Documentation:FIB book/Review of Side Impact Anthropomorphic Test Devices (ATDs)

Review of Side Impact Anthropomorphic Test Devices (ATDs)

Introduction

Background

Anthropomorphic test devices (ATDs) or “crash test dummies” are used extensively for crash testing in the automobile industry. They are designed to mimic the size and shape (anthropometry) and material response of human occupants during a crash (biofidelity) and are instrumented to collect data on the forces experienced during a crash test. Due to these requirements, different testing conditions often require alternate ATD designs. Dummies like the Hybrid III are currently used in frontal restraint evaluation by regulatory bodies such as the Federal Motor Vehicle Safety Standards (FMVSS), and are designed to be biofidelic primarily in frontal crash tests. While frontal crashes are more frequent than side impacts, injuries obtained through side impacts are often more severe making side impact tests an important area of study. Side impact ATDs are designed specifically for heightened biofidelity during lateral impacts. Therefore, they are used to test side impact protection in vehicles and are also recognized by regulatory bodies such as the International Standards Organization (ISO).

History

While the first ever crash test dummies were designed in 1949 by Sierra Engineering, these were not designed for any particular crash scenario. For 30 years mid-size adult male frontal crash test dummies such as the Hybrid II and III were used for a variety of crash tests until the first Side Impact Dummy (SID) was created in 1979 under contract for the US Department of Transportation, National Highway Safety Administration (NHTSA). SID was based on the Hybrid II, but included a steel rib and spine structure without arms. SID was eventually replaced in FMVSS regulations by SID-HIII which was altered from the original by replacing the Hybrid II head and neck structure with that of the Hybrid III’s.

In the late 1980s, a side impact dummy was developed in Europe (EUROSID) by the European Experimental Vehicle Committee (EEVC). Following thorough evaluation by governments worldwide and organizations such as ISO and the Society of Automotive Engineers (SAE), the EUROSID-1 was created and written into regulation in Europe, Australia, and Japan despite limitations to its biofidelity and instrumentation.

Due to the limitations of SID and EUROSID-1, another Biofidelic Side Impact Dummy (BIOSID) was designed in 1989 by the SAE. The BIOSID imitates the physical characteristics and lateral impact response of a fiftieth percentile male. In testing, the BIOSID showed “excellent biofidelity for head, neck, shoulder, thorax and abdomen responses” and includes arms and shoulders. Despite the improved biofidelity of the BIOSID, it has never been included in regulation due to the concurrent development of similar ATDs with improved capabilities such as the WorldSID and EUROSID-I.

In the 1990s, a small size side impact dummy (SID-IIs) with the physical stature of a 5th percentile female was designed to meet the scaled down ISO targets of a mid size male. SID-IIs incorporates parts from the Hybrid III female dummy and the scaled down abdomen of the BIOSID.

By 2001, an internationally funded Task Group designed the WorldSID dummy with increased biofidelity and instrumentation capabilities. The head and neck of the WorldSID is designed to be biofidelic both laterally and sagittally. It has also shown more potential than the SID, EUROSID and BIOSID as a far-side impact crash test ATD and is considered the most biofidelic of all mid-size male side impact dummies.

Other side impact dummies include the WorldSID Small Female dummy developed in 2004 and the Q3 dummy meant to represent a 3-year old child. The Q3 is proposed for use in FMVSS regulation in assessment of child restraint designs during a side impact sled test.

Side Impact ATD Overview

Types of Side Impact ATDs

Side Impact ATD Models

Four main families of side impact ATDs have been developed and evaluated to improve lateral crash testing: SID, EUROSID (and its modified forms, ES-2 and ES-2re), BIOSID, and WorldSID. Each generation has addressed limitations in biofidelity, instrumentation, and durability identified in its predecessors.

