SAE TECHNICAL PAPER SERIES 2005-01-1837 A New Database of Child Anthropometry and Seated Posture for Automotive Safety Applications Matthew P. Reed, Sheila M. Ebert-Hamilton, Miriam A. Manary, Kathleen D. Klinich and Lawrence W. Schneider University of Michigan, Transportation Research Institute Reprinted From: Air Bags,Occupant Protection and Child Passenger Restraints (SP-1933) 2005 SAE World Congress Detroit, Michigan April 11-14, 2005 400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760 Web: www.sae.org The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session organizer. This process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. For permission and licensing requests contact: SAE Permissions 400 Commonwealth Drive Warrendale, PA 15096-0001-USA Email: [email protected] Tel: 724-772-4028 Fax: 724-772-4891 For multiple print copies contact: SAE Customer Service Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA) Fax: 724-776-1615 Email: [email protected] ISSN 0148-7191 Copyright © 2005 SAE International Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. Persons wishing to submit papers to be considered for presentation or publication by SAE should send the manuscript or a 300 word abstract to Secretary, Engineering Meetings Board, SAE. Printed in USA 2005-01-1837 A New Database of Child Anthropometry and Seated Posture for Automotive Safety Applications Matthew P. Reed, Sheila M. Ebert-Hamilton, Miriam A. Manary, Kathleen D. Klinich and Lawrence W. Schneider University of Michigan, Transportation Research Institute Copyright © 2005 SAE International ABSTRACT Belt-positioning boosters have been shown to be effective in preventing these injuries. Durbin et al. This paper presents a laboratory study of body (2003) examined data from crashes involving 4243 dimensions, seated posture, and seatbelt fit for children children ranging in age from four to seven years in a weighing from 40 to 100 lb (18 to 45 kg). Sixty-two boys large, cross-sectional study of crash-involved children. and girls were measured in three vehicle seats with and The odds of injury, after adjusting for a variety of without each of three belt-positioning boosters. In potentially confounding factors, were 59% lower for addition to standard anthropometric measurements, children using belt-positioning boosters than for children three-dimensional body landmark locations were using only the seat belt. No injuries to the abdomen, recorded with a coordinate digitizer in sitter-selected and spine, or lower extremities were reported for children standardized postures. This new database quantifies using boosters, whereas children using belts alone had the vehicle-seated postures of children and provides injuries to all body regions. quantitative evidence of the effects of belt-positioning boosters on belt fit. The data will provide guidance for Using data from the same ongoing study, Nance et al. child restraint design, crash dummy development, and (2004) examined the incidence of abdominal injury in crash dummy positioning procedures. restrained children. Optimal restraint was defined as usage of a child safety seat with harness for children INTRODUCTION four years of age or less; lap-shoulder belt with a belt positioning booster for children four to eight years of Occupant protection for children who have outgrown age; and lap-shoulder belt for children older than eight harness-based restraints has received increased years. Any other restraint configuration, such as a four- regulatory and legislative attention in recent years. The year-old child in a vehicle belt alone, was considered U.S. National Highway Traffic Safety Administration suboptimal. An analysis of data from large-scale survey (NHTSA) recommends that children who are ages four of children involved in crashes showed that abdominal to eight and over 40 lb (18 kg) use belt-positioning injuries were three times less likely among children who boosters (also known as booster seats) with a three- were optimally restrained. The percentage of restrained, point (lap-shoulder) vehicle belt (NHTSA 2004). A belt- crash-involved children who sustained abdominal injury positioning booster is designed to position the child and peaked in the six- to eight-year-old range within which route the vehicle seatbelt such that the child is well the percentage of children who were optimally restrained restrained by the belt. (using belt-positioning boosters, under this definition) was less than ten percent. The recent emphasis on booster usage follows evidence from the field that children from four to eight years of age Arbogast et al. (2004) analyzed data from the same are less likely to be properly restrained than vehicle study to show that children four to eight years old are at occupants in any other age group. Many children move significantly greater risk of moderate or severe directly from harness restraints to vehicle belts even