August 2000
The National Highway Traffic Safety Administration's Office of Plans and Policy is planning to obtain crash data and statistically evaluate the effectiveness of antilock brake systems (ABS) and rear impact guards for heavy trucks. The agency welcomes your review and suggestions on the study.
Data collection NHTSA proposes to work with the State police from at least two large States. They will send data to NHTSA on every crash they investigate that involves a tractor-trailer, a bobtail tractor, or a medium or heavy single-unit truck. The data will include the basic State crash report plus a supplemental form identifying if the truck or trailer are ABS-equipped (as evidenced by presence of the malfunction indicator lights). The data will comprise approximately 10,000 tractor-trailer crashes and 5,000 single-unit trucks. On the subset of approximately 1,000 truck-trailers and 700 single-unit trucks that were hit in the rear by the front of a passenger vehicle, police will fill out a second supplemental form describing the rear impact guard on the trailer and the damage pattern on the passenger vehicle. Data collection will start in January 2001, or as soon as feasible after that, and run for two years. NHTSA believes these samples will be adequate for statistically evaluating ABS and rear impact guards.
Analysis of antilock brake systems The purpose of ABS is to help maintain directional stability and control during braking, and possibly reduce stopping distances on some road surfaces, especially on wet roads. ABS could reduce crashes involving jackknife, loss-of-control, run-off-road, lane departure, or skidding, or where trucks with conventional brakes were unable to stop in time to avoid hitting something frontally. On the other hand, ABS is unlikely to affect a control group of crashes where the truck was standing still, moving too slowly for ABS activation, or proceeding straight ahead when another vehicle unexpectedly hit it in the side or rear. The ratios of the various crash types where ABS has potential benefits to control group crashes will be compared for tractor-trailers where both units are equipped with ABS versus tractor-trailers where neither unit is equipped; also for ABS-equipped single-unit trucks vs. non-equipped trucks.
Analysis of rear impact guards The goal of a rear impact guard is to arrest the forward motion of the striking passenger vehicle and prevent a damage pattern called "underride with passenger compartment intrusion (PCI)" that is dangerous for occupants of the passenger vehicle. The proportion of rear impacts that result in underride with PCI will be compared for trailers with guards that meet NHTSA and/or industry standards versus older trailers with guards that do not meet NHTSA or industry standards. Since the NHTSA standard does not apply to single-unit trucks, the analysis for these trucks will be limited to estimating the overall incidence rate of underride with PCI in rear-impact crashes.
For your comments or to obtain additional information The two attached documents describe each evaluation in more detail. You may provide your comments to John Jacobus (202-366-2586 or jjacobus@nhtsa.dot.gov) or Charles Kahane (202-366-2560 or ckahane@nhtsa.dot.gov), Plans and Policy, NHTSA or ask them for additional information on either evaluation.
Antilock Brake Systems (ABS) are required on air-brake equipped truck-tractors manufactured on or after March 1, 1997 and on semi-trailers and single-unit trucks equipped with air brakes and manufactured on or after March 1, 1998(1). ABS is required on hydraulically braked single-unit trucks over 10,000 pounds Gross Vehicle Weight if they are manufactured on or after March 1, 1999(2). ABS was also available as an option for some of these vehicles for several years before those dates.
The presence or absence of ABS on a vehicle can be identified by:
The purpose of ABS is to help maintain directional stability and control during braking, and possibly reducing stopping distances on some road surfaces, especially on wet roads. ABS is potentially effective in any crash where the driver braked hard enough to activate the system, and where conventional brakes would have contributed to directional instability and/or excessive stopping distance. The types of crashes where reductions may be anticipated with ABS are those involving jackknife, loss-of-control, run-off-road, lane departure, or skidding - to the extent that these phenomena may have been caused by brake-related directional instability - and those where trucks with conventional brakes were unable to stop in time to avoid hitting something frontally. On the other hand, ABS is unlikely to prevent crashes where the truck was standing still, moving too slowly for ABS activation, or proceeding straight ahead when another vehicle unexpectedly hit it in the side or rear.