SID and SID-HIII

The first SID, developed in 1979, was derived from the Hybrid II midsize male dummy with a modified thoracic structure and no arm or shoulder assemblies. Its chest consists of a hydraulic shock absorber connecting five steel ribs to the spine. Shortcomings included the absence of a shoulder load path, unrealistic thoracic compliance, and limited instrumentation since it lacked sensors for neck, shoulder, abdominal loads, and rib deflection. As more biofidelic dummies became available, SID was phased out of FMVSS 214 in 2012.

The SID-HIII was later developed by replacing the Hybrid II head and neck with those of the Hybrid III, improving head–neck kinematics and measurement precision.

EUROSID Family

In 1986, European research organizations collaborated under the EEVC to create the EUROSID dummy. The improved EUROSID-1 was released and adopted for compliance testing in multiple countries after evaluation by ISO and SAE through 1987-1989.

EUROSID-1 expanded instrumentation in the ATD. However, it exhibited limited biofidelity from rib deflection “flat-topping” (where the data appears to plateau) due to piston rod binding, and lacked load measurements at the neck and in the pelvic region. Thus, ES-2 was developed, adding additional load cells (upper/lower neck, clavicle, T12, and femur) and redesigned rib guides to reduce binding. NHTSA’s modified version, ES-2re, incorporated rib extensions to prevent intruding into the foam of the vehicle seatback. ES-2re was adopted into FMVSS 214 in 2007 and is one of the most widely used regulatory dummies for side impact testing.

BIOSID

The BIOSID was developed by SAE in 1989 to improve side impact realism after evaluations of SID and EUROSID. It was the first dummy designed to meet ISO impact response biofidelity targets for key body regions including the head, neck, shoulder, thorax, abdomen, and pelvis. The BIOSID demonstrated good repeatability, durability, and instrumentation performance. However, despite its technical strengths, it was never formally incorporated into regulatory use and is used mainly as a research tool. This may be due to the high costs associated with changing regulation standards and testing, and due to the development of new models with higher biofidelity, such as the WorldSID.

WorldSID

WorldSID

WorldSID development began in 1997 under ISO TC22/SC12, with international funding and representation from Asia-Pacific, Europe, and North America. The first prototype appeared in 2001. WorldSID features a highly symmetric thoracic and abdominal design, biofidelic head and neck response in both lateral and sagittal directions, and anatomically realistic limbs. With over 170 potential data channels, it provides unprecedented measurement capability. According to ISO biofidelity ratings, WorldSID exhibits the highest overall biofidelity among midsize male side impact dummies and serves as the current global standard for advanced side impact testing.

Female and Child Side Impact ATDs

The development of adult and child side impact ATDs such as SID-IIs, WorldSID 5th Female, and Q3s has also expanded our ability to assess how occupant size, sex, and age affect injury outcomes in side impact collisions. These dummies were created to address research and regulatory gaps, providing more representative models for female bodies and children. However, similar to earlier generations, limitations in scaling, biofidelity, and anatomical accuracy remain, especially in replicating the diversity of real-world human responses.

SID-IIs

The SID-IIs represents a 5th percentile adult female and was developed to evaluate side impact protection countermeasures such as airbags. It incorporates Hybrid-III small female components and a scaled BIOSID thorax, abdomen, and pelvis structures. Although the SID-IIs demonstrates good biofidelity and instrumentation performance, its overall anthropometry and stiffness are not entirely representative of a small female, due to it being slightly lighter than the Hybrid-III female equivalent. The dummy lacks fine-tuned representation of female body composition and tissue distribution, which may affect the accuracy of injury predictions for real female occupants. It was approved for use in FMVSS 214 compliance testing alongside ES-2re in 2007.

