Hacker News new | comments | show | ask | jobs | submit login
Smart Headlights Let Drivers See Through Rain and Snow (scientificamerican.com)
35 points by Sukotto 1906 days ago | hide | past | web | 15 comments | favorite

I like what one of the commenters on the original post suggested:

How about having this intelligent system locate the drivers of oncoming vehicles and not shine light directly into their faces? Same thing for cars being followed, avoiding their rear view mirrors. If headlights were smart, we could make them brighter without blinding anybody.

This already exists, Gentex's SmartBeam[1] product has technology to shutter headlights on high-beam mode so that the light does not hit the windscreens of oncoming cars.


  > The researchers hope to cut system latency by developing faster ways of
  > transferring images from the camera to the microprocessor
  > running their algorithm and from that microprocessor to the headlight.
They should collaborate with John Carmack; this is quite similar to the latency problems he is working on for head-mounted displays.


I was hoping this would be a polarizing filter or modification to standard headlights.

While cool, it's unfortunately an entire electronics system dedicated to the situation. The last thing I want is a frozen sensor detecting the entire road was 'rain' and shutting off my headlights.

If you can think of this concern in 8 seconds or so, I assume automotive engineers can too.

It falls back to current behavior of just having the headlights on in the case of an error like that. There was another article on these a week or two ago on a different website that specifically mentioned something like that.

Manual override?

I'm thinking they'll probably include a fail-safe of some kind.

Yeah... most everything on the wiring harness goes back to a computer, these days... and modern ECUs will ignore data that doesn't fall within expected parameters.

Anyone know what the end product for something like this would cost?

Seems to me like this won't make it into popular use. At least not for a long time simply due to the cost of installation and upkeep.

I find that it is very rare that I find driving conditions so bad that it is a danger to drive. When I find visibility difficult I either slow down, take a safer road (easier to drive at 30 in low vis than 70) or I just stop driving.

The technology looks great. However, it will be difficult to justify the cost when the actual technology won't be in use 99% of time.

Whilst interesting I still feel a HUD type visual enhancer would be better.

Too many factors with this system that need to be accomodated from a sudden wind gust, rain that bounces and in dense fog and lets not forget the classic one were the car infront or lorry kicks up spray. Too many dynamic variables and thats just a few from the top of my head. Love to be proven wrong, but a HUD visual enhancer system seems alot easier and safer when accomodating these many many variables.

I am asking this because I'm ignorant and fishing for additional detail. In your idea, where does the HUD get its data to display, from another modality like radar?

could be many forms but radar or what i was thinking is visual enhancment with camera's, that type of thing. You are also not ignorant by far as a ignorant people don't ask questions I have found, so don't put yourself down.

Wish there was a video clip of this in action.

Project page with many videos, including entire talks given by one of the authors here: http://www.cs.cmu.edu/~ILIM/projects/IL/smartHeadlight/

The project is still in the academic research phase and I don't think they're quite as far as actually implementing it in a car in the wild, although they're moving ahead with it as fast as they can, so they might have actually done this recently. It won an award at the last International Conference on Computational Photography (ICCP) because it's just such a cool idea, and it seems like it might actually be feasible to some degree (even despite all the issues people have mentioned here and elsewhere). At ICCP, it was still in simulation and prototype-with-rain-machine-in-the-lab mode.

Btw I highly recommend watching the ICCP talk, available to watch from that webpage, or for download here: http://www.cs.cmu.edu/~ILIM/projects/IL/smartHeadlight/media...

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact