Saturday, May 3, 2008

Review of NCHRP Report 500, Guidance for Implementation of the AASHTO Strategic Highway Safety Plan

I just took a look at NCHRP Report 500, Vol. 18, titled A Guide for Reducing Collisions Involving Bicycles - Guidance for Implementation of the AASHTO Strategic Highway Safety Plan. The lead authors are Craig Raborn of the University of North Carolina Highway Safety Research Center and Darren J. Torbic of the Midwest Research Institute.

In the introduction, the report states

Bicyclists are recognized as legitimate roadway users. ... The safety interests of bicyclists are sometimes in conflict with the interests of motorists. This conflict arises primarily from the substantially different characteristics of the two modes of transportation. Although bicycles can be ridden on most types of roads, the design interests of accommodating higher motor vehicle traffic volumes and speeds during peak hour congestion may create conditions that are less safe for bicyclists. This guide includes road treatments, countermeasures, and other options that support a balanced transportation system.

Safety concerns can significantly influence a person’s decision to bicycle for transportation or recreation. Bicyclists inherently understand that they are vulnerable road users. However, understanding bicyclist safety issues has proven difficult for engineers, planners, and facility designers. Traditionally, safety problems have been identified by analyzing police crash reports, and improvements have been made only after crashes have occurred. Such methods are not sufficient to fully understand and effectively address bicyclist safety concerns; waiting for crashes before responding with countermeasures carries a high price because many bicycle crashes tend to be severe.

First the report says that "bicyclists are legitimate roadway users," then it says that the "design interests of ... higher motor vehicle traffic ... speeds ... may create conditions that are less safe for bicyclists." So the report's stated motivation is safety for bicyclists. (The report says nothing explicit about slower bicyclists delaying higher speed motor vehicles.)

The closest thing I could find of an analysis of why higher traffic speeds create a hazard for bicyclists was this on page III-1:

Speed influences both the severity of crashes that occur as well as the likelihood of occurrence, and has been identified as a contributing factor in all types of crashes. ... Bicyclists are vulnerable road users, and the impact of higher speeds on crash severity is obvious.

Also, in the discussion of Share the Road signs:

The safety effectiveness of shared roadway signs has not been evaluated, and their overall use is thought to be decreasing. Some experts now feel that they are only appropriate in “pinch point” locations where roadway facilities may not fully accommodate both bicyclists and motorists. ... These signs are typically placed along roads with significant bicycle traffic but relatively hazardous conditions for riding, such as narrow travel lanes with no [paved] shoulder, roads or streets with poor sight distance, or a bridge crossing with no accommodation for bicycles.

So hazardous conditions for bicyclists include narrow travel lanes with no paved shoulder. The report does not explain why, other than to say, "the impact of higher speeds on crash severity is obvious."

In analyzing the risks of various types of bicycle accidents, the report states that in a study done in 1996, motorist overtaking cyclist accidents made up 8.4% of total reported bicycle crashes.

The report describes this development of crash type methodology:

The crash typologies developed by Cross and Fisher, by NHTSA, and in the FHWA study evolved into the development of an automated crash typing software, the Pedestrian and Bicycle Crash Analysis Tool (PBCAT) (Harkey et al., 1999), which is currently being further refined for version 2.

Crash Type 11 is

Motorist overtaking bicyclist Description—The motorist was overtaking the bicyclist at the time of the crash.

One of the accompanying figures shows a motorist at night coming upon a lighted cyclist riding on the right edge of a narrow lane in a sharp right hand curve. The implication is that the motorist is about to strike the bicyclist from the rear. In the countermeasures for such an accident scenario, nothing is said of the bicyclist controlling the lane or of the motorist violating the basic speed law by driving too fast for conditions.

The other figure shows a car starting to pass a curb hugging cyclist on a narrow two lane road with a car coming the other way. Again, the report does not propose controlling the lane as a countermeasure. And the report says nothing about motorists thinking that is OK to pass bicyclists anywhere, any time, under any circumstances.

Because the report contains no analysis of the risk of a cyclist controlling a narrow lane being hit from behind, I consider it flawed. It repeats the mistaken analysis that because bicyclists who are hit from behind are at greater risk of serious injury or death, all bicyclists riding in narrow lanes are at greater risk. And because it doesn't analyze overtaking accidents in more detail, it does not address the possible countermeasure of controlling a narrow lane. More seriously, it does not address the countermeasure of integrating bicyclists with traffic.

Instead, the report focuses on countermeasures such as bike lanes and bike boxes that segregate bicyclists from traffic. Until we can gain some influence in the preparation of research reports like this, we will have a limited influence on unsafe practices like the installation of bike boxes in Portland and NYC.

Bob Shanteau

Robert M Shanteau, PhD, PE
Consulting Traffic Engineer
13 Primrose Cir
Seaside, CA 93955-4133
Voice: (831) 394-9420
Cell: (831) 917-0248
FAX: (831) 394-6045

Sunday, April 6, 2008

Report on April 2 AB 1581 Subcommittee meeting

The AB 1581 Subcommittee held its second meeting on April 2. Recall that AB 1581 requires that new and modified traffic actuated signals in California be able to detect bicycles and motorcycles, but only after Caltrans adopts guidelines for detection and signal timing. The AB 1581 Subcommittee will be presenting its recommendations to the

California Traffic Control Devices Committee to consider at its next meeting on May 29, 2008, in the CSAA building in San Francisco. The CTCDC in turn will recommend changes in the California Manual on Uniform Traffic Control Devices, which is administered by Caltrans. So actual changes are still many months and many meetings away. But the AB 1581 Subcommittee pretty much finished its work at the April 2 meeting.