WorldSID Small Female

The WorldSID 5th Female dummy, developed through the European Advanced Protection Systems (APROSYS) program in 2004, is a more sophisticated representation of female occupants compared to SID-IIs. Based on the geometry of the midsize male WorldSID, it integrates two-dimensional IR-TRACC sensors (tracking rib displacement) and enhanced instrumentation across the body. While the WorldSID small female achieved good overall biofidelity, it remains primarily a scaled version of the male dummy rather than a uniquely biomechanically characterized model. This geometric scaling fails to capture key sex differences in bone structure, muscle mass, and pelvic anatomy, leading to inaccuracies in torso mass distribution and pelvic response during impacts.

Q3s

The Q3s represents a 3-year-old child and was developed to improve upon the original Q3 dummy used in frontal impact crash testing, in terms of lateral biofidelity and structural robustness. With redesigns to the neck, shoulder, thorax, and hip joints, the Q3s provides enhanced measurement capability, including lateral shoulder deflection and tri-axial force sensing at the shoulder and pelvis. However, despite its improved instrumentation, discrepancies remain when comparing Q3s data with pediatric cadaver results, particularly in shoulder and spine stiffness.

Structure & Instrumentation

Biofidelity is controlled by the design, material, and instrumentation of the ATD. This section will focus on contrasting the construction and instrumentation of 50th percentile male side impact dummies: SID, SID-HIII, Eurosid-1, EuroSID-2/ES-2, ES-2re, BioSID, and WorldSID. Since the early development of the SID and SID-HIII, the structure of side impact ATDs have seen many changes and improvements. SID and SID-HIII are based mostly on the construction of the Hybrid-II and Hybrid-III 50th percentile male frontal impact ATDs, with improvements in biofidelity of the head and neck of the SID-HIII. The development of ATDs of EuroSID and BioSID brought unique neck, shoulder, and thorax designs. Finally, the recent WorldSID ATD brought about the most significant design changes and advancements in lateral impact injury assessments in vehicles.

Head

The head and neck construction has evolved significantly over the course of side impact ATD development, improving biofidelity and data collection during complex loading from modern automotive airbags and vehicle structures. The first side impact ATD, SID, utilizes the same head as the previously developed frontal impact dummy, the Hybrid-II. The SID-HIII improved biofidelity by using the head of the Hybrid-III 50th percentile dummy, which provides better head-neck responses (i.e. less stiff than SID head). The standard Hybrid-III head consists of an aluminum shell covered by a pliable vinyl skin that can be instrumented with triaxial head accelerometers and upper neck force transducers. The EuroSID, EuroSID-2/ES-2, and ES-2re heads are also based on the Hybrid-III head. The WorldSID ATD introduced the most advanced head design to date, with a polymer skull and polyvinyl skin offering highly biofidelic material properties and more advanced instrumentation. The head is designed with a cone shaped core to house triaxial linear accelerometers, three rotational accelerometers, and a six axis neck load cell to measure forces and moments near the occipital condyle joint. There is also a dual axis tilt sensor to aid with positioning of the dummy during test preparation.

Neck

The necks of the SID and SID-HIII were also reused from the Hybrid-II and Hybrid-III dummies respectively. In addition to improving the head-neck response from the Hybrid-II, the Hybrid-III dummy introduces upper neck load measurements. While the EuroSID-1 also used an Hybrid-III head, it incorporated a unique neck design, in which a central column of rubber connects the head and spine via metal discs or interface plates. The head-neck and neck-torso interfaces utilize a half-spherical screw, providing rotation at the top and bottom of the neck. Improving on the Eurosid-1, the EuroSID-2 or ES-2 added a lower neck load cell. The WorldSID neck consists of a rubber component with aluminum end plates and two spherical joints at the top and bottom. The range of motion and joint stiffness is controlled by four rubber buffers at each end.

Thorax

Vehicle pillar structures A, B, and C

Two WorldSID dummies in Side Impact test

The thorax, being the shoulders and chest, is a critical region for injury assessment with vehicle structures such as the B-pillar (a mid-positioned structural member that supports the roof and protects occupants during side impact and rollover), often interacting with these regions during lateral impact. The design and construction of the thorax varies significantly between side impact ATDs.

