Red means stop

I can't believe I need to write this, but I guess that the time had to come. Yes, bicyclists need to stop at stop signs (except in Idaho) and red lights. Yes, they need to stay stopped at traffic signals until the signals turn green or until conditions allow for a legal turn on red.

I write this in response to a comment posted by reader Freedom Bikes:

"I saw this on a bicycling forum. Any truth to this?

  'A lot of forward thinking bike communities (L'ville, KY for instance) totally advocate running reds/stops safely and have quantifiable data as to why it is safer to do so.' "

Nope, that's pure urban legend. One prominent bicycle advocate in Louisville (my friend Jackie Green) urges cyclists to ignore red lights and stop signs "when safe." All of the relevant local government officials, and all of the local bicycle safety instructors, and everyone on the board and staff of Bicycling for Louisville, disagree with him. Jackie sets forth his "as soon as safe" doctrine for leaving intersections, regardless of the presence of stop signs or the phase of traffic signals, here. He justifies it with a list of snippets from news articles about chain reaction car crashes that injured or killed innocent bystanders. Neither Jackie nor anyone else has performed any analysis of the relative safety of running red lights and stop signs "safely" versus obeying them. The anecdotes shared on his website merely show that cyclists and pedestrians sometimes get hurt by motor vehicles struck by other motor vehicles. They do not show any differential in danger between intersection and non-intersection locations or between whether or not the bicyclist or pedestrian was stopped at an intersection when hit.

Kentucky traffic law clearly requires bicyclists to obey stop signs and traffic signals in the same way as motorists must. Given the frequency with which motorists complain to me about bicyclists running stop signs and red lights, it seems to me quite likely that this behavior contributes strongly to the anti-bicyclist sentiment that leads to road rage assaults against bicyclists.

According to the League of American Bicyclists, 8% of car-bike crashes resulting in injuries are caused by the bicyclist running a stop sign or red light. Focusing on getting out of the intersection quickly will inevitably result in bicyclists spending less time evaluating the traffic conditions, more mistakes, and more crashes. At a stop sign or red light, I have much greater concern about getting hit by vehicles who have the legal right to go (that is, the cross traffic) than by the vehicles who have the legal obligation to stop (that is, the ones behind me).

Jackie bases his revisionist view on a Louisville ordinance stating that the traffic law applies to bicycles "... except those provisions of this traffic code which by their very nature can have no application." Even in the unlikely event that a bicyclist could get a judge to believe that the stop sign and red light laws by their very nature have no application to bicyclists, the Kentucky code contains no such provision and the local ordinance cannot supersede the state law. A bicyclist in Kentucky who crashes while running a stop sign or red light has thrown away most of her or his legal rights by having run the stop sign or red light.

If you've read this blog over the past several months, you know that I am no fan of stop signs and traffic signals. I consider other means of traffic control more appropriate in a majority of circumstances. With the well-informed and experienced bicycle advocates of Portland, Oregon seeking an Idaho-style yield-and-roll law for bicycles at stop signs and turning right on red, I am inclined favorably toward that option. Kentucky law clearly prohibits rolling through red lights and stop signs, though, and I believe in the benefits of everybody following the law.

When we make up our own rules, others on the road do not know what to expect of us. This results in confusion, chaos, and destruction - especially for us, the most vulnerable road users. When motorists feel compelled to abide by inconvenient traffic laws and bicyclists ignore those laws, motorists understandably resent our behavior. Angry, resentful motorists are not good for my health as a bicyclist! Even with their flaws, our traffic laws are worth following. We can't expect motorists to obey speed limits when we can't bother to obey stop signs and red lights. To borrow a slogan from San Francisco Bicycle Coalition, we need to "Give Respect to Get Respect." That starts by obeying the laws - as they are now, not as we wish they were. Cleaning up the scofflaw reputation of bicyclists will go a long way to strengthening our hand when we go to the state legislature to reform the traffic laws.

Seeing red, part 4: trying but not succeeding

Louisville Metro government has had bicyclists serving in some key bicycle-related planning and engineering jobs since 2000. They have understood most or all of the problems discussed in the past 3 posts about traffic signals. They have tried at least three solutions thus far. Why, then, do we still have problems with red lights here?