Attending the meeting in person were Ahmad Rastegarpour, Kai Leung, Richard Haggstrom, and David Priebe from Caltrans, James Lombardo from ABATE (motorcycle advocacy), Damon Curtis from San
Francisco, and me. David Roseman from Long Beach and Sean Skehan from the City of Los Angeles attended by telephone.

The meeting was very productive and we agreed on the following recommendations:
  1. Bicycle/motorcycle detection will be performance-based. At least 95% of bicycles and motorcycles that stop within the detection zone (defined below) shall be detected.

  2. The default detection zone shall be a 6'x6' box at the the stop line centered within the lane, except that if the lane is more than 12' wide, the edge of the box shall be 3' from the lane line. If the detection system is capable of detecting bicycles and motorcycles anywhere within the default detection zone, then no marking is necessary.

  3. A detection system that is capable of detecting the reference bicycle/motorcycle plus rider in a detection zone at least 3' wide but less than 6' wide shall be allowed, but only if a Bicycle Detector Symbol is placed at the center of the narrower detection zone.

  4. The reference bicycle/motorcycle and rider shall consist of a folding bicycle with non-ferromagnetic frame and 16" wheels with aluminum rims and stainless steel spokes, such as the Dahon Curve SL, plus a person 4' tall weighing 90 pounds standing over the bicycle.

  5. To be allowed for use at actuated traffic signals in California, a detection system will need to be certified by its manufacturer to meet the performance requirements of detecting the reference bicycle/motorcycle and rider stopped in either the default detection zone or the narrower detection zone at least 95% of the time under actual operating conditions.

  6. For the purposes of setting up the detection at a new or modified signal, a signal technician may choose to use an equivalent rim/rider consisting of a 20" BMX aluminum rim mounted in a wooden frame and set vertically 1.5" above the pavement plus a plywood cutout representing a person 4' tall weighing 90 pounds standing over the rim.

  7. Caltrans will approach manufacturers of carbon fiber rims to request that they either include a shorted coil of wire in the rim during the manufacturing process or else attach a sticker notifying the user that the rim will not trip inductive loops at traffic actuated signals.

  8. The minimum green time at traffic actuated signals shall be long enough such that, when combined with the yellow and all-red times, most bicyclists starting from a stop at the beginning of green will have enough time to substantially clear the intersection.

  9. The Subcommittee decided not to address the all-red time as a separate issue. Their reasoning was that CVC Section 21451 already requires that drivers facing a green signal "shall yield the right-of-way to other traffic and to pedestrians lawfully within the intersection or an adjacent crosswalk." Thus bicyclists who enter an intersection toward the end of yellow (which is legal in California) may not have enough time to substantially clear the intersection before conflicting traffic receives a green signal.
I presented these recommendations to the California Bicycle Advisory Committee the following day. CBAC made some helpful suggestions, one of which resulted in a refinement of the reference bicycle/motorcycle definition. CBAC also decided to address the issue of the many traffic actuated signals that are not addressed by AB 1581 by recommending to the CTCDC that the CA MUTCD provide for a 10 year compliance period for all signals in California to meet the recommended bicycle/motorcycle detection guidelines.

Report on March 4 AB 1581 Subcommittee meeting

At its January meeting, the California Traffic Control Devices Committee (CTCDC) created the AB 1581 Subcommittee to develop recommendations for the CTCDC to consider at its next meeting on May 29, 2008, at the CSAA building in San Francisco. The AB 1581 Subcommittee held its first meeting on March 4 after two cancellations and a reversal of the second cancellation.

The on-again off-again scheduling, along with some incorrect email addresses, hurt the attendance. Attending the meeting in person were Ahmad Rastegarpour, Kai Leung, and Heather Loebs from Caltrans, James Lombardo from ABATE (the motorcycle advocacy group that sponsored the bill), Damon Curtis from San Francisco, and me. The invitation to David Roseman from Long Beach went to the wrong email address but fortunately he was reached by telephone. Neither Ken McGuire, the Caltrans Bicycle Coordinator at headquarters nor his assistant David Priebe attended, which I later learned was because neither one received word that the March 4 meeting was back on again. An invitation went to Sean Skehan from the City of Los Angeles but I later learned that it was sent to an outdated email address. Unfortunately, Mr. Skehan could not be reached by telephone that day.

Here is the agenda that was emailed with the invitations:
  • Introductions
  • Identify Deliverables
  • Document Content (Standards, Specifications, and Guidelines)
  • Type of Document
  • Level of Detail
  • Current Detection Standards (Type "D" Inductive Loop, VDS)
  • Other Standards
  • Adopt/Develop Detection Standards, Specifications, and Guidelines
  • Schedule
Ahmad Rastegarpour chaired the meeting. He identified the deliverable as a recommendation to the CTCDC for changes to the California Manual on Uniform Traffic Control Devices (CA MUTCD) to address the changes required by the passage of AB 1581, specifically with regard to detection of motorcycles and bicycles as well as signal timing for bicycles at new and modified
traffic actuated signals. Existing traffic actuated traffic signals for which the detection is not being modified are not within the subcommittee's scope.

Damon Curtis discussed the problems that the City and County of San Francisco has with inductive loops including: (1) bicyclists and motorcyclists not knowing where to stop within a loop in order to be detected, and (2) loops being short-lived because of utility cuts and pavement failures. Video detection is limited by fog and a lack of clarity in coastal areas. Therefore SF is experimenting with Sensys Networks wireless detectors. These are fist-sized battery powered magnetic detectors that are placed in a 6" hole cut in the pavement and covered with sealant. The internal battery lasts about 10 years after which another hole is cut and a new detector installed. When a vehicle passes over the detector, the detector sends a radio signal to a roadside sensor that then forwards the detection to the signal controller.

