at the same time as me. He was in the outer lane. That lane merges with my lane in a very short distance. He was trying to beat me to the merging point. No way he would make it in time.
I had two choices.
1: Hit the brakes. That would then put the person behind me far closer to me, tailgating and stacking up all the traffic behind me.
2: Pedal to the metal. This convinced the other driver to the right of me to slow down and it also gave him plenty of room to merge behind me.
I then nailed it to the speed limit. A bit further there were some workers in the center of the lane divider working on shrubbery and grass. There was no lower speed limit posted but I slowed down to half the speed limit anyway just in case somebody made a mistake and tripped/backed into the actual traffic lane. I was driving the same as most police officers.
The man behind me dropped well back to a very safe following distance.
I drive like an unofficial ghost car. I learned how effective this can be when I once had a vehicle that was the same make and model as the real ghost cars. I also had the same haircut as the officers and would wear a jacket that looked similar. I also stuck a fake antenna on the top of my vehicle that looked the same as theirs.
I was legally impersonating a police officer. I knew many of the officers since I was teaching them data recovery and safety. I also showed them how I could easily hack their LIDAR by playing proper music on a retroreflective loudspeaker just behind the grill at the front above the (dirty) retroreflective licence plate.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