Shoulders

While SID is mostly based on the Hybrid-II dummy, it lacks a shoulder structure and therefore a shoulder load path. SID also lacks the instrumentation to measure neck and shoulder loads. The EuroSID-1 shoulders utilize a cam block attached to the top of a spine box that allows two polypropylene clavicles to rotate forward around it. This allows the arms which are attached to the clavicle to rotate. A shoulder cap made of foam is attached to the top of the cam block. The EuroSID-2/ES-2 improved on this design by reducing friction to prevent clavicle binding and created a more biofidelic clavicle response by limiting vertical motion. Other modifications include reshaping the foam shoulder cap, and adding a clavicle load cell. The BioSID utilizes a pin/clevis joint to attach the arm to the shoulder to allow lateral rotation. The WorldSID 50th percentile ATD has a shoulder yoke design, which attempts to simulate the human arm range of motion. However, the yoke allows for only two degrees of freedom, whereas the human shoulder has three degrees of freedom. The WorldSID 5th percentile ATD included a spherical shoulder joint, which added the final third degree of freedom.

Chest and Ribs

The SID chest structure consists of a hydraulic shock absorber that links five steel ribs to the spine. A major deficiency of the SID is that it has no instrumentation to measure abdominal loads and rib deflection. The EuroSID-1 chest consists of a spine box and three identical rib modules. Each rib module is made of a spring steel hoop covered in plastic and attaches to a piston/cylinder and spring/damper system. The EuroSID-2/ES-2 added a back plate load cell and T12 spinal load cell. It also improved on a key issue of the EuroSID-1 by adding needle bearing rib guides to prevent rib binding. The ES-2re further improved on the rib-back plate connection by adding a smooth transition surface to prevent foam from penetrating the large gap in the rib-back plate connection of the ES-2. The BioSID utilizes a "far mounted" rib structure, in which steel ribs are rigidly attached to the spine box on the side opposite the impact. This is intended to make the rib structure deform symmetrically about the coronal plane. Hybrid-III polymer-based dampening material is used on the fore and aft surfaces of the rib for viscous bending resistance and energy dissipation. The WorldSID also introduced a more complex thorax design, with several rib modules made of inner and outer ribs. The thorax is composed of three rib modules on each side, while the abdomen is represented by two rib modules. Each rib is equipped with IR-TRACC sensors to measure deflection and acceleration.  

Pelvis and Lumbar Spine

There is little public documentation on the SID and SID-HIII pelvis and lumbar construction, other than that it is based largely on Hybrid-II and Hybrid-III. The EuroSID uses a pelvis with two aluminum iliac wings linked by a sacrum block. The hip joints are ball joints attached to the iliac. The EuroSID-2/ES-2 utilizes a very similar pelvis structure, but with a slightly simplified pubic attachment hardware. The BioSID pelvis structure is also very similar to that of the EuroSID. The WorldSID pelvis has a polymer bone structure with polyvinyl chloride (PVC) skin. The left and right structure is connected to central sacro-iliac and pubic load cells. The pelvis also houses a rotational accelerometer (x-direction), a linear triaxial and a dual axis tilt sensor. A ball and socket joint is used at the hip.

The lumbar spine of the EuroSID is a standard FMVSS 208 part number 572 unit. The part is designed with a solid rubber column with top and bottom metal plates and a steel cable running through it. The EuroSID-II/ES-2 and ES-2re utilize a similar lumbar design. The BioSID lumbar spine is composed of a rubber cylinder with a braided wire running through the center and a 5 axis load cell can be installed at the base of the lumbar-pelvis interface.The WorldSID uses a novel omega-shaped lumbar design to allow more biofidelic shear motion between the pelvis and the upper body regions.

Extremities

The extremities, being the arms and legs, are generally not as heavily instrumented as the thorax or pelvis in side impact ATDs.