Mr. Curtis reported on some tests that SF had conducted on the Sensys detectors with bicycles. He later sent out a report and a spreadsheet on the test. He found that at a sensitivity setting that just rejected vehicles in the adjacent lane, a conventional bicycle with metal wheels as well as a Dahon folding bicycle could be detected 18" away and a simulated carbon fiber bicycle 6" away.

Kai Leung reported that he also had performed a test on the Sensys detectors, but at the default sensitivity setting, which was quite a bit lower than the setting that Damon Curtis had used. Mr. Leung found that a conventional bicycle could be detected 6" away but a "fancy" bicycle [which turned out to be a Trek 5200 with a full carbon frame and fork and aluminum rims] owned by a co-worker could not be detected until a ceramic magnet was placed in its water bottle holder.

David Roseman encouraged the subcommittee to recommend performance-based standards rather than be specific to a particular technology. He said that when loops work, they are 99.9% accurate but that video detection is not that good. He did not want to preclude adoption of other detection technology besides inductive loops. Ahmad Rastegarpour said that the technology could be handled in Caltrans Standard Plans and Specifications rather than in the CA MUTCD.

Mr. Roseman discussed the need for detection of bicycles and motorcycles to be added to the preventive maintenance checklist performed by signal technicians. Various levels of preventive maintenance are performed at different intervals, and it was generally agreed that detection of bicycles and motorcycles should be checked at least once a year.

I brought up the need to adjust the sensitivity of detectors to pick up bicycles and motorcycles when an actuated traffic signal is first turned on. I noted that currently the usual procedure is to start at the lowest sensitivity and increase it until cars are detected reliably. That results in poor bicycle and motorcycle detection. I recommended that instead the procedure should start at the highest sensitivity and decrease it until vehicles in the adjacent lane are no longer detected. Then a check should be made to insure that a typical bicycle is detected. David Roseman said that Long Beach sets the sensitivity the way I described and was pleased with the results.

The subcommittee then discussed the need to define a "reference vehicle" for the purposes of detecting bicycles and motorcycles. I volunteered to investigate this issue and report back at the next meeting. Among other things, I will check with Sensys Networks to find out just how their detectors work and what materials their detector can pick up.

I then gave the presentation that I had planned to give at the January CTCDC meeting but could not because it was not a public input item. I stated that bicyclists are having trouble being detected now and are not in favor of waiting for decisions on signal timing before implementing changes in the CA MUTCD to improve bicycle detection. James Lombardo, lobbyist for the motorcycle advocacy group ABATE agreed.

I went over how inductive loops detect vehicles like bicycles and motorcycles, which appear to a loop as a vertical conducting plate, as opposed to cars and trucks, which appear as a horizontal conducting plate. And because a bicycle's or motorcycle's wheels are closer to the ground than the frame, the inductive loop detects the metal rims rather than the frame or other components. Probably the hardest vertical vehicle to detect with a loop is a bicycle or motorcycle with small wheels, since the effect of the induced electrical currents would be smaller than with a larger wheel. A folding bicycle or small motorcycle should therefore be the reference vertical vehicle. I agreed to do some testing not only with loops but also with video detection and other detection technologies to see if that is true. Reno A&E, a manufacturer of loop detectors, provided test equipment and 3 detectors to me a few years ago that I have used to conduct testing of bicycle detection in the past and can use again to conduct this testing.

I covered how bicycles and motorcycles are detected best over a conventional Type A (6' square) loop when they stop on top of a sawcut and cannot be detected when they stop in the center of the loop. For a conventional loop that is not visible, a marking such as a Bicycle Detector Symbol is needed to show the bicyclist or motorcyclist where to stop in order to be detected. David Roseman said that he had not received any complaints from motorcyclists despite the lack of markings, and I suggested that may be because motorcyclists are trained to ride in either the right or left wheelpath and therefore just happen to stop on top of the sawcut of a Type A loop. James Lombardo says that he hears numerous complaints from motorcyclists in the Palm Springs area that they are not detected. I asked him to request motorcyclists there to send me photos of locations where they are having trouble so I can perhaps diagnose the problem.

Currently the only loop configuration in the Caltrans Standard Plans that will pick up a bicycle or motorcycle with metal rims across its width is a Type D loop. This is a diagonal quadrupole created with straight sawcuts. I presented several other designs for diagonal quadrupole loops, including the quadracircle, which is a circular loop with one or two diagonal sawcuts with the wire placed in a figure-8. I noted that the detection zone of a diagonal quadrupole loop falls off very quickly outside the loop, so that it can be made larger than a conventional square or round loop and still reject vehicles in the adjacent lane.

A non-metal wheel, such as one made of carbon fiber or plastic, cannot be detected by an inductive loop. Such a wheel can be made to be detected, however, by wrapping several turns of fine wire around the rim and splicing the ends together. I indicated that bicycle and motorcycle advocates will need to educate manufacturers, repair shops and riders of this fact.

The last major issue in my presentation was the location of the detection zone. Currently, the Caltrans Standard Plans show no inductive loops wider than 6', so bicyclists and motorcyclists who stop outside such loops are not detected. But diagonal quadrupole loops can be constructed larger than 6' and still reject vehicles in the adjacent lane, thus creating a larger detection zone for a bicycle or motorcycle to stop in.

I ended my presentation with the following recommendations: (1) Immediately require Bicycle Detector Symbols at all traffic actuated signals, (2) Eventually retrofit all traffic actuated signals either with diagonal quadrupole as the head loop, appropriately located and sized, or out-of-pavement bicycle detection, (3) Educate bicyclists and bicycle manufacturers of need to use several turns of wire on non-metal bicycle wheels, (4) Include diagonal quadrupole loops in future research, (5) Include additional diagonal quadrupole loops (e.g., quadracircle) on Caltrans standard plans, and (6) Continue research into inductive loops, out-of-pavement detection and extended green intervals for bicyclists.