NHTSA's Regulatory Impact Analysis cites a study by D. Otte et al predicting that ABS could reduce overall crash rates of tractor-trailers by 10 percent and single-unit trucks by 6 percent(3). Since approximately of crashes involve trucks that were standing still, moving slowly, or unexpectedly struck while moving straight ahead, ABS would have to prevent 15 percent of the remaining of tractor-trailer crashes and 9 percent of the remaining single-unit truck crashes to reach the overall benefits of 10 and 6 percent, respectively. In other words, any evaluation of ABS ought to include enough crash data on tractor-trailers that an observed 15 percent crash reduction would be statistically significant (and enough data on single-unit trucks that a 9 percent crash reduction would be significant)(4).
Tractor-trailer combinations can belong to four groups with respect to ABS: (1) tractor and trailer both ABS-equipped; (2) ABS tractor, non-ABS trailer; (3) non-ABS tractor, ABS trailer; (4) neither equipped. Crash data will need to identify the ABS status (i.e., "present" or "absent") of the trailer as well as the tractor. Crash statistics need to be computed separately for each group, and effectiveness of configurations (1), (2) and (3) must each be calculated relative to (4).
Analysis plan NHTSA's past evaluations of ABS for light trucks(5) and passenger cars(6) employ a statistical method that will also work here. Crash data are never detailed enough to identify exactly which crashes involved braking problems. But they allow crashes to be subdivided into generic groups that sometimes or often involve brake-induced loss-of-control and/or excessive stopping distance, and could benefit from ABS. They also permit creation of a control group(s) that is unlikely to be affected by ABS. Ratios of crash involvements in the various potential-benefit groups relative to the control group will be compared for ABS and non-ABS vehicles, as in the following hypothetical example that shows a statistically significant reduction of jackknife involvements when the tractor and trailer are both equipped with ABS:
|
Tractor: |
No ABS |
ABS |
||
|
Trailer |
No ABS |
ABS |
Reduction with ABS |
Chi-Square |
|
Jackknife Crashes |
100 |
40 |
||
|
Control Group Crashes |
500 |
400 |
||
|
Ratio |
.2 |
.1 |
50% |
12.51 |
The generic potential-benefit groups of crashes could include:
The control group could include crashes where the truck was parked, stopped prior to the crash, driverless, backing up, moving less than 10 mph, struck in the rear while moving straight ahead, or, perhaps, struck in the side while moving straight ahead. These are crashes where brakes were evidently not applied, or where the vehicle was moving too slowly for brake application to activate the ABS.
The data will generate a matrix of crash involvements by type of crash and the ABS status of the tractor-trailer (four groups as defined above). A similar matrix will be generated for single-unit trucks, but here there are only two groups of ABS status (equipped, not equipped).
Separate analyses will also be performed for crashes on wet roads and, if possible, for lightly loaded trucks. These are two situations where ABS is believed to have high potential effectiveness.
Data plan NHTSA proposes to obtain the crash data from State police in two large States over a two-year period (2001-2002). The State police would furnish the basic crash report, plus a supplementary crash report developed by NHTSA. They would fill out the forms for every crash involving a tractor-trailer, a bobtail tractor, or a medium or heavy single-unit truck. NHTSA anticipates a total sample of 10,000 tractor-trailer cases and 5,000 single-unit truck cases for the two States over two years. (Double and triple trailer combinations would be excluded from the study because sample sizes would be insufficient for analyses.)
NHTSA recently concluded a similar data collection to evaluate the effectiveness of conspicuity tape for semi-trailers. Florida and Pennsylvania filled out a supplemental form on every trailer, specifying if it was equipped with the tape and describing the condition of the tape. The data collection was highly successful and, in combination with the basic crash reports, the data were just right for the evaluation.
