F1technical.net

Thought you guys here would find this interesting

I'm surprised Siemens put their name on that POS.

Was it only using GPS to drive?!

That is a bit crap.

Drunk Siemens autonomous Mustang... Cars and coffee anyone?

I'm not sure what they were trying to prove with that GPS guided AV? Maybe that GPS only is not the way to do it...

The comments remind me a bit of the old saying just before television took over the world. "Radio with pictures? It'll never work!"

Greg Locock wrote: ↑

14 Jul 2018, 06:21

I'm surprised Siemens put their name on that POS.

The benefit would be in the training.
No doubt next year it will be far better.

Meanwhile in Cambridge a car is taught to follow a lane in 20 minutes, or so they claim.

https://www.cnet.com/uk/news/an-ai-lear ... 0-minutes/

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Big Tea wrote: ↑

14 Jul 2018, 12:05

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Individually numbered road aggregate that's mapped and sensors on every vehicle

AJI wrote: ↑

14 Jul 2018, 16:33

Big Tea wrote: ↑

14 Jul 2018, 12:05

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Individually numbered road aggregate that's mapped and sensors on every vehicle

That sort of thing. Not controlling the car, but making sure it knows exactly where it is, to a couple of mm.
Area traffic information and re-directs to miss hot spots etc, ensuring there is parking before letting it in a parking area etc. Even to the extent of stationary cameras waring of undecided pedestrians.
Belt and braces.

Big Tea wrote: ↑

14 Jul 2018, 12:05

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Something like this?

henry wrote: ↑

14 Jul 2018, 16:55

Big Tea wrote: ↑

14 Jul 2018, 12:05

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Something like this?

https://www.velocetoday.com/wp-content/ ... 57-570.jpg

Yes, except done properly there would be 2 empty lanes. One for large vehicles and one for emergency

Oh! and the kids would have their faces in hyper cartoons getting wound up

Or, mixing smart dust into the blacktop/ concrete...

https://people.eecs.berkeley.edu/~pister/SmartDust/

Big Tea wrote: ↑

14 Jul 2018, 16:55

AJI wrote: ↑

14 Jul 2018, 16:33

Big Tea wrote: ↑

14 Jul 2018, 12:05

I know it would be akin to laying tracks, but is there any reason a wire can not be embedded in the road surface?
Or for that matter a bright red line painted mid carriageway. The vehicle does not need to blindly follow it like rail tracks, but it would be there for reference.

When you consider the other changes that will need to be made to infrastructure, this would be minuscule for much return, especially if the 'wire' radiated data to the vehicle.

Individually numbered road aggregate that's mapped and sensors on every vehicle

That sort of thing. Not controlling the car, but making sure it knows exactly where it is, to a couple of mm.
Area traffic information and re-directs to miss hot spots etc, ensuring there is parking before letting it in a parking area etc. Even to the extent of stationary cameras waring of undecided pedestrians.
Belt and braces.

Totally doable. Referenced against GPS and terestrial telco towers and peer to peer transponders and cameras...
You could even put a compass in there!

Morteza wrote: ↑

14 Jul 2018, 02:42

https://www.youtube.com/watch?v=1NsrijKBbK8

It's as though it has simple path creation. Straight lines within a bounding box that is programmed as the road. Pinball mode. Point to point based on sparse GPS loc updates? Makes me wonder how far a vehicle could get based on odometer and steering angle data alone. Preset guidance as with the old V-2 rockets. Might make for a good academic competition.

Thinking to the future; imagine if vehicle automation becomes total. Typical traffic measures would be eliminated. Roads would become blank slates. No lines, no lights, no signs, no signals. Just vehicles moving at high speeds, each aware of the other. It would look chaotic to our eyes. Vehicles crossing intersections at speed in all directions without slowing down. Constant movement, reduced energy lost to braking. Steady speeds for the majority of travel, regardless of congestion. Masses of vehicles moving in synchrony, which would have otherwise been traffic if each was controlled by individual human actors. Huge energy savings potentially.

Next thing to perfect would be deployable ramps at intersections. If traffic density exceeds minimum spacing required to interlace cross-intersection traffic at high speed, the ramps would elevate on one road and send the cars leaping over the perpendicular traffic stream. Cheaper than building overpasses or underpasses inside of cities maybe. Or build limited STOL into the cars to further shift the cost away from infrastructure. 'Mundane' flying cars.