David Roseman in particular was interested in the quadracircle loops that I showed. He asked me if there were any operating examples of quadracircle loops with two diagonal sawcuts, and I told him that I had seen one in Monterey that had been cut by a contractor. When I saw it, I recognized that it was an improved version of the Caltrans Type D loop.

We agreed that for all new and modified signals, the "default detection zone" would be a 6' square centered in the lane at the limit line, in which case no special marking would be required. For a bike lane, the default detection zone would be 5' wide. If the actual detection zone were smaller than this in either a regular lane or a bike lane, a Bicycle Detector Symbol would be required, in which case the detection zone would be at least 3' wide, centered on the Bicycle Detector Symbol.

The subcommittee was concerned about how close the Bicycle Detector Symbol should be placed adjacent to a high speed lane out of fear that a waiting cyclist might be struck by a passing car or truck. It was decided that the Bicycle Detector Symbol would be placed no closer than 2.5' from the middle of the lane stripe. Furthermore, the preferred position for the Bicycle Detector Symbol in a through or left turn lane would be 1.5' to the right of the center of the lane.

Ahmad Rastegarpour reported that Caltrans was working on a Traffic Operations Directive to be sent to the Districts that would direct them to install Type D loops or video detection in all lanes at all new or modified traffic signals and to install Bicycle Detector Symbols in the rightmost left turn lane and, if there was no bike lane, in the rightmost through lane, or if there is a bike lane, in the bike lane itself.

Ahmad Rastegarpour reported that PATH had been unsuccessful in its attempts to distinguish between bicycles and other vehicles for the purposes of signal actuation and that Caltrans was redirecting PATH to the problem of signal timing for bicyclists. PATH is settling on a startup time of 2.5 sec and an average speed of 12 mph to cover the majority of bicyclists, as they reported to the CTCDC in January. They are concentrating on the length of the initial green interval rather than on the length of the all-red, however.

At the end of the meeting, Kai Leung asked whether bicyclists and motorcyclists would be willing to place radio frequency (RF) tags on their machines for the purposes of traffic signal actuation. This would fit in with a new effort among transportation professionals to automate the transportation system using Vehicle Infrastructure Integration (VII). Traffic signals would be equipped with transponders that detect any RF tags in the area. This would eliminate the need for inductive loops to detect motorcycles and bicycles. If other motor vehicles had RF tags as well, it would eliminate the need for loops altogether, which is an important objective for Caltrans. Obviously, this change would not happen overnight, but Caltrans is looking 10 or 20 years down the road to make sure that motorcycles and bicycles are not left out of the next detection technology like they were with inductive loops from the 1960's to now. I agreed to coordinate the effort to collect the feedback and report back to the subcommittee and its April 2 meeting.

Sunday, February 10, 2008

Barrier posts (bollards) on bike paths in California are overused

The California MUTCD reads:

Section 9C.101(CA) Barrier Posts on Class I Bikeways

Before a decision is made to install barrier posts, consideration
needs to be given to the implementation of other remedial measures, such
as Bike Path Exclusion (R44A(CA)) signs (see Section 9B.07) and/or
redesigning the path entry so that motorists do not confuse it with
vehicle access.

It could be necessary to install barrier posts at entrances to bike
paths to prevent motor vehicles from entering. When locating such
installations, care needs to be taken to assure that barriers are well
marked and visible to bicyclists, day or night (i.e., install reflectors
or reflectorized tape).


An envelope around the barriers should be striped as shown in Figure
9C-2. If sight distance is limited, special advance warning signs or
painted pavement warnings should be provided. Where more than one post
is necessary, 1.5 m (5 ft) spacing should be used to permit passage of
bicycle-towed trailers, adult tricycles, and to assure adequate room for
safe bicycle passage without dismounting. Barrier post installations
should be designed so they are removable to permit entrance by emergency
and service vehicles.

Generally, barrier configurations that preclude entry by motorcycles
present safety and convenience problems for bicyclists.

Such devices should be used only where extreme problems are encountered.

Note that Figure 9C-2 shows no barriers are shown in the path itself.

Chapter 1000 of the Highway Design Manual is still current (it has not
been superseded by the California MUTCD). It states:

Topic 1003 - Design Criteria

1003.1 Class I Bikeways

(15) Barrier Posts. It may be necessary to install barrier posts at
entrances to bike paths to prevent motor vehicles from entering. For
barrier post placement, visibility marking, and pavement markings, see
the MUTCD and California Supplement, Section 9C.101.
Generally, barrier configurations that preclude entry by motorcycles
present safety and convenience problems for bicyclists. Such devices
should be used only where extreme problems are encountered.

It would appear that based on the guidance in both the California MUTCD
and the HDM, barrier posts (bollards) are overused on bike paths in California.

Saturday, February 9, 2008

Two bicycle items (plus one) at January 31, 2008, CTCDC meeting

Three items were on the agenda California Traffic Control Devices Committee (CTCDC) that were of interest to bicycling advocates:

1. The Committee considered a proposal to amend the policy of the Speed Limit Sign contained in the California Manual on Uniform Traffic Control Devices (MUTCD). This applies to bicycling because one of the tools for improving traffic conditions for bicyclists is to control the traffic speeds. Here is the agenda item.

The proposal would clarify the procedures for Engineering and Traffic Studies (E&TS) to set the speed limit within 5 mph of the 85th percentile speed of free flowing traffic, providing that the speed limit could be set 5 mph lower if indicated by unusual conditions not readily apparent to drivers, but in any case not below the 50th percentile speed.

Several cities raised objections to the proposal based on the fact that speed distributions are typically so narrow that the first 5 mph increment above the 50th percentile speed may also be above the 85th percentile speed. Since previously the speed limit was set at the first 5 mph increment below the 85th percentile speed, they would be required to raise the speed limit on a substantial percentage of their streets.