Arms

WorldSID Side Impact Dummy showing "half arm"

The SID has no arm structure and thus no arm response or instrumentation. Other side impact ATDs utilize a “half arm”, in which the arm is approximately half the anatomical length and terminates at the lower thorax. The EuroSID-2/ES-2 and ES-2re both utilize a plastic arm skeleton covered in foam with a plastic skin. The arms utilize a high energy absorbing foam at the upper arm and softer foam at the lower arm. On the BioSID, a half arm extends to the bottom of the third rib. The arm structure consists of only a flat steel plate molded in foam with an upper arm load cell. Similarly, on the WorldSID, the half arm consists of a polymer bone  molded inside PVC skin filled with polyurethane elastomer with upper arm and forearm force load cells and accelerometers. The shoulder-arm joint can adjust its arm friction through three screws.

Legs

Similar to the pelvis and lumbar spine construction, there is little public documentation on the SID and SID-HIII leg construction but expected to be largely based on Hybrid-II and Hybrid-III. The legs of the EuroSID-1 utilize a standard FMVSS 208 part number 572 unit, which consists of a metal skeleton covered in plastic. EuroSID-2/ES-2 and ES-2re use the legs of the Hybrid-II dummy, but with biofidelic improvements to the mass distribution in the rigid bone structure and soft tissue structures. The EuroSID-2/ES-2 also saw the first introduction of the femur load cell to the EuroSID series. The BioSID leg has been mostly reused from the Hybrid-III dummy, but with minor adjustments to the weight distributions of the upper thigh. The WorldSID upper leg is instrumented with three axis load cells at the femoral neck and mid femur. The knee joint is instrumented with lateral load cells and accelerometers, while the lower leg is derived from the Hybrid-III construction.

Comparison of Instrumentation across 50th Percentile Side Impact ATDs

A table summarizing the instrumentation capabilities of several aforementioned ATDs is shown below.

Table 1: Comparison of Side Impact Dummy Instrumentation (adapted)

Measurement	SID	SID-HIII	BioSID	ES-2 / ES-2re	WorldSID
Head
Accel. (Ax​,Ay​,Az​)	Yes	Yes	Yes	Yes	Yes
Neck
Head/C1 (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	No	Yes	Yes	Yes	Yes
C7/T1 (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	No	No	Yes	Yes	Yes
Shoulder
Force (Fx​,Fy​,Fz​)	No	No	Yes	Yes	Yes
Deflection (δy​)	No	No	Yes	No	Yes
Upper extremities
Upper arm accel. (Ax​,Ay​,Az​)	No	No	Yes	No	Yes
Upper arm (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	No	No	No	No	Yes
Forearm accel. (Ax​,Ay​,Az​)	No	No	No	No	Yes
Forearm (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	No	No	No	No	Yes
Thorax
Spine accel. (Ax​,Ay​,Az​)	No	Yes	Yes	Yes	Yes
Rib defl. (δy​)	No	No	Yes	Yes	Yes
Rib accel. (Ay​)	No	Yes	Yes	Yes	Yes
Back plate force (Fx​,Fy​,Mx​,My​)	No	No	No	Yes	No
Abdomen
Force (Fy​)	No	No	No	Yes	No
Rib defl. (δy​)	No	No	Yes	No	Yes
Lumbar (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	Yes	Yes	Yes	Yes	Yes
Pelvis
Accel. (Ax​,Ay​,Az​)	Yes	Yes	Yes	Yes	Yes
Iliac wing (Fy​)	No	No	Yes	No	Yes
Sacrum (Fy​)	No	No	Yes	No	No
Acetabulum (Fy​)	No	No	Yes	No	Yes
Pubic (Fy​)	No	No	Yes	Yes	Yes
Lower extremities
Femoral neck (Fx​,Fy​,Fz​)	No	No	No	No	Yes
Femur (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	Yes	Yes	Yes	Yes	Yes
Tibia (Fx​,Fy​,Fz​,Mx​,My​,Mz​)	No	No	Yes	No	Yes

Biofidelity Assessments

Biofidelity refers to how accurately an anthropomorphic test device (ATD) replicates the response of a real human body during a crash. It includes how the joints move, and the forces and accelerations that the on-board sensors read. A biofidelic ATD is desirable since it increases the reliability and realism of crash test results. This increases the accuracy of injury prediction.

