Inductive Loop Vehicle Detector Gets Modernized

Much like George Lucas and the original Star Wars films, many of us may find that our passion projects are never quite finished, especially when new technology comes around or we just want to make some improvements for their own sake. [Muris] was featured a while back for a vehicle detecting circuit, but is back with some important upgrades to his project. (Which, luckily, do not include any horrible CGI aliens.)

For starters, the entire project has been reworked from the ground up. For anyone unfamiliar with the original project, the circuit detected a vehicle via an inductive loop and was able to perform a task like opening a gate. It now has two independent channels which are polled separately, yet has a reduced parts count which should make construction simpler. The firmware has also been reprogrammed, and in addition to sensing a vehicle’s presence can now also measure the speed of any vehicles passing by.

The complete list of improvements can be found on the project page, and an extensive amount of documentation is available on this if you want to try to roll out your own inductive loop vehicle detector. Of course, this isn’t the only way to detect a vehicle’s presence if inductive loops aren’t really your style.

41 thoughts on “Inductive Loop Vehicle Detector Gets Modernized

  1. It’s whe the inductive sensors start interrogating the cars computer when they stop at the traffic lights and they use them instead of speed cameras. Or maybe send a command to the car to reduce the engine speed.

  2. Seems like a good a time’n’place as any to inquire about this. I ride motorcycles and both of my bikes are adventure tourers with a lot of ground clearance and are fairly lightweight. During the daytime with traffic around, never an issue. But if I’m out out at night and no one’s around I always get a few lights that will never trigger for me.

    I’ve tried all the tricks in motorcycle lore about how best to trigger them, parking myself close to the metal lines that protrude through the asphalt. Putting the kickstand down right on top of the metal strips. Rolling myself forwards and backwards trying to find the traffic signals G-spot. But often to no avail and after sitting there 5 minutes watching the signal change over and over but never in my favor. Eventually I hit the ‘F this’ barrier and run the red light (as safely as possible and triple checking Johnny Law isn’t around).

    Other than dragging a cart full of angle iron behind me, anything to be done to make these induction loops detect me?

      1. Certainly have and certainly doesn’t seem to work. I personally believe this is ‘hack folklore’; something that seems plausible that we hear often enough to the point we assume it to be true. Happy to be convinced otherwise.

      2. I find it hard to believe a permanent magnet is going to work; it’s an inductive loop without core material, so it doesn’t respond to DC, and there is no core to saturate. Maybe if the electronic design is really crappy, any induced voltage could trigger the circuit.

        What could work, is placing a few loops of shorted copper wire as low as possible and parallel to the ground; an eddy current will be induced in these, and because the conductivity of copper is much higher than iron, the current should be higher. A bigger coil would be better, since it would couple with more of the magnetic field of the detection loop.

        If that doesn’t work, you could try actively generating a field in a coil close to the ground, but the optimal frequency might depend strongly on the make and model of the detector, and because the inductive loop is probably part of a resonant circuit, it would probably depend on the size of the loop as well. You could use an inductive loop and a good multimeter to measure the excitation frequency of some different detectors, and use a frequency sweep across the frequencies you find.

        Where I live (in the Netherlands), it does happen, and we have a special number we can call to request adjustments. Also, if you get a ticket, it usually is dropped if you argue that the detector wasn’t working.

          1. Argh posted something I didn’t intend to post.

            Taking a short extension cord and placing it over the loop (if you know where it is). Then plug the two ends together.
            This will effectively short the primary and trigger the loop.

    1. Loop detectors work by skin effect, so the surface area facing the loop is what matters, not volume. And of course proximity matters as well. A bike just doesn’t have a big metal surface area, so if the loop detector is not tuned to be super sensitive there isn’t really anything you can do. Stopping on top of the wires is your best bet, though. You could carry a metal plate with you that you put over a corner of the loop, but that’s not super practical.

    2. Your best bet is to use aluminium sheet under the bike, parallel to asphalt. Bigger aluminium sheet surface – better detection. Magnets do not work with inductive loops, but there are also magnetic vehicles on the market and those should be trigerred with magnets.

      1. Actually, with a microcontroller, that should be fairly trivial. It would just take a simple analog front-end and a few MOSFETs or a motor driver like a L6203 or L298 or something similar. The front-end doesn’t need to be much more complicated than a few resistors and a comparator, that will automatically generate a square wave for the microcontroller to measure if there is a sufficiently strong field. If there is no significant field, the output would be either stuck high or low, or highly irregular, both are easy to detect for the microcontroller. After detecting a stable frequency, the microcontroller could generate the same frequency (or be slightly off on purpose) for a few seconds. If you’re off by a little, the phase of the signals will continuously shift, which probably helps detection. You could also sweep slightly below that frequency, because some detectors use the drop in frequency when the loop is loaded for detection.

