Turn your bike horizontal to trigger the green light

Nobody who drives a car ever has to think about the various technologies used to trigger traffic signals. You just drive up to the intersection, and within at most 120 seconds (but probably less) you’ll have a green light.

Anybody who rides a motorcycle or bicycle on the road has encountered “defective” traffic lights — those signals that don’t detect the presence of your bike.

Some of us learn the deep mysteries of traffic signal technology, others rely on myth and magic, while the rest of us just shrug and either run the light or hobble to the sidewalk to push the pedestrian beg button. I think it’s especially pitiful when bike advocates cheer the placement of beg buttons that are nominally placed for cyclists, as if somebody is doing us a big favor. But I digress.

Bike accessible signal light button

Probably the most common vehicle detector used at traffic actuated traffic lights are inductive loops. These are coils of wires embedded into the pavement. You can often tell the location of these loops by seeing the rectangular cuts in the pavement at intersections.

These loops generate a magnetic field, and a chunk of metal traveling through this field induces eddies in the magnetic field, kind of like pitching a rock into a smoothly flowing stream induces eddies in the current. Big rocks, like big cars and trucks with big chunks of metal, induce big eddies. Small pebbles — like bicycles — induce barely detectable ripples.

Bikes are detected because the wheel (not the frame, as is commonly believed) generates these eddies. You can maximize these eddies by turning your bike horizontal. I demonstrate this at an intersection near my home. This works as well for my fancy carbon fiber road bike as well as the steel Schwinn I ride in this video. I should have placed the camera closer to me so you may need to go full screen to see what’s going on.

A word about magnets

That’s the “myth and magic” I alluded to before. Three or four years ago, I bought a set of fancy neodymium rare-earth magnets exactly like these bar magnets. I placed four of these magnets at various places directly over the induction detection loop. Nothing happened. I expected this because of how induction detectors work, but so many people swear by the effectiveness of their magnets I had to test this empirically.

The fact that I expected the test to fail, however, means my test was likely poorly designed. If you suggestions for improving this test, we should do a Mythbusters style video showing the effectiveness of magnets.


  1. Not all signals are triggered. Some, unfortunately, are on timers. No amount of flinging the bike around will make the light change earlier than the timer is set for. So, I digress here: Long ago during college years, our Traffic Engineering prof asked the class, “How long should a ‘red phase’ of the signal be?” One student raised his hand and answered, “Two minutes.” The prof was a little suprised that someone had guessed correctly and he asked, “How did you determine that?” Answer: “That’s about how long it takes me to roll a cigarette.”

  2. Many stop lights in my city (Palo Alto) are now triggered by cameras instead of pavement loops. Laying your bike down makes you less visible to the camera. The cameras are easily visible across the street (during daylight at least). If you’re having a hard time triggering one, stand up straight and square your shoulders toward the camera. Maybe even hold your bike up and wave it at the camera.

  3. @FR: Our local traffic engineer explained to me that while my bike would trigger the signal, the timing of the light change was subject to the timed rotation of signals set by the traffic controller. One signaled intersection (camera detection) I use is a two-lane collector crossing a 4 lane arterial. The light triggers immediately for me at 6 AM on a Sunday morning with no traffic present. At 7:30 AM, Monday morning, I stand there for 120 seconds, give or take, watching 4 lanes of traffic wiz by.

    Two years ago or so the traffic guys in our fair city helpfully responded to my request to tweak the induction loops for several intersections on one of my routes; it worked fantastically for my 700C wheeled hybrid. Then I bought a Brompton folding bike. Now with the Brompton, I can trigger some of the loops on the route with my much smaller wheel, but not the others. Gah. Fortunately, the City seems to be moving to all camera controlled signals. But again, detection doesn’t necessarily mean a rapid signal change will follow. For some signaled intersections, bicyclists should have a separate set of lights & controls.

  4. Wow, it’s interesting to me Richard. Turning a bike horizontal is very interesting. It depends on a variety of triggers though. 🙂 Thank you for your video too.

  5. @Ken: Traffic signals with video cameras work by detecting contrast, and sometimes have trouble finding bikes at night. I’ve had good luck by pointing my bike lights directly at the camera, but they are pretty easy to calibrate just like loop detectors so if it doesn’t work without a special effort it is best to notify the city’s public works as well.

    Another technology that is even more promising (and cheaper than video) is microwave detection, since it doesn’t have to be calibrated as often and can even distinguish between cars and bikes, potentially giving more green time to the latter.

    Important also to keep in mind that for most intersections , either loops or cameras will not be set up to detect bikes all the way at the right curb or in/past the crosswalk, which is where many cyclists end up waiting. Progressive cities are painting those little bike stencils over the “sweet spot”, but many of the cyclists I have talked to see them but don’t realize what they mean. Proper signal technology should really just work for all road users, with no knowledge of the mechanics required.

  6. Ive also thought that some type of indicator light should come on when the loop or video picks up a waiting car or bike at the intersection, so there is no guessing/waiting game regarding how long one should wait before considering it broken and running the red. Beyond that, any countdown timer should start at however many seconds it is going to take before the signal changes, and not just the last 20, 10, or whatever. My take is that most people would be willing to wait and not run the light if they knew it was definitely going to change for them, and if they could tell exactly how long a wait they were in for. From watching both cyclists and pedestrians, with no indicator most people jaywalk or run the red right around 45 seconds every time.

  7. Thanks for the additional feedback, all. Isn’t it wonderful we have to know all of this just to ride across the street? 🙂

    @Wuss912: The workaround is to wind some transformer wire around the rim a couple of times and solder the ends together between the rim and rim strip. More special knowledge!

  8. So would there be a way to make these sensors more sensitive, so that bikes and motorcycles could also be detected? It wouldn’t affect whether cars would trigger the light, but it sure would help anyone else using the roads.

  9. Any new traffic-actuated signal installations in California must detect bicycles and motorcycles per AB 1581. That’s part of what motivates the use of microwave and optical sensors that others have mentioned in the comments. There are thousands upon thousands of traffic lights in California that predate this 2008 law, however, so we get to live with them in the meantime.

    Simply cranking the sensitivity up isn’t always the answer, because then you get traffic signals triggered by cars in adjacent lanes and even random fluctuations in the magnetic field.

Leave a Reply

Your email address will not be published.

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