The measurement of biofidelity has evolved over time. The first methods involved qualitative observations of the criteria mentioned earlier. This was subjective and impossible to standardize. An early framework for standardizing biofidelity was published in 1989 by the International Organization for Standardization. Today, the Biofidelity Ranking System (BRS), also called BioRank, is used by the National Highway Traffic Safety Administration (NHTSA). Regardless of the method, biofidelity is split in two categories. External biofidelity, also called environmental biofidelity, is concerned with the overall motion and interaction of the human body with the environment during an impact. Internal biofidelity, also called injury prediction biofidelity, is concerned with the forces and deflections measured by the ATD. Furthermore, a biofidelity score can be given to the anatomical regions of the ATD individually and also as an overall score. Under the BRS system, the lower the score, the higher the biofidelity. This is not to be confused with the ISO and correlation and analysis (CORA) systems where a higher score indicates higher biofidelity. For the BRS system, the score represents the number of standard deviations the mean ATD response is from the mean cadaveric response, this score is calculated as shown using the below formula. A BioRank score of less than 2 is good, less than 1 is excellent.

$BioRankScore=\left|{\frac {\mu _{ATD}-\mu _{PMHS}}{\sigma _{PMHS}}}\right|$

$Mean:\mu$

$StandardDeviation:\sigma$

*Note that the above is a simplified version of the BioRank score calculation which illustrates the methodology

This method of scoring gives a very good representation of the biofidelity of an ATD. Dividing by $\sigma _{PMHS}$ means that, for example, if the mean ATD response is far away from the mean PMHS response in a test where the PMHS variance is high, then the ATD’s biofidelity score will not be impacted as much. Conversely, in a test where the PMHS variance is low, the ATD’s biofidelity score will be impacted more.

The original BioRank system was published in 2002. Some criticism of the ranking system at that time was that the “Test Condition Weights” were arbitrary. In 2009, the BioRank system was updated so that all tests used for evaluating ATD biofidelity are of equal value.

Rhule et al., 2009 and Kim et al., 2016 compare the biofidelity of WorldSID and ES-2re. Rhule et al., 2009 using the BioRank method and Kim et al., 2016 using the CORA method. In both studies, the WorldSID ATD presented better biofidelity than the ES-2re ATD. Although the development of front impact ATDs predates that of side impact ATDs, the WorldSID ATD shows good levels of biofidelity; 1.4 (Overall Internal), 2.2 (Overall External).

Kim et al., 2016, which used the CORA method, states that pre-impact deviation of the PMHS’s spine from the upright posture introduced phase error in responses. This caused the test data to become un-suitable for use with the BioRank system. For future tests, a potentially beneficial research area is to compare the biofidelity scores achieved on the same test between the BioRank and CORA methods. Such a study could identify strengths and weaknesses of each method as well as find any convergence areas between the two methods.

Quantitative Comparisons Between Side Impact ATDs

Side-impact ATDs are critical for evaluating occupant protection in lateral collisions. Commonly used dummies include SID-50th, BIOSID, EUROSID, WorldSID, and ES-2re, each with differing levels of instrumentation and joint design. Across several studies, side impact ATDs, including BIOSID, EUROSID, ES-2re (the EUROSID modification) and WorldSID, have been quantitatively assessed using sled tests, pendulum impacts, and full-vehicle crash scenarios to evaluate their biofidelity.