Some of the newer inductive loop detectors at Louisville traffic signals use the quadrupole loops recommended in the website on bicycle-sensitive traffic detectors. You can tell the quadrupole loops by the three parallel black lines in the pavement along the direction of travel (photos of traditional dipole and new quadrupole designs, called "figure 8 loops" on that page). You can see them, for example at the intersections of Pee Wee Reece Road and Taylorsville Road (near Bowman Field - map), E. Liberty Street and Baxter Avenue (map), and Spring Street and Payne Street (map).  Unfortunately, the design details of the quadrupole loops in Louisville result in at least half of them (including two of these three) failing to trigger for bicycles. Some that will detect bicycles (such as the one at Liberty and Baxter) require a bicyclist to place one wheel at exactly the right location, often beyond the stop bar. Merely switching from dipole to quadrupole loops does not solve the problem - the physical and electrical details of the loops need to be right.

A couple of years ago, Metro installed experimental bicycle-sensitive loop detectors on Spring Street on both sides of the intersection with Mellwood Avenue (map). Spring Street, a marked Bicycle Route, has bike lanes on each side. Metro installed the bicycle-sensitive loops in the bike lane, which makes sense only until you recognize that a bicyclist heading northwest (toward downtown) on Spring Street needs to merge left into the main travel lane to keep from getting hit by motorists turning right from Spring Street onto Mellwood Avenue. I don't like a design that forces bicyclists to choose between triggering the light and risking getting right-hooked. In addition, Metro never marked the pavement or in any other way informed bicyclists of the purpose or location of the bicycle sensors. Most local bicyclists I know have no idea the sensors exist until I mention them.

This year, a block away and with equally little fanfare, Metro installed pushbuttons on Spring Street/Adams Street at its intersection with Story Avenue (map). A bicyclist facing a red light on Spring/Adams can pull over to the curb, hit the button, and wait only a few seconds for the light to change. (The buttons work.) Alas, this again requires the bicyclist to ride to the right-hand edge of the street. For a bicyclist riding NW on Spring Street, this causes no problem because traffic on the intersecting street (Story Avenue) goes one-way to the left. For a bicyclist riding SE on Adams Street, though, motorists coming from behind may turn right and hit the bicyclist who has just ridden from the curb after pushing the button. To use the button, the bicyclist must take an unsafe position on the road. As with the bicycle-sensitive loop detectors a block away, no signs or pavement markings alert bicyclists to the existence or function of the button.

In all of these cases, I suspect that it would have cost Metro no additional money to install a bicycle-sensitive loop detector as the traffic detector in the middle of the right lane. It would have cost a bit more to add the necessary bicycle logo (scroll to the bottom of the page) to mark where bicyclists should stop to trigger the light. The problem: some traffic engineers don't believe that bicyclists and motorists can ever be trained to accept the validity of a bicyclist riding in the middle of a vehicular traffic lane, even if only at an intersection. It's time to try a well-designed, adequately funded educational campaign to raise public awareness and change behavior of motorists and bicyclists. Sometimes, education can accomplish goals that engineering cannot. 

Seeing red, part 3: Sensor-actuated signals

Take a look at the road surface when you approach a traffic signal. If you see a tar outline of a rectangle with cut-off corners, the road holds a coil of wire that acts as a metal detector to trigger the traffic signal. If a car stops on top of that metal detector (a.k.a. inductive loop sensor), the car will trigger the light to turn from red to green. But what if a bike stops on top of the sensor? Usually, nothing at all.

When an electrical engineer's explanation of bicycles and sensor-actuated traffic signals first appeared online in 2003, I brought it to the attention of Metro traffic officials. The article explains how to make the loops sensitive enough to detect bicycles without getting triggered by larger vehicles in adjacent lanes. It also tells how to mark streets to show bicyclists the "sweet spot" of the sensor. Several cities have made bicycle-sensitive loop detectors their standard. The Louisville Metro Complete Streets manual approved in 2007 mentions bicycle-sensitive signals once, but does not require or recommend making all sensor-actuated signals sensitive to bicycles.

How much of a problem is this? If you don't care about whether bicyclists stop at red lights, then you probably won't rank it high on your list of necessary improvements to our bicycling environment. I make a point of riding according to traffic law, and these bicycle-ignoring signals drive me crazy. Here's a short top-of-the-head list of signals that I can't trigger, no matter where I place my bicycle on the sensors: Spring Street at Payne Street; Payne Street at Baxter Avenue; Country Lane and Brownboro Road (in front of Doll's Market); North Bellaire Avenue at Frankfort Avenue; and Hillcrest Avenue at Frankfort Avenue. The light at Hillcrest and Frankfort is even more crazy-making, because it flips back from green to red 2 or 3 seconds after a motor vehicle gets off the sensor. A bicyclist waiting behind a motorist at this red light can make it through on green only by tailgating the motorist and sprinting.

