Single ended breakbeam

Discussion in 'The Projects Forum' started by jbord39, Nov 4, 2011.

  1. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    Hey all,

    I am trying to make a sensor or system which functions like a breakbeam. The only catch is that there can be no receiver on the other side of the beam.

    Ideally the detection width should be small, and the distance is about 10-20 feet. It should also have a fast enough response time to output from a passing car.

    Here are the ideas I have considered so far:

    X-band motion radar detector.
    Passive Infrared Sensor (maybe the http://www.optexamerica.com/content/12/CX702.pdf)

    Ultrasound-- This is my least favorite idea as it is hard to keep the beam thin and response time of sensors is around 10-20Hz.

    Thanks for any help,

    John
     
  2. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,771
    971
    must it be a beam... How about pressure switch? Typical traffic counters use a sealed rubber tube that is run across the road and a pressure sensor to detect when a tire drives over it.
    http://www.drivewayalert.com/pressure-tube
     
  3. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    It can't be a pressure switch because we need to stay off the road. To be in the road requires permits and we would rather not deal with that.
     
  4. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    2,574
    230
    Could you use IR beam with a passive reflector on the opposite side of the road?

    Ken
     
  5. John P

    AAC Fanatic!

    Oct 14, 2008
    1,634
    224
    How about using a cheap TV camera? You can get little ones for less than $50, and you could use an LM1881 Video Sync Separator to pick apart the output signal. Then just look for a change in (let's say) a place corresponding to the center of a central line of the image. Or you could put up an array of alternating black and white marks on the far side of the road, and verify that the pattern is or isn't visible. No extra light source needed, a long as it's daylight.
     
  6. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I would say cars are reasonably reflective, so you could try an IR laser beam coupled with a lens and camera/ir sensor.

    Also you can start with visible laser to get the optics pointing the right way and then switch to IR so that you don´t dazzle the drivers.

    Also, are there multiple lanes of traffic or just one? How about the other direction, is it close enough to cause recognition?
     
  7. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    Only one lane of traffic. Don't you think some oddly shaped cars could not reflect the laser beam back into the detector?
     
  8. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    2,574
    230
    Maybe a rather tight beam, aimed at an angle to a mirror (parallel the the road) across the road, and back to a receiver at a complimentary angle further down the road on the same side as the transmitter.

    Ken
     
  9. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    Thanks, that would work great. The only thing is we are trying to minimize the amount of set up for any human operator and also make it very portable. Preferably nothing should be on more than one side of road.

    Do passive infrared sensors have a drawback? I am trying to get an idea before I order sensor parts.
     
  10. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    2,574
    230
    Not sure what you are referring to.

    Ken
     
  11. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    I mean, is there a reason that a PIR with a narrow viewing width couldn't function as a breakbeam?
     
  12. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    2,574
    230
    Interesting! Since they work on a change in temperature from background that may or may not work. Be worth a try.

    Ken
     
  13. crutschow

    Expert

    Mar 14, 2008
    13,023
    3,236
    A passive IR sensor, such as used in motion sensors may work, if it's response times can be made fast enough. This circuit, for example, has a LP rolloff time-constant in the first stage op amp of 100ms, so that might have to be reduced to get the response time you need (how fast will the vehicles be moving?). Of course that may increase the noise as there's likely a tradeoff between response speed and sensitivity.
     
  14. Adjuster

    Well-Known Member

    Dec 26, 2010
    2,147
    300
    Will passive (PIR) sensors respond quickly enough to detect vehicles passing at the speeds you expect? These are usually pyroelectric devices, and therefore inherently slow.
     
  15. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    You could use a passive visual light system, ie a camera or other passive light detecting device like photodiode which may be faster and cheaper.

    If you point it at a slight downward angle, looking at the road surface (at the far side of the one lane) it will visually detect a car in that lane but not in other lanes.

    If you could use two sensors on the one tripod, say a foot apart and parallel, that could also return a vehicle speed as well as detection, and the speed data would also be useful in eliminating false readings like a bug or leaf that moved in front of a sensor.
     
  16. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    We are expecting up to 70 mph traffic. We also can neglect any passing cars in other lanes and just focus on one.
     
  17. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    Thanks, this is actually what I am trying to do (use two 'breakbeams' to calculate velocity using distance over time).

    My question is, how could a photodiode be used to effectively do this? We need the sensor to be reasonably far from the road (preferably) and to calculate the speed within 10%. Would we have to sent out a signal which the photodiode would sometimes (when a car was in the way) pick up reflections from?

    Another question is this: I know most police radar speed detectors must be pointed INTO the moving vehicle (for the correct doppler shift and speed measurements). However, I know that baseball speed guns can be held perpendicularly to the moving object. How can this work since they both use radar?

    Thanks again,

    John
     
  18. crutschow

    Expert

    Mar 14, 2008
    13,023
    3,236
    Where did you get that information? Baseball speed guns work just like police radar and the ones I'm familiar with must be used at or near home plate so that the ball is coming directly at the gun to detect the speed accurately.
     
  19. John P

    AAC Fanatic!

    Oct 14, 2008
    1,634
    224
    Now that you say you want to measure vehicle speed, I'm even more convinced that the way to do it is with a TV camera. As I recall, 30mph is 44 ft/sec, so 60 is 88 ft/sec, so let's say the top speed is 100 ft/sec. The scan rate of a camera is 30 frames/sec, so between TV frames a car would travel a max of about 3 feet. If you had a time to start looking for a transition from dark to light or light to dark (time since the beginning of the TV frame, that is) could you compare that transition in two successive frames? I think you could. What you need to do is place the camera at a distance where its field of view gives one scan line a reasonable length of road to cover, based on the length of time variation you feel like measuring. What I mean is do you want 3 feet to correspond to the whole length of a scan line or just a fraction? That's a judgment call. There's also the issue of whether the cars will be seen as light or dark compared to the background. Another design decision!
     
  20. jbord39

    Thread Starter Member

    Mar 31, 2010
    41
    0
    That was my initial idea as well. The problem is that this would require some sort of computer vision software. To detect the edge in an environment with moving clouds and ambient lighting changes could be difficult. That is why I was hoping for a sensor-based approach. What implementation would you use to detect the edges?
     
Loading...