The basic crash report will describe the crash configuration, the movement of each vehicle prior to the crash, and the point of impact on each vehicle. In many cases that will be enough information to assign the trucks to crash groups, as described above. The basic report will indicate the model year and VIN of the truck-tractor or single-unit truck. That is often enough to reveal if the truck was ABS-equipped.
The supplemental form will enable police to record information about trailers needed to determine if they have ABS, including the presence of an exterior Malfunction Indicator Light (MIL), the model year, the build date, and the VIN. For truck-tractors/single-unit trucks, the presence of the dashboard MIL and the build date will also be recorded, to supplement the VIN on the basic report.
The supplemental form will also include a few questions about the crash configuration where the basic report does not provide all the data needed to group crashes. For example, in a State where the basic form does not distinguish between trucks that were "stopped" or "slowing" before the crash, this would be asked on the supplemental form. Police would specify whether a truck departed from its travel lane prior to the impact. Where possible, police would ascertain whether trucks or trailers were lightly or heavily loaded.
The field-collected data should be more than adequate, in most cases, to identify whether trucks or trailers are equipped with ABS. For the limited number of cases where that is impossible (pre-standard models where ABS was an option), NHTSA hopes manufacturers would be willing to analyze these VINs or consult their build sheets to determine if the vehicle had ABS.
Statistical adequacy of this data plan Based on vehicle-age distributions in 1997-98 Florida and Pennsylvania crash data, it is anticipated that a sample of 10,000 tractor-trailer cases collected in 2001-2002 would likely include approximately 3,400 where neither the tractor nor the trailer were equipped with ABS, and 3,000 where both were equipped(7). Based on the distribution of crash groups in those 1997-98 data (as best as can be inferred from the basic police report) and the analysis plan illustrated in the hypothetical example above, an observed 25 percent reduction of jackknife incidents, relative to the control group, would produce a chi-square that is statistically significant at the two-sided .05 level. A 33 percent reduction of jackknife incidents would be significant at the two-sided .01 level.
Similarly, a 15 percent reduction of rollovers and other run-off-road crashes would be significant at the .05 level, a 20 percent reduction at the .01 level. For rear-end-striking impacts, the critical reductions are 16 and 21 percent, respectively. For all types of collisions with passenger vehicles in which the tractor-trailer was primarily responsible for the crash, the critical reductions are 13 and 18 percent respectively.
Since NHTSA anticipates ABS to be especially effective in reducing jackknife incidents, and, based on the Otte study, about 15 percent effective in reducing other types of non-control-group involvements by tractor-trailers, the proposed sample of 10,000 crashes ought to be quite adequate, yet not excessive, for evaluating the effectiveness of ABS on the tractor and the trailer.
On the other hand, the proposed data plan will only generate a sample of about 5,000 single-unit trucks. Only about 1,500 of them are likely to be equipped with ABS. NHTSA projects that an 19 percent reduction of rollovers and other run-off-road crashes would be significant at the .05 level, a 24 percent reduction at the .01 level. For rear-end-striking impacts, the critical reductions are 20 and 26 percent. For all types of collisions with passenger vehicles in which the truck was primarily responsible, the critical reductions are 16 and 21 percent.
Since the Otte study only predicted a 9 percent reduction of non-control-group crashes of single-unit trucks with ABS, it is likely that the proposed sample will not produce statistically significant results. That raises a question whether single-unit trucks should be included in the study. Their inclusion is justified because it would involve little additional cost, and the data might conceivably reveal some important, unanticipated effect for ABS.
Inadequacy of alternative data plans The National Automotive Sampling System - Crashworthiness Data System (NASS-CDS) does not currently include heavy trucks as "case" vehicles. It picks up records of about 150 crashes per year involving tractor-trailers and 50 involving single-unit trucks that collide with light passenger vehicles. Even if NASS-CDS included a representative sample of heavy-truck crashes, it would not produce the sample of 10,000 crashes needed for a statistically meaningful evaluation of the effect of ABS.