abdominal injuries than older or younger children, as though the children are often too small to obtain good indicated by a score of 2 or greater on the Abbreviated restraint from the vehicle belts alone. The serious Injury Scale (AIS). Among this age group, children who consequences of poor belt fit include injuries to the were restrained by a vehicle belt alone (lap or lumbar spine and abdomen as the lap portion of the belt lap/shoulder) were about 34 times more likely to sustain slides off the pelvis and loads the abdomen during a an AIS2+ abdominal injury than those using belt- frontal impact. This pattern of occupant motion relative positioning boosters or harness restraints. to the belt is often referred to as submarining. In the past three years, at least 22 U.S. states have dimensions made the dummies much less susceptible to passed legislation that addresses occupant protection submarining than children. Submarining occurs when for children over four years of age, and similar legislation the pelvis slides below the lap portion of the belt, has been introduced in the remaining states (SAFE allowing the belt to load directly onto the abdomen. KIDS, 2004). Most state child-restraint laws classify These limitations made the dummies insensitive to children based on age, typically requiring child restraint differences in booster design and belt routing that would use up to six years of age. Most parents and caregivers significantly affect safety for child occupants. Using a rely on legislation as their guide to appropriate restraint modified P3 dummy, the authors showed that good use, so legislation is seen as a critical step for reducing booster seat design requires lap belt guides that hold the injury for this population. However, because children of belt flat on the upper thighs of the child, rather than the same age vary widely in size, age-based regulation targeting the anterior surface of the pelvis. The data with too low of a cutoff age may not achieve the goal of from Chamourd et al. are limited in that the child properly matching children and restraints. anthropometry was measured in postures that are not representative of child postures in vehicle seats or In 2002, the U.S. Congress passed legislation known as boosters, and three-dimensional measurements of Anton’s Law that directs NHTSA to take a variety of skeletal posture were not made. In particular, the seated actions addressing safety for older children. Among the pelvis orientation for the child volunteers and the routing major provisions, NHTSA was directed to revise FMVSS of the belt relative to the skeleton were not reported. 208 to require lap-shoulder belts in all rear seating positions, to incorporate in regulation a crash dummy Other data on child body dimensions have been used for representing a 10-year-old child, and to study the child restraint design. Weber et al. (1985) extracted data effectiveness of vehicle-integrated child restraints. In from a large-scale survey of standard and functional addition, NHTSA was directed to “consider whether to child anthropometry (Snyder et al. 1977) to recommend establish performance requirements for seat belt fit when dimensions for child restraints. These data are limited, used with booster seats and other belt guidance however, by the differences between the vehicle-seated devices.” postures and the measurement postures, and by the lack of detailed information on the three-dimensional Few previous studies have examined belt fit for children positions and orientations of key skeletal structures, in booster seats. Klinich et al. (1994) studied belt fit for such as the pelvis. 155 children between seven and twelve years of age in three rear vehicle seats and three belt-positioning The current study was conducted to address some of the boosters. Belt fit and posture were evaluated limitations of previous research on children in booster subjectively using multi-category scales by examining seats. In particular, the objectives of the current study videos of the children in each of the test conditions. The are to: boosters were found to improve belt fit significantly for most children in most of the tested combinations of 1. develop a detailed database on vehicle-seated booster and vehicle seat. The researchers noted that anthropometry, posture, and position for one cause of poor belt fit in the absence of a belt- children and adolescents sitting in harness positioning booster was the tendency of the children to restraints, belt-positioning boosters, and slouch rather than to sit maximally rearward on the seat. vehicle seats with three-point belts; Discomfort associated with extended knees was identified as the impetus for the slouched posture, and 2. develop a positioning procedure for six-year- the boosters were shown to allow postures with greater old and 10-year-old crash dummies that knee flexion and less slouching. Poor belt fit resulting provides representative posture and position from slouching has been implicated in abdominal injuries on booster seats and vehicle seats; (Glassman et al. 1992; Johnson et al. 1990), particularly when children are restrained by a lap belt alone. 