The Committee asked about the leeway that police give in enforcing speed limits. I told them about a meeting I had with the Traffic Commissioner when I first started as Traffic Engineer for the City of Monterey in 1990 to go over our E&TS procedures. He said our procedures were fine, but then I asked him about the latitude he gave for speeding citations in his court. He said he gave 12 mph over the speed limit. I asked why such a large number, and he said that traffic citations are handled as criminal matters and the burden of proof was beyond a reasonable doubt. He said that there was considerable doubt about the accuracy of radar, and this doubt required a large latitude.

The Committee members were taken aback that the latitude allowed by traffic court judges was this large, but other traffic engineers in attendance corroborated my story. The co-chair asked if the 12 mph latitude applied to freeways with speed limits of 65 mph, and I had to admit that I did not know, because Monterey's highest speed limit was 35 mph.

I then said that we could think of speed enforcement as a team consisting of traffic engineers, police, and judges, but that although engineers were setting speed limits honestly and police were doing the best they could to enforce them, the judicial system was dropping the ball. I said that if the traffic court judges used a smaller latitude, then we would not be having this argument. I said that the judges were basing their decisions on some anti-radar propaganda, including the myth of a radar gun measuring a tree moving at 40 mph, and said that we need to educate the judges on the accuracy of radar.

During the next break, George McDougall, Statewide Radar Coordinator for the CHP introduced himself to me. He told me about the newest radar guns that take two readings, one of the vehicle and one of the ground, and then calculate the difference. This allows for enforcement from moving vehicles. It is also extremely accurate, certainly to within 3 mph.

The CTCDC approved the new wording, including the 50th percentile floor. This next goes to Caltrans which will make the decision of whether to include the wording in future editions of the CA MUTCD.

2. San Francisco applied for permission to install signs to warn bicyclists that motorists may be making illegal right turns. Here is the agenda item.

Jack Fleck, SF's Traffic Engineer, and Damon Curtis, Associate Engineer Bicycle Program, attended the meeting. Here is Mr. Fleck's presentation.

Mr. Fleck described the problem as bicyclists who pick up speed on this downhill part of Market are being right-hooked (my term, not his) by drivers making illegal right turns. His presentation contained pictures of no right turn signs, delineators (flexible cylindrical posts) and a concrete median to try to prevent motorists from turning right at Octavia. Despite all this, he said that some automobile drivers are slowing to a crawl and making the turn anyway in front of bicyclists who are traveling fast on this downhill and passing cars on the right that they do not think will be making an illegal right turn at Octavia.

He proposed two alternates for signs, one with wording that says: BICYCLISTS WATCH FOR PROHIBITED RIGHT TURNS, and the other that says: BICYCLISTS WATCH FOR CARS MAKING PROHIBITED RIGHT TURNS.

I had questions about the wisdom of having a bike lane at that location, particularly because bicyclists could not leave the bike lane to pass slow motor vehicles on the left because of the delineators and median. I recommended a wide outside lane instead to allow bicyclists to integrate with motor vehicles. Jack Fleck responded that the SFBC considered that idea but rejected it out of concern for cyclists who were not comfortable "mixing it up with traffic."

The CTCDC approved SF's request for experimentation, with the proviso that the experiment include several alternative schemes, including red light cameras to catch illegal right turns (which would be allowed by AB 23).

After I got home, I thought of a different legend for a sign from the one proposed: BICYCLISTS - NO PASSING ON RIGHT IN INTERSECTION, perhaps with a picture of a car making a right turn in front of a bicyclist, such as this sign on the Stanford University campus:

I think that the message BICYCLISTS - NO PASSING ON RIGHT IN INTERSECTION plus the picture above get the message across to the bicyclist exactly what the hazard is and what to do about it. If a car in the right travel lane is slowing down approaching Octavia, then anticipate that it is about to make a right turn and STOP!!!!

I have suggested to SF and SVBC that a sign with the wording and picture above be included in the experiment.

3. Caltrans presented its plans to address AB 1581 concerning detection of bicycles and motorcycles at traffic actuated signals. Here is the agenda item.

The CTCDC chair said that this was a discussion item only and that public input would occur at their next meeting on April 24.

Ahmad Rastegarpour from Caltrans' Electrical Systems Branch gave this presentation on Caltrans' plans.

Mr. Rastegarpour stressed the Caltrans emphasis on distinguishing bicycles from motor vehicles for the purpose of providing additional minimum green time.

Then Dr. Steven Shladover from UC Berkeley gave this presentation on bicycle crossing times.

Dr. Shladover's email transmitting his presentation contained the following proviso: "I would caution that these presentation slides were not intended to be a stand-alone document for general distribution and are not entirely self-explanatory, but were meant to be shown with a narration that fills in additional information not shown here."

Members of the CTCDC expressed alarm over the implications of Dr. Shladover's findings on traffic signal operations, particularly on the potentially large (8 sec was mentioned) all-red times.

I was not allowed to give the presentation that I had prepared, although printed copies that I prepared were distributed to the Committee.

Instead, I said that neither the bicyclist nor motorcyclist communities had been given an opportunity to provide input on Caltrans' plans. I said that I had given a presentation in October to the Electrical Systems Branch, but that Mr. Rastegarpour was not there that day, and that I had given a copy of the presentation to Caltrans during the November Bicycle Advocacy Summit with the Director. I pointed out that a member of the motorcycle community was also at the meeting (I had notified ABATE and AMA, and Chuck Pederson from ABATE was there). I said that there were errors in Caltrans' presentation, including their assertion that motorcycles were detected by existing loops.

The CTCDC chair then directed that a subcommittee be formed to work on the plan. Mr. Rastegarpour will chair the committee, and members will include representatives from the Cities of Los Angeles, Long Beach, (San Francisco?), along with representatives from the bicycling community (me) and the motorcycling community (to be selected).

