Advocacy & Safety - Traffic signal bike detection

The Governator of California has recently signed a bill that will require all new traffic actuated traffic lights to detect bicycles and motorcycles. Caltrans is working on new standards for magnetic induction loop actuators for this new law.

Cyclist and Transportation Engineer Bob Shanteau will present a paper to Caltrans next Tuesday on what he believes the new standards should incorporate. A brief overview and the PDF of the paper is available at my website (http://www.cyclelicio.us/2007/10/traffic-lights-and-bicycles-technical.html). Bob has asked for constructive feedback before his presentation next Tuesday.

The paper only addresses inductive loop detectors in the pavement. Video detectors are not part of Bob's presentation. Also, carbon fiber bikes and wheels are not part of the presentation -- carbon cannot be detected by loop detectors.

...carbon cannot be detected by loop detectors.

Yet another reason to keep my old steel bikes... ;)

Also, carbon fiber bikes and wheels are not part of the presentation -- carbon cannot be detected by loop detectors.
This is true even in Portland, where we have loop detectors that are some of the most sensitive around to picking up bikes. They don't pick up bamboo, either.

They don't pick up bamboo, either.

Supposedly, the frame material doesn't really matter -- it's the wheel that's important. For carbon wheels, Bob-the-traffic-engineer suggests winding an insulated copper wire around the rim (under the tire) a couple of times then soldering the ends together for a continuous circuit. He has no idea if that actually works, though.

I've 'tested' the inductive sensors close to where I live by just using my front Aluminum wheel. 90% of them trigger with that alone. These are all Q type I think and I put the wheel directly over the middle cut, same place I stop by bike when riding.

As to the painted marking to show the optimum place, that is great. I have noticed that after resurfacing that the cuts show up thru the resurfacing about 1 week later and become very obvious by 4wks later.

One nit in the paper (which I only glanced thru) is that I stop in the left side of left turn lanes when making a left, not the right side - that is done so I can see oncoming traffic without being blocked by the vehicles in the opposing left turn lane. I also stop in the left side or middle of thru lanes and avoid stopping on the right side to avoid right hooks.

One thing that would be fantastic is if there was a positive feedback that the sensor has been triggered. This is not easy, but perhaps LEDs on light posts could be arranged to be visible to cyclist without causing confusion with being a red/green light, but how to arrange them to show if thru or left turn lane is triggered, I don't know.

Al

Bob's paper looks good.

Much like noisebeam's train of thought, I suggest changing the following:

Even though a Type Q loop is excellent at rejecting vehicles in the adjacent lane, the Caltrans Standard Plan shows that Type Q loops are 6' wide and centered in the lane, the same as Type A loops. This means that a cyclist riding in a travel lane who stops toward the right side of the lane will be outside the loop and therefore not be detected. But since a Type Q loop rejects vehicles in the adjacent lane much better than Type A, B or E loops, it can be built wider and located closer to the lane line and therefore in a location where a bicyclist can be expected to stop.

I recommend that cyclists stop near the center of the lane at most signalized intersections, not the right side. The right side invites a right hook by right-on-red traffic if there is no right-turn-only lane, and invites too-close passing if the lane is not extra-wide, i.e. under typical conditions. A bike detection stencil positioned over the center of the lane on top of the center wire of a quadrupole loop encourages better positioning of the cyclist in the lane when arriving at intersections.

If a cyclist avoids the center of the lane for some reason associated with other traffic in that lane, the other traffic will trip the sensor.

One other point:
Something I learned from a seminar given by George Palm in 1987 was always to connect loops in
series.....The implication for bicycles is that connecting loops in parallel reduces the change in inductance from a
bicycle by about half, making it harder to detect. Therefore loops should always be connected in series.

Multiple loops in series will also make the bicycle harder to detect compared to a single loop attached to a typical deltaL/L detector. If the loops are small enough, this will not be a problem. However, if one or more of the loops are large, the percentage of the magnetic field interrupted by the bicyclist will become insufficient for typical detectors. Here in NC, the standard loop sizes on the approach are around 60' long - too long for series connection with a small head loop using a typical deltaL/L detector. I recommend that this paragraph caution the reader about any combined wiring (series or otherwise) and refer the reader to the later paragraph on detector types.

I also second the recommendation that if multiple loops are to be used, especially if they are large compared to the bicycle footprint, separate detector circuits should be used, with their outputs logically "OR"ed together.

Carlsbad and Encinitas generally use circular inductive loops. I can trip most of these signals by stopping with each rim over a point on the last circle and then lowering a pedal over a third point on the circle. (I use aluminum rims, steel frames, and aluminum caged / steel axled pedals exclusively.) If I cannot trigger a light using this technique, I send an email to the traffic department and have had 50 to 70% success in getting the situation remedied.

The Governor's endorsement is a most welcome additional tool in my fight to make all intersections bicycle-friendly.

Brilliant. Now to get this in Louisiana. I love sitting for ten minutes at a light, watching it cycle every way but mine. I have a steel road bike and a steel mountain bike, and no amount of leaning the bike will trip the sensors. Hell, you'd have a hard time convincing me they could see my old Suzuki Samurai.

Time to lash a Ford to the back of my bike, so I can get through town.

Brilliant. Now to get this in Louisiana. I love sitting for ten minutes at a light, watching it cycle every way but mine. I have a steel road bike and a steel mountain bike, and no amount of leaning the bike will trip the sensors.

What kind of inductive sensors do you typically find where you live?
Al

What kind of inductive sensors do you typically find where you live?
AlI was referring to the bill. I have no idea what sort of sensors they put in the road here. All I know is that there are squares cut in the pavement (I might take a photo later), and the light only changes when a car shows up behind me.

I was referring to the bill. I have no idea what sort of sensors they put in the road here. All I know is that there are squares cut in the pavement (I might take a photo later), and the light only changes when a car shows up behind me.

You can usually tell by the cut out shape - a double loop is Q(uadrapole) a single is D(ipole). This page gives advise on where to place your bike to most likely trigger the sensor depending on the type:
http://www.humantransport.org/bicycledriving/library/signals/green.htm

The paper referenced in the OP does as well, but less clearly.

Al

You can usually tell by the cut out shape - a double loop is Q(uadrapole) a single is D(ipole). This page gives advise on where to place your bike to most likely trigger the sensor depending on the type:
http://www.humantransport.org/bicycledriving/library/signals/green.htm

The paper referenced in the OP does as well, but less clearly.

Al

That sort of works... except in my area where the old loop is still in the street and displaying its pattern, and the new loop has been patched in above it, and now displays its pattern... it can be somewhat difficult to tell what loop is what, and where the sweet spot exists.

Of course in my case it hardly matters as few of the local loops actually will trigger the light even with my heavy old steel cruiser bike. (and yes, I have notified the street department)

I can't wait to see if the Govenator's edict will have any effect on the local loops.

You can usually tell by the cut out shape - a double loop is Q(uadrapole) a single is D(ipole). This page gives advise on where to place your bike to most likely trigger the sensor depending on the type:
http://www.humantransport.org/bicycledriving/library/signals/green.htm

The paper referenced in the OP does as well, but less clearly.

Al
Thanks for that link. It looks like we have diagonal quadrapoles around here. Last time I tried reporting a situation like this, I ended up in "other department" calling hell. I think these people hide from the public - for good reason, I suspect.