3. develop a physical belt-fit assessment procedure for children from 40 to 100 lb that Chamourd et al. (1996) identified significant problems uses crash dummies (six-year-old and 10- with the pelvis and thigh dimensions of the 3YO, 6YO, year-old Hybrid III); and and 10YO P3 dummies and 3YO and 6YO Hybrid-III dummies. Pelvis measurements from 54 children sitting 4. identify differences between children and on the ground with their backs against a wall were crash dummies that could adversely affect the compared to dummy dimensions in similar postures. fidelity of crash-test assessments of belt- Radiographs of seven children in standing and seated positioning boosters. postures were used to assess the fidelity of the crash dummy pelves. They authors concluded that the This paper reports on the development of the child combination of unrealistic thigh flesh and pelvis anthropometry database. Seated posture and belt fit were measured for 62 boys and girls weighing between Vehicle Belt Configurations 40 and 100 lb in a range of vehicle- and booster-seat conditions. Three-dimensional data on body landmark The vehicle mockup was equipped with a three-point belt and seatbelt locations were obtained using a portable system with a sliding latchplate and emergency locking coordinate measurement machine. The data will be retractor obtained from a late-model sedan. The made available in a relational database to aid in the retractor and D-ring were mounted to an adjustable improvement of restraint systems for children. fixture that provided a large range of fore-aft, vertical, and lateral adjustability. Each seat was equipped with a METHODS webbing-mounted buckle. In seat B, the location of the buckle anchorage could be varied over a wide range. Reconfigurable Vehicle Mockup The outboard anchorage was located in the same position with respect to H-point for each seat. The Testing was conducted in an UMTRI laboratory using a outboard anchorage location routed the belt near the reconfigurable mockup of a vehicle rear seating area seat bight, simulating typical rear-seat belt geometry. shown in Figure 1. Four vehicle rear seats, shown in Figure 2, could be interchanged in the mockup. Two Child Restraint Systems (CRS) seats were obtained from sedans and one (seat 2) from a minivan. The seats were mounted high enough from Figure 3 shows the four child restraints used in the the floor that none of the children were able to touch the study. CRS-A is a backless belt-positioning booster floor with their feet while sitting all the way back on the equipped with fixed-position belt guides. A flexible seat, reproducing the typical situation for children in rear attachment provided with the restraint that is used to vehicle seats. Testing was conducted in the right improve torso belt fit was not needed with the tested belt outboard seating position on each seat, except that geometries and hence was not used. CRS-B is a high- testing with seat D was conducted only in center seating back belt-positioning booster with an adjustable back position, which was equipped with an integrated harness angle. The back is height adjustable and is equipped restraint. The H-point location, seat back angles, and with a path for the torso portion of the belt. During seat cushion angles were measured for each seat using testing, the back height was adjusted for each participant the procedures in SAE J826 (SAE, 2004). All three seats to place the torso belt guide above the shoulders, were mounted so that the seat cushion angle (SAE A27) following the manufacturer’s instructions. The lap belt was 14.5 degrees, a typical value for vehicle seats. The guides, which are height adjustable, were tested at their seat back angle was varied across test conditions. lowest setting. CRS-C is a rigid-shell combination restraint that can be used with an integrated five-point harness or as a belt-positioning booster. Testing was conducted using the restraint as a belt-positioning booster with the harness removed. Open slots in the shell provide a lower belt path and a three-position clip on the upper portion of the back is provided to assist in routing the torso portion of the belt. CRS-D is a forward- facing-only restraint with a five-point harness that is rated by its manufacturer for children up to 36 kg (80 lb). CR-E is a restraint with a five-point harness that is integrated into a minivan bench seat. The vehicle owners’ manual recommends the restraint for children from 10 to 18 kg (22 to 40 lb). Figure 1. Reconfigurable vehicle rear-seat mockup in laboratory. The FARO Arm coordinate measurement device is visible at the right side of the picture. Figure 2. Vehicle rear seats