WorldSID vs. ES-2re

WorldSID demonstrates high biofidelity while ES-2re is the most widely used regulatory dummy for side impact testing. As a result, these two models are commonly compared to each other in side impact studies. Rhule et al. used the updated Biofidelity Ranking System (BRS) to compare WorldSID and ES-2re, processing ATD data identically to human subject data, including filtering, time-zero alignment, and subsampling. Drop tests, pendulum tests, sled tests, and shoulder and thorax impactor tests were performed on both ATD models. Using the BRS, lower scores denote higher biofidelity in body regions.

The WorldSID consistently achieved better external biofidelity ranks across most body regions. Both dummies had similar thorax external scores, reflecting limitations in replicating human thoracic timing, especially in sled tests such as NHTSA’s rigid low-speed conditions. WorldSID exhibited better shoulder and pelvis responses, while ES-2re tended to respond earlier in the thorax and displayed stiffer pelvis behavior.

Internal biofidelity rankings also favored WorldSID. Excluding the abdomen, WorldSID achieved a 1.2 overall biofidelity rank versus 1.7 for ES-2re. Including the abdomen, which could not be measured for ES-2re due to instrumentation limitations, WorldSID scored 1.4. WorldSID demonstrated particularly strong neck and shoulder biofidelity, while its abdominal response remained moderate at 2.4. Head internal responses were similar for both dummies. Overall, WorldSID demonstrated higher biofidelity both externally and internally, than ES-2re.

Table 2: External and internal biofidelity ranks for WorldSID and ES-2re (BRS Values)

Paper	Body Region	External Biofidelity		Internal Biofidelity
		WorldSID	ES-2re	WorldSID	ES-2re
Rhule et al., 2009	Head	-	-	0.3	1.0
	Neck	-	-	0.8	2.2
	Shoulder	1.0	2.1	0.9	1.3
	Thorax	3.2	3.1	2.0	2.4
	Abdomen	1.9	2.7	2.4	n/a
	Pelvis	2.7	3.5	1.8	1.5
	Overall Average (with Abdomen)	2.2	2.8	1.4	-
	Overall Average (without Abdomen)	-	-	1.2	1.7

Sutterfield et al. further examined the ES-2re against the WorldSID in component sled tests, pendulum tests, and drop tests.

The ISO Biofidelity Classification score quantified the biofidelity of the ATDs in the study. Developed based on the corridors from ISO 9790, the ISO classification ranks the biofidelity of an ATD based on the weighted sum of the biofidelity scores assigned to a specific body region. The body regions given different weightings include: head, neck, shoulder, thorax, abdomen, and pelvis. Each body region receives its biofidelity score based on the weight sum of a given test and the corresponding test response. The test response receives a score of 10, 5, or 0, depending on whether the response meets its requirements, falls outside of the requirement but lies within one corridor width of the requirement, or falls outside the requirement by more than one corridor width, respectively. The ISO classification categories are: excellent, good, fair, marginal, and unacceptable.

Table 3: ISO Biofidelity Classifications

Paper	Category	Score Range
Scherer et al., 2009	Excellent	> 8.6 to 10.0
	Good	> 6.5 to 8.6
	Fair	> 4.4 to 6.5
	Marginal	> 2.6 to 4.4
	Unacceptable	0 to 2.6

The findings of Sutterfield et al., corroborate Rhule et al., showing that WorldSID better replicates human response in all body regions tested.

Table 4: Biofidelity Comparison to WorldSID (ISO 9790 Values)

Paper	Body Region	ES-2re	WorldSID (Production) (NHTSA, 2005)
Sutterfield et al., 2005	Head	5.0	10
	Neck	4.2	5.6
	Shoulder	4.5	7.1
	Thorax	4.2	8.4
	Abdomen	3.8	7.8
	Pelvis	3.3	6.1
	Overall	4.1	7.6

SID-50th, BIOSID, EUROSID, WorldSIDp

Earlier side-impact dummies, SID-50th, BIOSID, and EUROSID, and a prototype of WorldSID (WorldSIDp) have also been evaluated in sled and pendulum tests, providing a baseline for improvements in designs.