It is essential for the evaluation data to identify the ABS status of both the tractor and the trailer. State crash files do not record the VIN or the model year of trailers. That makes it impossible to determine if trailers were ABS-equipped. (Exception: Florida encodes the VIN and model year for fewer than half the trailers involved in crashes, about 2000 per year. After five years there might be enough Florida data for a statistically meaningful analysis.)
Similarly, national files essentially compiled from State crash data, such as the Fatality Analysis Reporting System (FARS), the Motor Carrier Management Information System (MCMIS) or the NASS General Estimates System do not have trailer VINs or other information needed to ascertain the ABS status of trailers.
On the other hand, selected State files and FARS may prove satisfactory for the evaluation of ABS in single-unit trucks, since trailer information is not an issue. ABS status of the truck can be ascertained from VINs, with assistance from manufacturers on certain pre-standard VINs. However, this alternative (or supplementary) approach may entail some weaknesses or State-to-State inconsistencies in the identification of "potential-benefit" and "control-group" crashes, and it is likely to require a more extensive VIN-analysis effort by the manufacturers.
Footnotes for Evaluation Of Antilock Brake Systems For Heavy Trucks
1. Code of Federal Regulations, Title 49, General Printing Office, Washington, 1999, Part 571.121.
2. Code of Federal Regulations, Title 49, General Printing Office, Washington, 1998, Part 571.105,Nt.
3. Preliminary Regulatory Impact Analysis - Notice of Proposed Rulemaking - FMVSS Nos. 105 & 121 - Stability and Control While Braking - Requirements for Medium and Heavy Vehicles, Submission to NHTSA Docket No. 92-29-N02-001, Washington, 1993, p. 38.
4. Also see Final Economic Assessment - Final Rules - FMVSS Nos. 105 & 121 - Stability and Control During Braking Requirements and Reinstatement of Stopping Distance Requirements for Medium and Heavy Vehicles, February 1995, Submission to NHTSA Docket No. 92-29-N05-002.
5. Kahane, C.J., Preliminary Evaluation of the Effectiveness of Rear-Wheel Antilock Brake Systems for Light Trucks, Submission to NHTSA Docket No. 70-27-GR-026, Washington, 1993.
6. Kahane, C.J., Preliminary Evaluation of the Effectiveness of Antilock Brake Systems for Passenger Cars, NHTSA Technical Report No. DOT HS 808 206, Washington, 1994.
7. There would also be approximately 2,600 cases with ABS on the tractor but not the
trailer, and 1,000 cases with ABS on the trailer but not the tractor.
Federal Motor Vehicle Safety Standards 223 and 224 set minimum requirements for the geometry, configuration, strength and energy absorption capability of rear impact guards on full trailers and semi-trailers over 10,000 pounds Gross Vehicle Weight manufactured on or after January 26, 1998(1). However, the Truck Trailer Manufacturers Association (TTMA) had already issued a voluntary Recommended Practice RP 92-94, Rear Impact Guard and Protection in April 1994, including all the essential elements of the subsequent NHTSA standards except for the energy absorption requirement (2). Between January 1, 1952 and January 25, 1998, trailers and semi-trailers were Federally regulated by Federal Motor Carrier Safety Regulations (FMSCR) that mandated rear impact guards, but allowed substantially smaller guards than the NHTSA standard and the TTMA recommended practice, and imposed no strength tests on the guards(3). The FMSCR has also applied to medium and heavy single-unit trucks (with certain exceptions) from January 1, 1952 to the present, but FMVSS 223/224 and RP 92-94 do not address single-unit trucks.