Afterwards, Chuck and I spoke with Mr. Rastegarpour outside the meeting room. Mr. Rastegarpour asked me to call Ken McGuire and tell him about the formation of the subcommittee, which I did. I told Mr. Rastegarpour that I would be in Sacramento next Thursday for the CBAC meeting and Friday afternoon for the Bicycle Advocacy Summit, and requested a meeting with him on Friday morning.

Monday, January 21, 2008

FAQ Rec.Bicycles Frequently Asked Questions Posting Part 4/5 Traffic Detector Loops

A traffic loop detects metal objects such as cars and bicycles based on the change in inductance that they induce in the loop. The loop is an inductor in an LC circuit that is tuned to resonate at a certain frequency. A metal plate over the loop (like a car) causes the magnetic flux to be shorted, reducing the inductance of the loop. This causes a change in resonant frequency, which is detected and sent to the signal controller. One of the ways of testing a loop is to create a loop about 2 feet in diameter with several turns of wire (connecting the ends) and placing the test wire in the middle of the traffic loop. The test wire should cause a detection, if all is working.

The same effect is seen with a vertical piece of metal, such as a bicycle, but is weaker. Because aluminum conducts electricity quite well, aluminum rims help. Steel rims are OK. Non-metal rims cannot be picked up at all. A bicycle with aluminum rims will cause about 1/100 the change in inductance of a car.

It is always possible to set a detector's sensitivity to pick up a bicycle. The trade-off is in longer detection times and the possibility of false detections from vehicles in adjacent lanes. Most people who set signal detectors use the lowest sensitivity setting that will pick up cars reliably.

I advocate using the highest setting that will avoid picking up vehicles in adjacent lanes. Digital circuits used in modern detectors can use high sensitivity settings without unacceptable increases in detection times. Unfortunately, there are still a lot of old detectors out there, and most people who work on signals use principles based on the performance characteristics of old detectors.

In any case, bicyclists should, as a general rule, place their wheels over one of the slots to maximize their chance of being detected. That is where the magnetic field perpendicular to the wheels is strongest. Bouncing the bike or moving it back and forth does no good. If you have a metal frame, another tactic that may work is to lay the bicycle down horizontally inside the loop until the light turns green.

Advancements are under way that may make traffic loops obsolete some day. In particular, radar, infrared and sound detectors have been introduced. Systems based on video cameras are especially promising. Such systems can easily detect bicycles. Such a system may even be able to detect pedestrians some day.

Wednesday, January 16, 2008

Merging into a bike lane is not the same as turning across it

In a recent column in Oregon Cycling, Bikes & The Law: The Right Hook, attorney Ray Thomas states, "the law clearly requires motor vehicles to first yield the right-of-way to bicyclists occupying the bike lane, just as vehicles changing lanes on a multi-lane roadway must first yield the right-of-way to other vehicles occupying the lane the driver would like to enter."

Not quite. Merging into a lane (changing lanes) is not the same as turning across a bike lane. A driver merges into a lane with other drivers going the same direction. A driver turning across a bike lane is crossing the bike lane, not merging into it. As the driver turns across the bike lane, in fact, he is going in a different direction from through bicyclists in the bike lane.

There is no name in traffic engineering for the act of crossing a through lane to make a turn, because such an act violates the principle of positioning your vehicle before you make a turn. Yet that is exactly what proponents of the Portland bike lanes are saying they want to happen. They want through bicyclists to keep to the right of right turning vehicles. I was taught never to pass a right turning car on the right during a group ride when I first started to ride seriously over 30 years ago. A little while later I was involved in the development of the bike lane law in California that was passed in 1976. We specifically designed that law to try to prevent right hook accidents.

On a freeway, slower traffic is supposed to keep to the right. But exits are located on the right side. Does that mean that a fast driver is supposed to take an exit directly from the left lane? No, you can get a ticket for that. Instead, a fast driver is required to merge into the right lane first, then exit, even if he has to slow down to match the speed of traffic in the right lane. That is exactly what the California bike lane law requires of drivers turning right from a street with a bike lane.

From what I can tell, both bicyclists who were killed recently in right hook accidents in Portland had pulled up to a red light next to a stopped truck. They were following what I have been told is the bike lane law in Oregon, which invites such right hook accidents. The Oregon law would be like expecting fast drivers on a freeway to exit directly from the left lane, being careful to yield to drivers in the right lane. Such an expectation is clearly unrealistic, so it is not allowed. Why should we expect a similar maneuver on a street with bike lanes to be reasonable or safe?

Mr Thomas also states, "if bicycles in bike lanes weighed the same as locomotives on railroad tracks (where the legal right-of-way principles are quite similar) there would be fewer motorists cutting us off because the result would be catastrophic for the motorist." When railroad tracks run parallel to a highway, the resulting highway/railroad grade crossings and highway/highway intersections are handled in a special way, usually with traffic signals. See the Manual on Uniform Traffic Control Devices.

Light rail lines that run in highway medians have a similar problem. Light rail trains are heavier than cars, and drivers turn left into the paths of such trains way too frequently. So it is not the weight of the train that makes the difference, it is the driver's expectation that they are not turning across the path of through traffic. The same principle applies to bike lanes to the right of right turning cars. It is the fact that bike lanes to the right of right turning vehicles violate a basic principle of traffic and thus violate a driver's expectation that is the problem, not the failure of such drivers to yield to bicyclists in the bike lane.

Friday, January 4, 2008

Report on my Bicycle Detection presentation to Caltrans HQ Electrical Systems Branch

On October 16, 2007, I gave a presentation, "Detecting Bicycles and Motor Vehicles Using the Same Loop Detector" to 3 members of Caltrans HQ Electrical Systems Branch: Theresa Gabriel, the head of the branch; David Maurin and Bashir Choudry. I recently sent most of you an article with what I intended to present, asking for comments. Thank you to those of you who provided input; it was quite valuable.

