The City of San Diego, California, USA published its Traffic Signal Bicycle Detection Study Final Report in 1986. I have scanned a copy of the document and posted it online.
Cover of the San Diego Traffic Signal Bicycle Detection Study Final Report
This report’s main topic is inductive-loop traffic-signal detectors. These are metal detectors with their antennas, loops of wire, spread out just below the surface of streets. The detectors serve to trigger traffic signals when vehicles are overhead.
Detection increases efficiency of traffic flow. A small street that crosses a main street can, for example, get a green light only when a vehicle is waiting. A timer does not have to stop traffic on the main street every minute or two, whether or not anyone is waiting in the side street.
An earlier type of vehicle detector was a pressure-sensitive device embedded in the street. From approximately 1970 onward, though, loop detectors came increasingly into use, having become practical with the advent of transistorized electronics. The traffic-engineering literature mentions them as early as 1966.
Pressure-sensitive detectors responded to bicycles, but many loop detectors were not sensitive enough to detect bicycles, as reported as early as 1975 by bicycling author John Forester. By 1985, several California communities had addressed this issue in one way or another. The report describes these efforts, and offers suggestions both for new installations and for retrofits.
The report includes several pages of research data establishing the sensitivity necessary to detect bicycles. Detection was easily achieved with the correct setting of the electronics in the signal control box, but raising the sensitivity of the detector would result in unwanted triggering by a vehicle in the next lane. Commonly, a vehicle exiting an intersection could trigger a signal. Installation crews addressed this problem by turning the sensitivity down as low as possible while still detecting a passenger car in the lane directly over the actuator loop.
The detector, so adjusted, will not detect a bicycle and often not a motorcycle. This deficiency encourages — in some cases forces — bicyclists and motorcyclists to run red lights. Most bicyclists, probably also most motorcyclists, do not understand why the lights stay red, fostering disrespect for traffic signals. Entering an intersection on the red can, as is well-known, result in crashes.
Making the loop detector antenna more directional — so it responds much more strongly to vehicles in the lane directly overhead than to ones in adjacent lanes — allows the sensitivity to be raised so that bicycles and motorcycles are detected reliably. The San Diego report shows how to accomplish this.
The San Diego report describes a workable, very low-cost solution: all that is needed is to lay the wire in a different pattern, adjust the sensitivity of the electronics, and in some cases, paint a marking showing where bicyclists need to wait. There was nothing new to on its topic to say for decades, until evolving technology brought detection using video cameras, light-emitting diodes, microwaves and ultrasonic sound emitters into consideration.
Nonetheless, implementation of reliable detection can be described as, at best, spotty, even 30 years after publication of the report. I address that history in another post.
References:
MGA Associates, City of San Diego, Traffic Signal Detection Study Final Report, ca. 1986. (The document does not list a publication date, but the latest dates mentioned in the document are in 1985.)
Forester, John, Effective Cycling, Second Edition, 1975, page 3.4-2
Tranoff, P. J. and P.S. Parsonson, Selecting Traffic Signal Control at Individual Intersections, online at https://trid.trb.org/view/172651 See for example references 18, 30, 31, 49, 52, 58, 63 (1974, which also mentions treadle detectors), 77, 114, 115 (the earliest reference to loop detectors, 1966).
The earliest discussion of induction loop insensitivity I have discovered is in the 1974 report “Bikeways: State of the Art.” The report notes that equipment is available to overcome the problem, but provides no specifics. The report focuses on European design treatments which are intended to actuate accurately when a bicycle is in a bicycle lane adjacent to an actuated traffic lane. The follow-up 1977 “Locational and Facilities Criteria, Part II: Design Treatments” does not follow this up, instead recommending separate pedestrian push-button type actuators for cyclists, again in the bike lane.
The 1975 edition of Effective Cycling also discusses the problem, and identifies it as a problem of loop design, but offers no solution, instead recommending that cyclists simply treat such signals as stop signs. I do not have a copy of the initial edition of Bicycle Transportation at hand so cannot comment on this.
Three or four TRB papers from the 1976-78 era continue the focus of Locational and Facilities Criteria in their emphasis on pedestrian-style push buttons. That treatment was eventually rejected due to a 20 percent right-turning car refusing to yield to the bicycle when the bicycle was stopped at the right road edge up at the stop line, the stop position mandated by the separate push-button type head.
The 1983 edition of Bicycle Transportation is the first extensive discussion I have found of induction loops in regards to bicycles that recommends both more attention to gating (gain) and the now familiar figure-8 loop pattern.
Bruce — Thanks. I have the 1977 edition of Cycling Transportation Engineering, comb-bound version published by Forester’s Custom Cycle Fitments. Section 3.2.11 covers presence detectors and recommends the quadrupole loop. It does not mention the D-type loop. I also have a copy of Bicycle Transportation, 1994 edition. I’ll crack it open later — I have time only for this brief message now.
Forester’s book Bicycle Transportation (MIT Press, 1994) mentions that there are other patterns which may work better than the quadrapole loop, but without describing them in detail. There is also a more extensive, updated discussion of the politics of detector installation.
Pingback: Traffic-signal actuators — some documentation | John S. Allen's Bicycle Blog
Pingback: Non-responsive signal actuators in Waltham | Street Smarts