This evaluation will compare the crash performance of guards on trailers meeting the NHTSA standard and/or TTMA recommended practice to the smaller "pre-TTMA" guards on trailers meeting only the 1952 FMSCR standard. All trailers built on or after January 26, 1998 meet the NHTSA standard. Most trailers built between April 1994 and January 1998 meet the TTMA recommended practice, but not necessarily all trailers, because the TTMA recommended practice is voluntary and some trailer manufacturers do not belong to TTMA. Most trailers built before April 1994 have the smaller, pre-TTMA guards, but some might have already met the TTMA recommended practice. The following visual cues and data can help place the trailer and its guard in the NHTSA/TTMA group or the pre-TTMA group:
The primary objective of a rear impact guard is to prevent an especially hazardous crash configuration known as "underride with passenger compartment intrusion (PCI)" that can occur when the front of a passenger vehicle, especially a small passenger car, contacts the rear of a truck trailer. The rigid structures in the front end of a small passenger car, such as the engine block, are often less than 30 inches above the ground, whereas the bed of the truck trailer is usually more than 45 inches off the ground(4). The passenger vehicle can underride the trailer, and the bed of the trailer will almost immediately intrude into the occupant compartment above and behind the hood, at great risk to occupants.
To protect against underride with PCI, trailers should be equipped with a guard that is low enough to guarantee engagement with the rigid structures of even the smallest passenger cars, wide enough to assure contact in off-center impacts and strong enough to arrest the motion of the passenger vehicle before its compartment area reaches the trailer bed. The rationale for the NHTSA regulation is that the NHTSA/TTMA guards ought to be more effective at preventing underride with PCI than the pre-TTMA guards, because those are substantially smaller and not required to meet a strength test.
Specifically, NHTSA's regulatory evaluation predicts that NHTSA/TTMA guards would reduce the incidence of fatal crashes involving underride with PCI by about 10 to 25 percent relative to pre-TTMA guards(5). Nonfatal crashes usually involve lower speeds where guards are more likely to succeed in preventing underride with PCI(6) (but not necessarily in reducing injuries). Thus, in State data comprising nonfatal as well as fatal crashes, it is reasonable to anticipate that NHTSA/TTMA guards could reduce the overall incidence of underride with PCI by 25-30 percent, or even more, relative to pre-TTMA guards. In other words, the evaluation ought to include enough crash data that an observed 25-30 percent reduction would be statistically significant(7).
Underride with PCI is also possible when the front of a passenger vehicle impacts the rear of a medium or heavy single-unit truck. Crashes of single-unit trucks will be included in the data collection for the purpose of estimating the incidence rate of underride with PCI. This rate could be used in future analyses of the potential benefits of upgraded rear impact guards in single-unit trucks. However, since these trucks are only covered by the earlier FMSCR, not FMVSS 223/224 or RP 92-94, there will be no "before vs. after" analysis of the incidence rate.
Analysis plan Records of crashes in which the front of a passenger vehicle struck the rear of a truck trailer will be classified according to the type of rear impact guard on the trailer (NHTSA/ TTMA, or pre-TTMA) and the level of frontal damage on the passenger vehicle (two levels: underride with PCI, all others). The proportion of crash involvements resulting in underride with PCI is computed for each of the three types of guards, and statistically compared for the NHTSA/TTMA guards and the pre-TTMA guards in this hypothetical example:
|
Pre-TTMA |
NHTSA/TTMA |
Reduction |
Z |
|
|
Underrides with PCI |
80 |
50 |
||
|
Total N of impacts |
200 |
250 |
||
|
Proportion underride with PCI |
.4 |
.2 |
50% |
4.65 |
By 2001-2002 almost all truck-trailers on the road will be equipped or retrofitted with conspicuity tape, even those with pre-TTMA guards. The analysis will not be confounded by the presence or absence of the tape (which helps prevent or reduce the severity of nighttime rear impacts).
Separate analyses will also be performed for crashes where the striking vehicle is a small car (where the NHTSA/TTMA guard may have the highest potential effectiveness).