After my presentation, Ms. Gabriel responded to my recommendations for changes in the Caltrans Standard Plans, Standard Specifications, Transportation Electrical Equipment Specifications, and the California MUTCD that would address problems that bicyclists are having being detected at actuated traffic signals in the State.

Here are my recommendations followed by Ms. Gabriel's responses:

1. Configure the head loop as a diagonal quadrupole

Caltrans has no interest in investing any additional time or money into improving loop detection. Loops are an obsolete technology that are not reliable for the detection of bicyclists. Furthermore, Construction and Maintenance personnel at Caltrans complain that loops result in premature pavement failure, particularly in asphalt concrete pavement. Therefore Caltrans is looking to non-invasive technologies for detection at actuated traffic signals. Caltrans has contracted with PATH at the University of California, Berkeley, to develop this new non-invasive technology. Any new detection method, however, would need to be compatible with Type 170 controllers (and presumably 332 cabinets) and be able to hold the call as long as the bicyclist was stopped.

Caltrans is interested in bicycle detection and Ms. Gabriel said that she had been actively involved in the development of AB 1581 as well as unsuccessful versions of the bill in earlier legislative sessions.

Caltrans also has problems with insufficient funding to maintain the approximately 5000 traffic signals that it currently has. Caltrans cannot afford to upgrade its traffic signals to meet ADA standards, much less to detect bicyclists. Currently, Caltrans will install one or more Type D bicycle loops at a new or rebuilt traffic signal only if bicycle traffic is anticipated. But even when a Type D loop is installed they still receive complaints from bicyclists. Only bicycle pushbuttons are guaranteed, and they are not practical. Any new detection method will need to guarantee that all bicyclists are detected, including those who are riding non-metal bicycles or from children whose bicycles have small wheels.

2. Locate the loop where bicyclists are expected to stop, but if this is not possible, use a Bicycle Detector Symbol

Ms. Gabriel feared that making a bicycle loop larger in order to locate it where a bicyclist is expected to stop would result in detection of vehicles in the adjacent lane and crosstalk with loops in the adjacent lane. Caltrans is not interested in investing any funds into finding out whether larger bicycle loops have the problems that she fears.

Even with a Type D loop, bicycle detection is not guaranteed to be reliable. Caltrans is unwilling to use Bicycle Detector Symbols unless reliability can be guaranteed, because if a bicyclist stops over a Bicycle Detector Symbol and is not detected and proceeds on a red signal and is injured or killed in an accident, then Caltrans could be held liable.

Besides, the decision to install Bicycle Detector Symbols at actuated traffic signals and whether to add Bicycle Detector Symbols to the drawing in the California MUTCD showing bicycle loop locations lies with the Striping Branch within Traffic Operations, not the Electrical Systems Branch.

3. Use elastomeric loop sealant (or perhaps hot-melt rubberized sealant)

Ms. Gabriel said that the major cause of loop failures was pavement failure, not loop sealant failure as I had stated.

Caltrans has stopped using asphalt emulsion and epoxy loop sealant. It used to supply a 2-part elastomeric loop sealant to construction contractors, but now specifies only hot-melt rubberized asphalt. Although 3M loop sealant is a 1-part sealant and is likely considered elastomeric, Caltrans is unwilling to specify products that are only available from a single source.

Besides, the Transportation Laboratory is in charge of the loop sealant Standard Specifications, not the Electrical Systems Branch.

4. Connect the loops in series

Caltrans uses parallel-series connections not to reduce the inductance at the input into the loop sensor (as I had speculated) but to eliminate the need for a trouble call when a single loop fails out of a group of four. Given the lack of funds, however, Caltrans cannot afford enough traffic signal technicians to perform routine maintenance on detector loops or the equipment inside signal cabinets, so the failed loop is usually found only as a result of a complaint. The only routine maintenance that Caltrans signal technicians perform is to ensure that all signal indications are actually working.

5. ΔL sensor units preferable over ΔL/L sensor units

Such a change would require further development into loop detection, in which Caltrans has no interest.

6. Best to use a separate sensor unit for bicycle loop, but if this is not possible, add only quadrupole loops

Adding quadrupole loops rather than square Type A loops on a loop retrofit would require additional sawcuts, which Caltrans is not willing to do because of potential pavement damage.

7. Set the sensitivity of the sensor unit at the highest setting that will detect bicycles and still reject vehicles in adjacent lanes

Increasing the sensitivity setting increases the risk of crosstalk and of detecting vehicles in the adjacent lane. Using a bicycle or bicycle rim while setting the sensitivity would require closing the lane, which oftentimes is impractical because of heavy high speed traffic.

I plan on helping with the development of new non-invasive detection methods that will reliably detect bicyclists. But I am still of the opinion that loops will be around for a long time, so I also plan on assisting local agencies in improving their detection of bicyclists.

Re: Caltrans response to questions about bicycle detection

A followup to my previous post:

On December 18, 2007, I had a telephone conversation with Jeff McRae, Chief of Caltrans' Office of Intelligent Transportation Systems Projects and Standards. He and a staff member of his wrote the letter that I included in my previous post. Here is the gist of our conversation:

1. I told him that the #1 problem that bicyclists have at traffic actuated intersections is being detected in the first place.

2. He said that Type D loops detect bicycles but that Caltrans is moving away from in-pavement detection because of safety and operational concerns.

3. I asked if he had read my presentation to the Electrical Systems Branch.

4. He said that he had read my presentation some time ago.

5. I asked him if Theresa Gabriel, head of the Electrical Systems Branch, had informed him that the Districts were telling her that bicyclists were complaining of not being detected in lanes equipped with Type D loops.