        I’d expect the frequency to be a few tens or hundreds kHz at most, which should be easy, even for an old 8-bit controller at 16MHz, like standard low-code Atmel/Microchip controllers (Arduino, etc.).

        1. Some 27mhz am transceivers coupled to class c amplifiers have been said to trip the lights. I’ve never tried it myself. I’m thankful for faulty traffic lights because after years of failed driving tests I tried an alternative centre and was stalled at a light for 10 minutes while the instructor fumed.

    3. I’m just looking at this national geographic show (back translated it’s called “system breakers”, don’t know the real english name) about this. He put a neodymium magnet on bottom of the bike and copper wire loop on the rim (better put it somewhere else than rim). Seemed to work for him.

    4. It’s really an installation issue. System designers should account for all types of vehicles likely to be common to the installation site. Figure 8 loops help mitigate the issue. Alternative detector types are also helpful, too. We pushed an AMR-based detector for this reason. With those, a sufficiently strong magnet actually would help with detection, since they look for changes in the local magnetic field. Like most things, a lot of design choice comes down to cost, so often these solutions aren’t implemented. As stated above, your best bet is to position yourself directly above the loop cut, since you actually get minimal change when in the center of the loop. Also, this problem has pushed many states to pass laws allowing motor bikes to forgo waiting for the signal if they ascertain that it is safe to do so, so check your local laws. Not an idea solution from a safety standpoint, but it’s good to know.

    5. For a bicycle, “laying it down” on it’s side trips stubborn sensors for me. I don’t literally have to lay it on the ground — just leaning over to the side without getting off usually works. This is perhaps not practical for a motorcycle. Presumably this works because the bike frame looks like SparkyGSX’s “loop of wire” to the inductive sensor.

    6. Maybe a a coil of wire (kind of like what is in the video) would work. A regular magnet won’t because it doesn’t produce a time varying magnetic field.
      When I started biking somewhat regularly, I found this:
      The idea of using an active device to create a magnetic field that mimics what a large metal object does to a loop sensor’s magnetic field sounds OK to me.
      But my city started putting in bike lanes and paving trails which meant I could avoid having to contend with traffic and loop sensors set up for cars. So I never built the device and can’t say how effective it is.

    7. Cyclist here. Look for the marks with a bike icon. They’re front of the loop, center, usually. Putting a wheel on one should be sufficient; your tires probably have metal in them, and the rim certainly is.

      The people who are truly screwed are people with modern, high end bicycles – carbon wheels are increasingly common not just among triathletes, the tires don’t contain any metal unlike motor vehicle tires, the frames are carbon fiber, the cranks can be, hell, even some bikes have wireless shifting and hydraulic brakes, meaning there’s very, very little on the bike that is metal save the chain, chainrings, and cassette.

    8. Pretty sure it will take an active loop with a minimal processor controlling it in order to handle the variety of loop frequencies I’ve found. State to state the systems vary widely, and vary even county to county or by the city. The entrepreneur that comes up with a working model will have a long line of bikers with wads of cash on their doorstep. A coil that detects the freq in use, then opposes, shifts freq, reinforces, as needed, cycling through methods as needed.

      You can discover documentation of some systems online, but there is no nationwide or even statewide set of standards. Hanging out at an intersection with an o-scope and loop is frowned on by officers… don’t ask me how I know…

    9. Someone told me long ago to kill the bike and restart. The electromagnetic fields thrown off by the starter should trip the inductive loop in the road. I’ve been doing it for many years and it seems to work.

      1. Actually it does work for lights. There are several companies that make camera systems for detecting the presence of traffic. Radar is also in common use today, and seems to be more reliable than cameras where weather varies strongly by season. (Think snow and white cars, for example.)

  3. Neat, with more data a training model (like with the facial or image recognition opensource software though using maybe the code algorithm for images, for the coil signal) and some sort of pattern recognition you could get the system more specific to whatever you’d like. May be other opensource spectral pattern recognition software I am not aware of. Even could make a simple routine to verify each signal against your known signals to identify specific vehicle signals without additional emitters.

    1. Where I am it is legal to violate the light and proceed with caution if 45 seconds has elapsed AND you are CORRECT that it is a controlled traffic light that IS FAILING to detect your vehicle. Now… just ‘splain that all to a judge! But never fear! There’s a legal firm here distributing cards explaining that law to motorcyclists and how to properly act at a malfunctioning light, plus what to say to an officer that doesn’t believe you, and their phone number for legal repre$entation when they ticket you anyways.

    1. Discussion turned to how to trigger the lights when by rights they should be triggered but fail to because the vehicle is not being detected as it should.

      IE: Motorcycles, bicycles, trucks with frame too high off ground.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.