x band radar+detecting car and obstacle in roads+1.5 kilometer range

hi guys,im working on a project in university.its about detecting cars and probably human (any obstacle) in the roads and it will alarm the driver (specially big machines like bus and truck) about the distance and the collision time (specially in case of overtake and pass) because the death rate of this kind of accident is high.the tip is the sensor range must above the 1 kilometer,perhaps 1.2 to 1.5 km.it think it must be a x band millimeter doppler radar.what sensor do i need?from which company can i purchase?

Unless you've got an absolutely needle-thin beam, there's going to be so much clutter in that range that I doubt such a sensor exists, and if it does, unless you have a defence budget, you can't afford it.

(Is there something wrong with your shift key?)

And for a 1Km range you'll need a fair amount of power - for which you won't get a license. Or you'll annoy a lot of serious people messing up their radar. You need clearance to even buy such stuff professionally.

And doppler would be no good - everything static would reflect the same amount depending on the speed of your truck. You'd need pulse or fm to identify the target from the background.

And X-band is 10GHz - 3cm wavelength - not mm waves. Anti-collision car radar works at 77GHz
and has a much shorter range. mm waves start at 30GHz - I've worked on such stuff.

And if you want to detect a 2m tall person at 1Km using 3cm you'd need a beam width of about 0.1 degree, which would need a dish much bigger than your truck... ( > 20m across ).

And that's presuming there's nothing else about to reflect back cluttering the return

And given your truck is moving it'll be bouncing about - how do you point this monster?

I don't think even a miltary budget can make this work.

Do some sums - finding out how I get these numbers would be good for you.

And tell your supervisor to get real

regards

Allan

Detecting cars and obstacles at over a kilometer away is very difficult. I would however imagine that a good camera with a suitable lens and a well written computer vision program (using e.g. OpenCV) could be able to do it.

Have to be very good to work in a moving ( bouncing about) vehicle..

regards

Allan

edit: think of trying to hold a pair of 10x binoculars steady in the passenger seat....

allanhurst:
Have to be very good to work in a moving ( bouncing about) vehicle..

regards

Allan

edit: think of trying to hold a pair of 10x binoculars steady in the passenger seat....

Excellent point

couple of additional point.

1st : there is rarely a road that has 1km straight line that is visible in that straight line, that is not a highway.
also, with that much straight distance, one would expect cross traffic and for it to clear before reaching that location.

2nd: nuisance tripping would make me turn it off.

3rd, speed sensitive distance, for assumed stopping distance would be very helpful for preventing nuisance alarms.
no need to worry about what is a Km up the road when you are doing 10 kph.

Lastly, the vision system would be much better, you can get video stabilization so you are not at the mercy of a bumpy road.

I agree you'd be better off with optics - but still impossible..

regards

Allan

ps if you ever get it to work let me know - it's worth a PhD!

Nothing is impossible, given sufficient resources;

  1. buy a surplus F4-F Phantom
  2. pull out the Look-Down/Shoot-Down Radar and Targeting systems
  3. buy an old CRAY X/MP to analyze the returns
  4. use an UNO and a 4*20 LCD to display the messages ("ALERT- SYSTEM FAILURE")