Data plan NHTSA proposes to obtain the crash data from State police in two large States over a two-year period (2001-2002), as an add-on to the data collection the agency is already planning to evaluate antilock brake systems (ABS). In the ABS study, the State police will furnish the basic crash report, plus a supplementary ABS-related crash report developed by NHTSA, for every crash involving a tractor-trailer or medium or heavy single-unit truck: about 10,000 tractor-trailer cases and 5,000 single-unit truck cases. In the subsets of approximately 1,000 trailers and 700 single-unit trucks that were struck in the rear by the front of a passenger vehicle (car, pickup truck, van or sport utility vehicle), the police will fill out a second supplementary form describing the impact guard on the truck and the damage on the passenger vehicle.
NHTSA recently concluded a similar data collection to evaluate the effectiveness of conspicuity tape for semi-trailers. Florida and Pennsylvania filled out a supplemental form on every trailer, specifying if it was equipped with the tape and describing the condition of the tape. The data collection was highly successful and, in combination with the basic crash reports, the data were just right for the evaluation.
The basic crash report will describe the crash configuration (front-to-rear, sideswipe, etc.), the specific damage locations on each vehicle (front-center, front-corner, etc.), data on each driver, and the make, model year, body type and VIN of the frontally impacted passenger vehicle. If photographs of the vehicle damage are part of the basic investigation, copies will be included with the report.
The supplemental form for this underride study will enable police to record information needed to classify the type of rear impact guard on the trailer, based on their inspection of the trailer, its VIN plate, and the guard. They will record the trailer make and type (van, flatbed, container, tank, etc.), the VIN number, and the build date. When feasible, they will look for the TTMA label on a guard vertical strut and the FMVSS Certification Label affixed on the horizontal cross member. Above all, they will have a set of pictures/diagrams/drawings of typical rear impact guard geometries and installations, ranging from FMVSS 223/224 types to very early models, and they can mark the picture(s) that most closely resembles the guard on the subject trailer. If the rear-impacted vehicle is a single-unit truck, they will record the type of truck and whether or not it was equipped with a guard.
The supplemental form will also enable State police to identify which of the passenger vehicles have underride with PCI. Specifically, the form will include a selection of top view and side view diagrams/photographs of underride damage to passenger vehicles (including no underride, underride without PCI, underride with PCI). State police will mark the diagram(s) that most closely resembles the damage to the case vehicle. The diagrams will also assist in identifying centered versus offset or oblique impacts on the guards.
The field-collected data should be more than adequate, in most cases, to identify the type of rear impact guard on a trailer. For the limited number of pre-FMVSS 223/224 trailers where identification is difficult, NHTSA hopes trailer manufacturers would be willing to analyze the VINs or consult their build sheets to identify the type of guard.
Statistical adequacy of this data plan As discussed above, a two-year census of crashes investigated by the State police in two large States is likely to include 1,000 cases that are initially classified as impacts by the front of a passenger vehicle with the rear of a tractor-trailer. A supplemental form would be obtained on each of those crashes. However, it would be appropriate to exclude about 320 of these cases from the analysis because they are basically sideswipes or angle collisions, with contact mainly on the corners of the vehicles, and to analyze the 680 cases that are genuine front-to-rear crashes likely to involve an impact into the guard.
Based on recent observations by NHTSA staff of the types of guards on trailers passing through weigh stations, NHTSA anticipates that by 2001-2002, about 469 of these 680 cases would involve a NHTSA/TTMA guard, while approximately 211 would involve a pre-TTMA guard.
In the 1989-99 National Automotive Sampling System - Crashworthiness Data System (NASS-CDS), 30 percent of all passenger vehicles that frontally impacted the rear of any t ruck-trailer had underride damage with PCI. NASS does not identify the model year or type of guard on the trailer, but since this is a mix of pre-TTMA and NHTSA/TTMA trailers, it implies the underride-with-PCI rate for the pre-TTMA trailers is at least 30 percent.