I expect that Louisville Metro and the surrounding counties have hundreds of intersections controlled by sensor-actuated signals. Maybe half of these signals will trigger for a bicycle placed at exactly the right spot; the others won't trip for an individual bicycle, ever. These sensors usually get replaced whenever the road is repaved. It costs little to make the new detectors bicycle-sensitive. City, county, and state traffic engineers: Please build signalized intersections that work properly for bicyclists. How can you expect us to obey red lights that don't recognize our existence?

Seeing red, part 2: timed lights

On several one-way streets in downtown Louisville, traffic signals are timed to allow vehicles to proceed at a steady speed through a string of intersections without needing to stop for any red lights. These synchronized traffic signals work well as long as the vehicles can maintain that set speed within a few mph.

In my experience, I can ride through 6 or 7 green lights in a row on some of these streets if I ride fast: at least 23 mph. On my commuting bicycle, I simply can't ride that fast. Yesterday, I did some measurements and calculations to learn how fast a bicyclist needs to ride to keep up with the traffic signals, and if any slower-than-car speed could allow a bicyclist to cruise through all green lights.

For the test case I used East Main Street, a common commuting route with a bike lane. The signals are timed to allow a driver going at 34 mph to go from a green light at one intersection to green lights at all of the following intersections. Riding at 24 mph, one would fall behind the signals by about 5 seconds each block plus another 5-10 seconds delay if you need to start from a full stop or a low speed. This fast rider would get through about 7 intersections before getting stopped by a red light. At 20 mph, a rider would get through about 5 intersections before encountering a red light; at 15 mph, 3 intersections; at 12 mph, only 2 intersections before needing to stop again. How slowly would you need to ride to encounter all green lights? You would need to poke along at less than 10 mph.

Someone riding from the east end of Market Street to 3rd Street, for example, would have red lights add 1 minute to the trip if she or he rode at 20 mph. Red lights would add 2-1/2 minutes to the travel time for a 12-mph bicyclist. That comes to a 30% time penalty for the 20-mph rider, and a 40% time penalty for the 12-mph rider.

If Metro re-synchronized the signals for a 25 mph speed limit (instead of 35 mph), the 20 mph bicyclist could get to 3rd Street or beyond without stopping, and the 12 mph bicyclist would arrive one minute sooner. But what about the motorists who could drive at only 24 (instead of 34) mph? The travel time difference for this 10-block trip would be 51 seconds.

Bottom line: If Louisville (or another city with synchronized downtown traffic signals) wants to make its downtown safer and more accessible to bicyclists and pedestrians, they need only re-time the traffic signals for a lower speed. This will decrease the number and severity of car crashes (including crashes with bicyclists and pedestrians), make it much easier for pedestrians to cross the street, and allow bicyclists much less delay due to red lights. The cost - a minor inconvenience to motorists.

Seeing red

Most major streets in greater Louisville meet at intersections controlled by traffic signals  ("signalized intersections" in traffic engineering lingo). In spite of the number and severity of crashes at signalized intersections, the inconvenience of stopping at red lights regardless of traffic volumes, and the high cost of building and maintaining these intersections, I took them for granted until recently. It seemed as though they must improve traffic flow and safety compared to other options (stop signs, for example) or else the traffic engineers would stop using them. I certainly don't want a more chaotic traffic environment than we have now.

As regular readers know, I prefer modern roundabouts (not traffic circles) to signalized intersections in many situations. I believe that one-lane roundabouts could replace many of the traffic signals in our area with benefits for motorists, bicyclists, and pedestrians. Of course, a transition from signalized intersections to roundabouts would cost money and involve controversy. Even in the best-case scenario, we will have lots of traffic signals for years to come.

As long as we use traffic signals, we need to make them functional for all legal road users. In Louisville, we have two basic types of traffic signals: signals timed to turn from red to green on a fixed schedule, which might differ for different times of day (for example, morning rush hour versus evening rush hour); and signals that stay green for traffic along the more major street until a sensor detects vehicles waiting on the smaller street. I'll call the first type "timed signals" and the second type "sensor signals." Locally, very few signals of either type work well for bicyclists. Tomorrow, I'll explain the problems and explore solutions.