Author Topic: Need help finding a suitable sensor to detect vehicles.  (Read 273 times)

0 Members and 1 Guest are viewing this topic.

Offline fantis1337

  • Contributor
  • Posts: 22
  • Country: is
Need help finding a suitable sensor to detect vehicles.
« on: January 09, 2017, 11:14:46 PM »
TLDR: Need a sensor to detect vehicles at 5-10 m horizontally and 8 m vertically away from the object.

I'm currently working on a long project for university, part of the project includes detecting traffic. Ideally, I would want it to detect all types of vehicles such as motorcycles, lorries and just regular cars. Additionally, I would love if I could sense bicyclists and pedestrians and distinguish them from vehicles but that is not a strict requirement.

I believe image analysis by a camera would be perfect as I could define what it detects but I do not want to use a PC with an operating system such as Raspberry Pi. Can this be implemented on a 32 bit ARM processor ( Teensy 3.2 - Cortex-M4 72 MHz)?

Other ideas include an ultrasonic sensor, radar, hall effect and PIR.  I reckon some work better than others and what I'm asking is for you to push me in the right direction.

I can provide other information If it is relevant.

Aleks

If you run into an asshole in the morning, you ran into an asshole.
If you run into assholes all day, you’re the asshole.
 

Offline McBryce

  • Frequent Contributor
  • **
  • Posts: 853
  • Country: de
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #1 on: January 10, 2017, 12:38:59 AM »
If the vehicles are passing through the sensor area, I'd suggest something very simple such as a vertical array of IR / Laser sensors with a reflector on the other side of the road. The shape detected should be enough to distinguish anything from a person to a lorry as long as there is only one lane of traffic.

McBryce.
 

Offline nfmax

  • Frequent Contributor
  • **
  • Posts: 373
  • Country: gb
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #2 on: January 10, 2017, 12:50:11 AM »
Have a look at the magnetic sensors & vehicle detection IC from http://www.speakesensors.com/ - you may be able to put something useful together at pretty low cost! (I've not tried these myself though - yet)
 

Offline BrianHG

  • Regular Contributor
  • *
  • Posts: 154
  • Country: ca
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #3 on: January 10, 2017, 12:57:57 AM »
When it comes to vehicle counting, I know that on my local streets when the city puts up a monitor, it's just a hollow rubber tube across the street with one side closed, the other side on an air pressure sensor connected to an MCU counting the pressure spikes at the tires roll over the tube.  The tube diameter is usually around 2/3 the diameter of a garden hose and squeezes and rebounds like a rubber band.

This solution works in all weather, day and night, but, to avoid counts on the sidewalk, you need to place the rubber hose across the road and tie it down at each end and in the middle of the road.  Usually a huge U shaped nail hammered into the asphalt.  You might be able to sense the pressure difference between a bike and car only because the bike is so light by comparison.  Multi-axle semi trucks will be counted more than once due to the extra wheels.


« Last Edit: January 10, 2017, 01:00:15 AM by BrianHG »
__________
BrianHG.
 

Offline snarkysparky

  • Regular Contributor
  • *
  • Posts: 66
  • Country: us
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #4 on: January 10, 2017, 01:17:06 AM »
through beam infrared is a very reliable method.  The blocked time could be used to estimate the type of passing vehicle.

http://www.optexamerica.com/security-products/ax-100tfr
 

Offline atflake

  • Newbie
  • Posts: 2
  • Country: us
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #5 on: January 12, 2017, 06:34:09 AM »
If your trying to implement some sort of security barrier/gate then I can add some lessons learned after i helped fix many systems that another company installed.

 Infrared Light curtains and inductive loops seem to work the most reliably in the field for detecting vehicles (moving or stopped). Have integrated many of these fixes in automated gate systems at many different facilities. The advantage with light curtains was cost and 99 percent of the time solved our problems along with a single ir beam as a last ditch saftey method to prevent a 1000lb gate from slamming down on some one at 30mph.

We used a micro controller to monitor the light curtains and inductive loops and the firmware was modified for each unique situation. Loop and light curtain placement was usually different at each location, so timings had to be adjusted to make a system work properly.

Physical location of light curtains (side looking vs top down) to get anti tailgating requirement work. Especially for pedestrians and motorcycles/bikes

Infrared light curtains were very effective at stopping tailgating when combined with in ground inductive loop sensors. (its all about the physical location and timing of each component in relation to each other)

Light curtain sensors must be as low to the ground as possible and account for the largest vehicle height possible (Vehicles with trailers can cause some interesting consequences when the 1000lb barrier arm comes down because the system thinks someone is tailgating!  |O Make sure you really understand the requirements. (every possible situation that can happen if possible). When a trailer hitch is lower than your light curtain the system sees a gap between the truck and the trailer and thinks the trailer is tail gaiting. Not good  |O. Good thing the customer defined the height requirement :-X

Inductive loop sensor placement is key, since sensitivity adjustment is degraded if they are too close together. Changing the number of wraps in your detection loop and specs of the loop amplifier/receiver hardware is also a variable. Also be ware of existing metal in the area before loop placement and take appropriate measurements before install (mock up loops and take inductance measurements onsite). Man hole covers or buried metal pipes can cause problems  |O

Cameras with recognition software were used in some installs, but cost can get prohibitive, and complexity goes way up. (network cable, hardware, expensive enclosures). Plus you have to deal with the notification process of  what to do with the information if it triggers some kind of security issue. If its used to just close the gate then your good, but usually someone wants to know whats going on and under what conditions can normal operation resume.

These are just a few things that come to mind.

-atflake
« Last Edit: January 12, 2017, 06:42:22 AM by atflake »
 

Offline neil t

  • Contributor
  • Posts: 28
  • Country: au
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #6 on: January 12, 2017, 09:02:05 AM »
I've noticed online cheap dopler radar sensors that work with Arduino and the like perhaps that may be worth looking into.
 

Offline fantis1337

  • Contributor
  • Posts: 22
  • Country: is
Re: Need help finding a suitable sensor to detect vehicles.
« Reply #7 on: January 12, 2017, 11:49:33 PM »
Thanks for the replies. Looks like I'm going with microwave detector as I'm mounting it on a street light.
If you run into an asshole in the morning, you ran into an asshole.
If you run into assholes all day, you’re the asshole.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf