Interpretation ID: nht92-4.41
DATE: August 11, 1992
FROM: Steven Henderson -- McGill University
TO: Paul Jackson Rice -- Chief Counsel, NHTSA
TITLE: None
ATTACHMT: Attached to letter dated 8/28/92 from Paul Jackson Rice to Steven Henderson (A39; Std. 108)
TEXT:
Thank you for your letter of June 29 concerning our motorcycle headlamp warning device. I certainly agree with you that the device contravenes the letter of DOT Standard No. 108 as it presently stands. Nevertheless, I ask for your patient indulgence in considering the counterarguments I present below.
I will argue that the potential benefits of our device are so congruent with both the congressional mandate of the NHTSA ("to reduce the mounting number of deaths, injuries, and economic losses ... on the Nation's highways") and the intent of the author(s) of Standard No. 108, as to warrant the inclusion in the standard of an exception referring specifically to the operating characteristics of the device. My understanding is that an exception may be incorporated into the standard when in specific situations the public safety would be increased by a rigorously specified departure from die standard. That is, the exception allows the spirit of the law to rule while protecting the letter of the law from violation. Nevertheless, an exception governing a device should only be granted if its use would not cause any difficulties that the original standard was framed to prevent. I will address these issues below. Your further point that "the safety benefits of the device are speculative and unproven" is very relevant and well taken, and again I ask for your tolerant consideration of my discussion on that point.
Briefly, my understanding of the issues you raise regarding the device's noncompliance with Standard No. 108 are:
a) the headlamp modulation is greater than 17% and so the device does not comply with S5.6.1(c). (Please note that modulation is between intensity levels within a single beam, and is available for use with either the high or the low beam.)
b) the headlamp flash rate of 10 per second supersedes the maximum allowable rate (S5.6.1(a)) of 280 per minute, or 4.67 hz.
c) the taillamp(s) would no longer be steady-burning, as required by S5.5.10(d).
d) the turn signal flash rate of 10 hz would not comply with the SAE specification of 60-120 flashes per minute (1-2 hz) referred to in Standard No.108.
Within my discussion of each of the four specific instances of noncompliance I will also attempt to infer the intent of the clauses cited. If the device violates the letter of the law while satisfying the spirit or inferred intent of the law in each case, I feel that the granting of an exception should be
considered by the NHTSA.
First, S5.6.1(c) calls for a modulation intensity of not greater than 17%, the presumed intent being to reduce distraction and annoyance to other drivers whenever such a modulating headlight system is in view. For a device that is used steadily (as is the device described in S5.6), when the presence of a motorcycle is recognized, the modulation becomes redundant and so carries no useful information. Without a limit on its magnitude, the modulation would likely interfere with the other visual processes needed for operating a motor vehicle. On the other hand, our device is only used intermittently and only at need, so that when the presence of the motorcycle is recognized by the car driver, the horn button is released and the modulation stops. As the device is only used to attract attention when necessary and at no other time, no redundant information from the device is ever present in the visual scene to distract other drivers. Therefore our proposed device does not conflict with the inferred intent of S5.6.1(c). Furthermore, to maximize the likelihood that the motorcycle will be seen and an accident avoided, the exception for the device should permit 100% modulation on the condition that modulation can only occur in conjunction with the sounding of the horn.
Second, S5.6.1(a) specifies a modulation rate of between 200 and 280 cycles per minute, or 4.67 hz, while our device has a modulation rate of 600 cycles per minute, or 10 hz. Despite the salience of the 10 hz flash rate, the likelihood of distraction to other drivers is nonexistent, as again the device is only used at need rather than continuously (see the counterargument given above). The 10 hz flash rate is particularly appropriate for use as a warning signal. Figures 1 and 2 (from Snowden and Hess, 1992) show that at every retinal eccentricity, light modulation of approximately 10 hz is more easily seen (or seen at a greater distance) than modulation of any other frequency. Finally, referring to Appendix A, the "brightness enhancement effect" (Bartley, 1939) also achieves maximum enhancement at a rate of about 10 flashes per second, the frequency corresponding to the alpha rhythm of the human visual cortex. As the 10 hz enhancement effect was reported in a document commissioned by the NHTSA and published as Report No. DOT HS 807 121 (1987), I believe that the intent of the author(s) of Standard No.108 (the NHTSA) would be to exploit the safety advantages to be gained by the use of 10 hz light modulation, given the knowledge gained from the visual system research that has been conducted since Standard No.108 was written, and given the safe means to use this knowledge. (Please also note that the 50% duty cycle of the proposed device yields maximum brightness enhancement, as shown by the figure in Appendix A.) I therefore contend that as the flash rate of 10 hz is not contrary to the inferred intent of S5.6.1(a), and further, that its use will not cause any of the difficulties that S5.6.1(a) was likely Written to prevent, that the requested exception should permit the use of a 10 hz flasher frequency, again on the condition that modulation can occur only in conjunction with the sounding of the horn.
Please observe that the first two points of noncompliance result from an attempt to apply to our device, an exception granted to allow the use of another device that is explicitly described in the standard. Our device is unique and clearly differs from the device referred to in S5.5.10(c) and described in S5.6. (In our attorney's opinion the patent search we commissioned found no prior art of sufficient similarity to jeopardize our
patent claims, as shown in the prior art section of the enclosed patent application - see Appendix B.) Therefore, the noncompliance of our device with an exception drafted in specific reference to a different device should not be taken as grounds for prohibiting our device. Rather, I contend that our device is not contrary to the spirit and intent of Standard No.108, and that introducing an exception for the device into the standard would be the most appropriate course of action for the NHTSA to take. Section 5.6 establishes a clear precedent for introducing an exception into the standard by the NHTSA, if public safety would thereby be expected to increase.
Third, when the horn button is pressed the taillamp flashes at a rate of 10 hz, rather than being steady-burning as required by S5.5.10(d). However, a taillight's purpose is to mark the rear of a motor vehicle during nighttime driving when it would otherwise be invisible. For this reason the law requires that taillights be lit at night. The law makes no such requirement during the day. The law does require that motorcycle headlights be lit during the day. Consequently, the taillight, being wired in parallel with the headlight, is also always lit during daytime riding, although not required to be by law. At night the taillight will always be steady- burning, as required by S5.5.10(d), because the flasher device is only able to induce taillight flicker during daylight hours due to the photocell circuitry incorporated to prevent the headlight from generating strobe effects at night. Therefore, the device is in compliance with S5.5.10(d), as it will cause the taillight to flash only at times that it is not required by law to be lit. At such times as the taillight must be lit to mark the rear of the motorcycle to following drivers, it can not be made to flicker by pushing the horn button, and so the device does not violate the intent of the standard.
Fourth, if the horn button is pushed while a turn is being signalled, the appropriate signallight will give out 2 hz bursts of 10 hz flicker (the other signallight remaining dark) and so will not comply with the flash rate of 1-2 hz "specified by SAE requirements incorporated by reference in Standard No. 108". Note first that the situation in which the turn signal and horn button are simultaneously in use will be relatively rare. In addition, the salient difference between a 10 hz flash rate and a 2 hz flash rate allows both to be seen simultaneously and with little or no interference between the two perceptual channels. In other words, an observer who sees a turn signal if a uniform fight is flashing at 2 hz will also see a turn signal if a 10 hz flickering fight is flashing at 2 hz, due to the independence of the temporal channels of the visual system. This channel independence has been most clearly shown by Hess and Snowden (1992) who state that:
The results for 0 and .3 c/d (Fig.4) (shown here as Fig.3) suggest that probes of 1 Hz are detected by a temporal mechanism with a low pass temporal characteristic while probes of 8 Hz are detected by a band pass temporal mechanism centered at 8-10 Hz...For stimuli of mid spatial frequency (1 and 3 c/deg) the results of Fig.5 (shown here as Fig.4) suggest the presence of at least 2 temporal mechanisms, one low pass underlying the detection of low temporal frequency probes and one band pass centered at 8 Hz and underlying the detection of higher temporal frequency probes (p.50).
I have also included their Figures 7 and 8 (shown here as Fig.5 and 6) from
the same publication to illustrate that the temporal frequency of 2.4 hz (above the highest signal light rate specified by SAE) is also detected by the low pass channel tuned to 1-2 hz frequencies, and not by the band pass channel tuned to 8-10 hz frequencies. This result demonstrates empirically that no perceptual interference will occur between the signal flash of 1-2 hz and the hazard flash of 10 hz. Thus, if the signal switch and the horn switch are used together, the SAE-specified turn signal flash of 1-2 hz will be perceptually present, the hazard signal flash of 10 hz will also be perceptually present at the same location, and the two signals will not interfere. Therefore, as widely separated temporal channels in vision are independent, our device is in compliance with the SAE specification of a 1-2 hz flash rate for turn signals, and the additional presence of a 10 hz flash rate for a hazard signal at the same location does not introduce a point of noncompliance between our device and Standard No. 108.
For the reasons given above, I feel that the action by which the NHTSA could most appropriately carry out its congressional mandate would be to write an exception into Standard No. 108, with the exception clause containing the specific assertion that the proposed device is legally permitted if the device can only be actuated by the horn button. The temporary granting of the exception (for an interval of one or two years) will allow the "speculative and unproven ... safety benefits" of the device to be evaluated. If such benefits exist, and the device could help save hundreds or even thousands of lives every year, testing must be carried out.
Before discussing two studies by which the device's benefits can be assessed, I wish to offer from personal experience some anecdotal evidence that demonstrates the need for this device. In 1971, while riding my Triumph motorcycle on a mountain highway in British Columbia, I had a head-on crash with a car at a closing speed of about 60 miles per hour. The car's driver had gradually crossed into my lane as he prepared to turn left onto a gravel road. I moved to the center line in an attempt to squeeze past his car, whereupon he saw a car about fifty yards behind me and attempted to return to his own lane. I was thrown over his car and trailer and sustained a ruptured spleen, some broken bones, and a concussion. The other driver subsequently testified in court that he never saw me, although the accident took place on a clear sunny summer afternoon. That accident would almost certainly not have occurred if I had had the use of the device I have described to you. Furthermore, the responses of other motorcyclists to whom I have described the device are uniformly enthusiastic. We have all felt the helplessness of seeing a motorist inadvertently threatening our lives because he or she has not seen us and we have no means to make ourselves more visible. This device would give motorcyclists the means to do so. If the NHTSA grants a temporary exception to allow safety tests, I believe that many motorcycle manufacturers and insurance companies would enthusiastically support and participate in the studies outlined below. The purpose of the first study is to generate videotape records of automobile drivers' responses to the device's warning flashes. Several motorcycles will be equipped with warning flashers and special purpose cameras. The cameras will be similar in principle to those used by business security firms, in that they will constantly record onto a thirty second videotape loop.
Fifteen seconds after a horn button is pressed, the entire loop's contents will be copied to a permanent videotape record. This record will show, for
each instance of flasher use, the circumstances making such use necessary, as well as the car driver's reaction to the warning flashes. A telltale light or the flasher itself will be within the camera's field of view, furnishing a flasher activation record against which the car driver's preflash and postflash behaviors can be categorized as either appropriate or inappropriate. Each record of a change from inappropriate to appropriate behavior that coincides with flasher activation will be prima facia evidence of the safety benefits conferred by the device. Although this study should generate compelling visual evidence for any safety benefits that exist, the study will not generate logically compelling statistical evidence for the effect, as the criticism can logically be made that the car driver could have seen the motorcycle and responded appropriately in any case (such as by coming to a halt after beginning to pull out from a side street, or by pulling back into the correct lane rather than attempting to complete an overtaking maneuvre), even without flasher activation. Although the cost of conducting the first study is not prohibitive, to carry out the study will certainly require funding support. Clearly however, no group or company will fund the study without assurances from the NHTSA that the device may be legally used if the research demonstrates that the device reduces accidents and saves lives.
The second study is an example of the statistically rigorous experimental design needed to incontrovertibly establish that the device does confer a safety benefit as well as to allow an estimation of the benefit's magnitude. This study requires that a motorcycle manufacturer install flasher systems onto several hundred of their motorcycles. As the motorcycle accident rate in North America in 1987 was approximately 1 accident for each 13 registered motorcycles (U.S. Census figures), a large number of participants is required to allow a statistically significant measure of the device's safety benefits to be made. For example, a safety benefit having sufficient strength to reduce the accident rate by 25 percent (a reduction of 1000 fatalities per year) requires 800 motorcyclists in each of two groups (only the experimental group being equipped with the flashers) to show a statistically significant effect at p less than .05, and 1380 motorcyclists in each group to show statistical significance at p less than .01. Again, no motorcycle manufacturer will conduct so costly a test - even with the support of motorcycle insurance companies - unless the NHTSA has signalled a willingness to allow the device to become available if a safety benefit can be shown.
My most fervent wish is that in light of the reasoning offered above regarding our flasher system, you will reconsider your initial position as stated in your letter of June 29, 1992, and introduce into Standard No. 10 an exception allowing the use of the device I have described. In addition, I again ask (further to my letter faxes of June 3 and June 9) that you request an advanced examination of the patent application filed with the U.S. Patent Office on April 1, 1992 by Steven Henderson and David Kernaghan under Serial No. 863686.
Thank you for your interest and your patience. I look forward to your reply.