used in testing. From left, seats 1, 2, and 3. A B C D E Figure 3. Child restraints shown with 6YO and 3YO crash dummies (CRS E is integrated in vehicle seat 4). Test Conditions Trials 13, 14, and 15 used child restraints B, C, and D on seat 1 with a fixed back angle of 23 degrees. In trial 16, Data were gathered in the 20 test conditions listed in children were tested on seat 2 without a child restraint. Table 1. The test conditions were selected to Trials 17-19 were conducted on seat 3 with no child investigate particular factors that were hypothesized to restraint and with three different buckle locations to affect child posture or belt fit. Most of the testing was quantify the effect of seatbelt geometry on belt fit. In trial conducted in vehicle seat 1, a low-contour seat from a 20, children who meet the weight requirements for the popular sedan. The first twelve test conditions examined integrated restraint were tested in seat 4 (restraint E). the effects of seat back angle, sitting instructions, and In test conditions 1-14 and 16 the buckle location was booster use. The first six “sitter-selected posture” trials set to position measured in the vehicle from which the were conducted at the start of each participant’s testing. test seat was obtained. The D-ring position was set to Tests with and without the CRS were blocked, and the location in the vehicle model from which each seat was order of the blocks and of trials within each block were obtained, except that the D-ring position was rotated randomized. The participant was instructed to sit in the around the H-point with seat back angle changes. vehicle seat or child restraint and to put on the belt. The child chose his or her preferred posture and buckled the Participants and Standard Anthropometry belt without coaching or further instruction. Test conditions 7-12 were conducted next, using the same Sixty-two children (27 girls and 35 boys) were recruited blocking and randomization approach used in trials 1-6. for testing by word-of-mouth, fliers, and newspaper In these trials, the participants were instructed to sit as advertisements. Table 2 summarizes some of the far back as possible on the seat with the arms and legs participant characteristics. The goal was to recruit sagittally symmetric and the hands on the thighs. This children who spanned the range of potential users of “standard” posture conditions are intended to produce belt-positioning boosters with respect to stature and less-variable postures that might be appropriate to weight, including the range between the masses of the reproduce with crash dummies. The investigator 6YO and 10YO Hybrid-III crash dummies (23 kg and 32 deployed and buckled the belt in the standard-posture kg, respectively). Anthropometric dimensions were trials. recorded using standard techniques (Roebuck, 1994). Figure 4 shows the measurement of knee height. Table 1 Test Conditions Seat Back Buckle Angle (SAE Condition Vehicle Seat CRS Posture Location A40) 1 1 A Sitter-Selected Standard 19 2 1 A Sitter-Selected Standard 23 3 1 A Sitter-Selected Standard 27 4 1 None Sitter-Selected Standard 19 5 1 None Sitter-Selected Standard 23 6 1 None Sitter-Selected Standard 27 7 1 A Standard Standard 19 8 1 A Standard Standard 23 9 1 A Standard Standard 27 10 1 None Standard Standard 19 11 1 None Standard Standard 23 12 1 None Standard Standard 27 13 1 C Standard Standard 23 14 1 B Standard Standard 23 15 1 D Standard n/a 23 16 2 None Standard Standard 23 17 3 None Standard Low 23 18 3 None Standard Middle 23 19 3 None Standard High 23 20 4 E Standard n/a 23 Table 2 Participant Characteristics and Standard Anthropometric Measures Measure* Mean S.D. 5th%ile 25th%ile 50th%ile 75th%ile 95th%ile Age (years) 8.4 1.9 5 7 9 10 11 Stature 1334 115 1121 1245 1354 1430 1483 Stature (in) 52.5 4.5 44.1 49.0 53.3 56.3 58.4 Weight (kg) 31.8 7.8 19.4 25.1 32.1 37.3 43.8 Weight (lb) 70.1 17.3 42.7 55.3 70.8 82.3 96.5 Head Length 177 9 164 171 175 182 191 Erect Sitting Height 706 52 609 675 719 744 770 Shoulder Height 455 58 381 417 461 478 504 Acromion Height 436 42 372 401 448 469 489 Knee Height 401 41 329 375 405 435 455 Bideltoid Breadth 328 34 275 302 332 351 385 Biacromial Breadth 248 29 198 231 249 268 294 Hip Breadth 268 29 220 246 272 290 310 Shoulder-Elbow Length 286 27 242 264 291 310 322 Elbow-Fingertip Length 355 36 298 332 359 382 401 Buttock-Knee Length 457 49 377 418 470 494 519 Chest Depth† 152 17 124 140 149 168 178 Abdomen Depth† 168 25 135 152 165 182 214 Chest Width† 224 23 189 205 227 239 257 Abdomen Width† 213 26 178 192 208 234 260 * Dimensions in mm unless otherwise noted. † Measured in hardseat in automotive posture (see text). Three-Dimensional Anthropometry The three-dimensional location of body landmarks was measured in this study using a FARO Arm coordinate digitizer (FARO Technologies, Lake Mary, FL). The procedures were very similar to those used previously