6. He said no, he was not aware of such complaints.

7. I told him that Type D loops are only 6' wide, the same as the square Type A loops that were introduced in the 1960's, so bicyclists who stop in the right side of the lane are outside of the loop and thus cannot detected.

8. I asked him if he knew George Palm, who I understand invented the quadrupole loop configuration (what Caltrans calls a Type Q loop). I told him that George worked for 3M and Canoga Controls but is now retired and that I had recently spoken with him.

9. Mr. McRae said no, that he was unaware of George Palm.

10. I said that George had taught that Type A loops needed to be located at least 3 feet from the lane line in order to reject vehicles in the adjacent lane. Since lanes are usually 12 feet wide, that led to the Type A loop being 6 feet wide. Caltrans has no records on the development of the Type D loop, so no one knows for sure why the developers of the Type D loop decided on a width of 6 feet. Most likely they selected 6 feet because that matched the width of the existing Type A loop.

11. I said that the big advantage of a quadrupole loop was that its magnetic field drops off much faster outside the loop than a conventional dipole (square, round or diamond) loop. That means that the sensitivity of the vehicle detector (Caltrans calls it a loop sensor) can be turned up to pick up bicycles and high-bodied trucks while still rejecting vehicles in the adjacent lane. The implication is that a quadrupole loop can be built wider than 6 feet and still reject vehicles in the adjacent lane.

12. Mr. McRae said that he had talked with experts on loops and that he understood that Type D loops larger than 6 feet wide would have a lower sensitivity for bicycles and be more subject to picking up vehicles in the adjacent lane.

13. I said that bicycles and motorcycles are like vertical plates of conducting metal and thus are detected if they stop on top of the wire in a loop slot and that they cannot be detected if they stop in the center of a Type A, B or E loop. Type D loops are diagonal quadrupoles and thus can detect a vertical metal plate anywhere within the boundaries of the loop. There are other possible configurations for a diagonal quadrupole than the Type D loop, which I discuss in my presentation.

14. He seemed to have a limited understanding of the physical principles involved in detecting bicycles with inductive loops, so I told him about my background of 3 years of electrical engineering at San Jose State before getting my BS in physics and then my MS and PhD in transportation engineering and subsequently specializing in traffic signals and bicycles. All four areas of my education, training and experience come together in helping me understand detecting bicycles with inductive loops.

15. He expressed concern that bicycles made of non-conducting materials are becoming more common.

16. I said that I am trying to spread the word that bicyclists whose bicycles are made of non-conducting material such as carbon fiber and who use them on the road for training or for transportation need to wrap a few turns of magnet wire around their rim, splice the ends and insulate the splice before installing the rim strip, tube and tire. He did not seem to follow how a bicycle wheel is made, so I asked if he was a bicyclist.

17. He said that he was not a bicyclist and that the information I gave him on how bicycle wheels are made was new. He also was not aware of what magnet wire is.

18. He said that Caltrans has decided to install Type D loops at all new and modified actuated traffic signals.

19. I said that if a bicyclist cannot see the Type D loop because it is covered over with asphalt pavement, then the knowledgeable rider will assume the worst case, which is that the loop is not a Type D loop and that he/she will have to stop the bicycle over the top of the loop slot containing the wire. But if he/she guesses wrong about where the slot is, then the bicycle will be out of the zone of detection and not be detected.

20. I said that one way to address the issue is to increase the zone of detection by making the Type D loop larger, decreasing the dead zone outside the loop. I said that the magnetic field of Type D loops, like Type Q (quadrupole) loops, decreases very quickly outside the loop, allowing the loop to be made larger and still reject vehicles in the adjacent lane.

21. He expressed doubt that a larger Type D loop would detect bicycles and still reject vehicles in the adjacent lane.

22. I said that I had asked for the documentation on the development of the Type D loop and that Caltrans had responded that it could find none. So there is no evidence to support his contention that larger Type D loops would be less sensitive to bicycles and more subject to picking up vehicles in the adjacent lane.

23. He said that although I was recommending larger Type D loops as a way of making the dead zone smaller, no matter how large a Type D loop is, it cannot cover the entire lane because of the necessity to avoid adjacent lane pickup. He then suggested that in my presentation to the CTCDC in Thousand Oaks on January 31, 2008, I focus on markings that tell bicyclists where to stop in order to be detected.

24. I agreed, telling him that one of the recommendations in my presentation was bicycle detector symbols for loops that cannot be seen I asked if he would be attending the CTCDC meeting in January.

25. He said no, he would not be attending the CTCDC meeting, but that one of his staff members would be there.

26. I told him that when I requested District 05, which contains Monterey County, to paint bicycle detector symbols at some nearby state owned and operated actuated traffic signals, their electrical engineer had replied no, out of concerns for safety of their crews and because of the cost.

27. He said that Caltrans is very concerned about both safety of their crews and cost, and that is why they are focusing on video detection.

28. I said that video detection is still rare at Caltrans signals, and that Caltrans owns and maintains thousands of traffic actuated signals with tens of thousands of loops. So any change to video detection is years into the future.

29. I told him that a Caltrans standard specification calls for the final lift of asphalt pavement be placed after the loops are installed. I understand that the reason for the specification is so that loop sealant on the pavement surface has a tendency to fail. I said that in my presentation I recommend the use of better loop sealants than the asphaltic material they are now using.

30. I said that at new or modified actuated traffic signals, the lane is already closed, so the placement of bicycle detector symbols can be done with no increased risk to their crews and little additional cost. If the final lift is to be placed on top of the loop, then the crew simply needs to note the location of the loop so it can paint the bicycle detector symbol after the loop is no longer visible.

31. He agreed that placing bicycle detector symbols during construction is simple. With that, he said he had another matter to attend to and signed off.