Given 211 pre-TTMA trailers, 469 NHTSA/TTMA trailers, a 30 percent underride-with-PCI rate in the pre-TTMA trailers and the analysis plan illustrated in the hypothetical example above, an observed 20 percent reduction in the underride-with-PCI rate in the NHTSA/TTMA trailers would produce a Z score that is statistically significant at the one-sided .05 level (i.e., > 1.645). A 28 percent reduction of underride-with-PCI would be significant at the one-sided .01 level.
Since NHTSA anticipates NHTSA/TTMA guards to be at least 25-30 percent effective in reducing underride-with-PCI in all types of crashes (fatal plus nonfatal), the proposed sample ought to be adequate, yet not excessive, for evaluating the effectiveness of rear impact guards.
The proposed sample of 700 single-unit trucks (including 550 genuine front-to-rear crashes) should be more than adequate for estimating the incidence rate of underride with PCI in the passenger vehicles that impact them.
Inadequacy of alternative data plans Data are useful for this evaluation only if they explicitly describe the type of guard on the trailer and the type of damage on the passenger vehicle. That fundamentally excludes unmodified State crash data files without VINs or other information on trailers (except the general trailer type). It also excludes files essentially compiled from State crash data, such as the Fatality Analysis Reporting System (FARS), the Motor Carrier Management Information System (MCMIS) or the NASS General Estimates System. In general, none of these files have much detail on the type of damage to the passenger vehicle, either.
NASS-CDS, as stated above, clearly identifies passenger vehicles that have underride with PCI, but has no details on the trailers. NASS-CDS could be modified in the future to document the trailer VIN and the type of guard, but the accession rate of 20 cases per year involving the front of a passenger vehicle hitting the rear of a tractor-trailers would not produce a statistically adequate sample (680 cases) for many years.
Only Florida crash data encode trailer VINs and model years, although for fewer than half the trailers involved in crashes. A weak analysis could be based on unmodified Florida data, using the trailer model year as a surrogate for the type of guard (possibly supplemented by manufacturer analyses of these VINs), and the injury rate of the drivers of the passenger vehicles as a surrogate for the proportion of passenger vehicles that had underride-with-PCI. However, the presence or absence of driver injury is a poor surrogate for a detailed damage pattern, and this analysis would do little more than describe the long-term trend in injury rates as a function of the trailer model year. It would be difficult to argue that a trend of decreasing injury rates is specifically due to improved rear impact guards.
Footnotes for Evaluation Of Rear Impact Guards For Truck Trailers
1. Code of Federal Regulations, Title 49, General Printing Office, Washington, 1999, Part 571.223 and 571.224. FMVSS 223 regulates the strength, geometry and energy absorption of the guard. FMVSS 224 regulates the configuration and installation of the guard on the trailer.
2. Rear Impact Guard and Protection, Recommended Practice RP92-94, Truck Trailer Manufacturers Association (TTMA), 1020 Princess Street, Alexandria, VA 22314, April 1, 1994. It uses the text of NHTSA's Supplemental Notice of Proposed Rulemaking on Rear Impact Guards and Rear Impact Protection (57 Federal Register 252, January 3, 1992), which did not include energy absorption requirements. RP 92-94 was revised in 11/94 to include tank trailers.
3. Code of Federal Regulations, Title 49, General Printing Office, Washington, 1999, Part 393.86.
4. Final Regulatory Evaluation - Rear Impact Guards, FMVSS No. 223 and Rear Impact Protection, FMVSS No. 224, Submission to NHTSA Docket No. 01-11-N10-003, Washington, 1995, pp. II-23 - II-29.
5. Ibid., p. V-5.
6. Ibid., p. V-4.
7. A secondary objective of the guards is to absorb some of the collision energy, providing some additional protection to occupants. FMVSS 223 has an energy absorption requirement whereas RP 92-94 does not. A study of this requirement is not included in the proposed evaluation because it is unknown how many 1994-97 guards could have met the FMVSS 223 test, and because there do not appear to be any crash data suitable for analyzing it.