in many studies of adult occupant posture and position (Reed et al. 1999). The investigator located the desired landmark by palpation, placed the tip of the FARO Arm probe on the landmark, and pressed a button to record the location. All data were expressed in a laboratory coordinate system with the X axis positive rearward, Y- axis positive to the right, and Z-axis positive upward. Landmark data were recorded in each of the test conditions in Table 1 and while the participant sat in a specially constructed laboratory hardseat. The hardseat, shown in Figure 6, is designed to produce a posture similar to a vehicle-seated posture but to provide access Figure 4. Fixtures for standard anthropometric measurements to posterior landmarks on the spine and pelvis. showing measurement of knee height. Table 3 lists the landmarks that were recorded during Figure 5 shows the stature and weight of the participants the vehicle-seat and hardseat trials. Reference points in the context of the 1990 U.S. population based on on the seat, CRS, and belt were recorded where NHANES III (see NCHS, 2004). Percentiles of weight applicable. The reference points allow the body were computed in 5-cm bins for children from 3 to 14 landmark data to be referenced to a seat or CRS years old in NHANES III using the appropriate sample coordinate system. Points were digitized on the seatbelt weights. The actual (unweighted) observations from or harness where it passed over the sternum, clavicle, NHANES III are shown in gray. As intended, the and the lateral positions of the left and right anterior- sample spans the reference stature and weight for both superior iliac spines (ASIS). The landmark data for the the 6YO and 10YO Hybrid-III dummies. participant were sufficient to define the three- dimensional locations of the major skeletal components, including the head, thorax, pelvis, clavicles, and the right humerus and femur. Figure 7 shows the landmarks schematically. In the hardseat, surface landmarks over the C7, T4, T8, T12, L3, and L5 spinous processes were recorded along with the locations of the left and right posterior-superior iliac spines (PSIS). Combined with the ASIS points, the PSIS points give the three- dimensional position and orientation of the pelvis. The spinous process landmarks were digitized twice and the pelvis landmarks four times to ensure that a consistent set of points was obtained. Figure 5. Stature and weight for the current sample compared with children from 3 to 14 years old in NHANES III (1990 U.S. population). Lines show quantiles in NHANES III based on 5- Figure 6. Participant seated in a laboratory hardseat used to cm-wide bins. The reference stature and weight for the 6-year- obtain posterior and anterior body landmarks in same posture. old and 10-year-old Hybrid-III ATDs are shown (Mertz et al. A slightly larger hardseat was used with the larger participants. 2001) Table 3 Body, CRS, and Seat Landmarks Digitized During Testing Hardware Reference Points Vehicle Seat Points (3*) Torso Belt on Clavicle (2) D-ring Location (3) Torso Belt on Sternum (2) Lap Belt Anchor Lap Belt at ASIS-R (2) Buckle Stalk (2) Lap Belt at ASIS-L (2) CRS Reference Points (5) Landmarks on Participant Top Head Lat. Femoral Condyle (R) Glabella Suprapatella Infraorbitale (R) Lateral Malleolus (R) Corner of Eye (R) Heel of Shoe Tragion (R) Toe of Shoe Figure 7. Schematic illustration of body landmarks digitized Acromion (R) Shoulder Clearance Point during testing. Posterior spine landmarks and PSIS were recorded only in the hardseat. Acromion (L) Hip Clearance Point Lat. Humeral Epicondyle (R) Elbow Clearance Point Pelvis Angle Measurement Wrist (R) ASIS (R) A new method was applied to quantify pelvis orientation Suprasternale ASIS (L) in the vehicle-seated conditions. A low-profile Substernale inclinometer capable of measuring orientation with respect to gravity on two axes was taped to the skin over Participant Landmarks Recorded Only in Hardseat the sacrum. Thin-film pressure transducers under the C7 PSIS (L) inclinometer plate provide compensation for changes in orientation due to deformation of soft tissue under the T4 PSIS (L) inclinometer. During the hardseat landmark T8 Abd. Midsagittal Profile (7) measurements, the inclinometer pressure compensation T12 Abd. Lateral Profile (7) was calibrated by pressing on the inclinometer with a range of pressure levels and gradients. Figure 8 shows L3 the inclinometer on a participant’s sacrum. The data L5 from the ASIS and PSIS locations measured in the * Numbers in parentheses indicate the number of points hardseat were used to convert the inclinometer- digitized in each category. (R) and (L) indicate right and left, measured angles to a three-dimensional representation respectively. of the orientation of the bony pelvis. Figure 8. Pelvis inclinometer on participant’s sacrum. Some tape has been removed to provide a better view of the sensor.