BIOSID and EUROSID showed improved thoracic and abdominal instrumentation over SID-50th, which exhibits higher thorax stiffness and limited measurement capability. EUROSID provided better shoulder and thorax responses relative to SID and BIOSID, making it more suitable for European regulatory testing. Thoracic forces in the WorldSIDp was closer to the PMHS corridors than forces from the other dummies. However, the WorldSIDp did not perform as well as other dummies in pelvic forces under all impacting conditions.

WorldSID, BioSID, ES-2, EUROSID-1, ES-2re

Scherer et al. also examined the biofidelity of WorldSID against the BioSID, ES-r, EUROSID-1, and ES-2re. Scoring is based on the ISO 9790 values and testing methods. The table below displays the average results of the WorldSID in a sled test conducted at 6.8 m/s, testing the response of the thorax in the fifth test out of five targeted tests to the thorax.

Table 5: Thorax test 5 - 6.8 m/s sled test results

Paper	Measure	Lower bound	Upper bound	Run		Rating
				1 to 13	Avg
Scherer et al., 2009	Thorax plate force (kN)	Plot	Plot	Plot	Plot	6.4
	Rating			All tests 10	10.0
	Peak upper spine lateral acceleration (G)	82	122		39
	Rating			2 tests 5, 11 tests 0	0.8
	Peak lower spine lateral acceleration (G)	71	107		52.0
	Rating			All tests 5	5.0
	Peak lateral acceleration impacted rib  (G)	64	100		86
	Rating			All tests 10	10

The lower and upper bounds are obtained through cadaver testing. The average of the test results are shown along with the ratings associated with the test, depending on whether the response fell within the corridors, within one width of the corridors, or outside one width. The weight of a given test and the test response is used to calculate the weighted sum for the body region.

Similar to Yoganandan et al., the EUROSID displayed good biofidelity in the shoulder and thorax, while WorldSID displayed good biofidelity in the thorax. WorldSID also displayed a more biofidelic response in the pelvis, contradicting the findings from Yoganandan et al. This may be due to the refinement of the WorldSID from the prototype tested in 2005.

Table 6: Mid Male Side Impact Dummy Biofidelity Comparison

		Biofidelity Rating
Paper	Model	Head	Neck	Shoulder	Thorax	Abdomen	Pelvis	Overall
Scherer et al., 2009	WorldSID production version	10	5.3	10	8.2	9.3	5.1	8
	BioSID	6.7	6.7	7.3	6.3	3.8	4	5.7
	ES-2	5	4.4	5.3	5.2	2.6	5.3	4.6
	EuroSID-1	5	7.8	7.3	5.4	0.9	1.5	4.4
	ES-2re	5	4.2	4.5	4	4.1	3.2	4.2

Overall Trends

When viewed across studies, WorldSID, achieves the highest overall correlation with human responses, particularly in thorax, shoulder, and head regions, followed by BIOSID, ES-2re, EUROSID, and SID-50th. Differences between ATD performance are most pronounced in thoracic deflection timing, abdominal instrumentation, and shoulder kinematics, highlighting the importance of multi-axis response measurement.

Future Development & Limitations

The development of side impact ATDs such as BIOSID, EUROSID, and WorldSID, has greatly helped us learn more about injuries and develop systems to mitigate these injuries. However, these ATDs have some limitations and controversies, thereby impacting our understanding of how the human body reacts in response to side impact injuries.

Strengths & Limitations in Literature

With each new ATD generation, there has been an improvement in design, biofidelity, and instrumentation methods. For example, early side impact dummies used stiffer thoracic components and had limited rib articulation, resulting in chest deflections that were quite different from cadaver testing. To improve this, BIOSID introduced softer, damped rib components along with a flexible spine box. This was further improved in EUROSID by using individual rib modules that had advanced measurement capabilities. The largest improvement came in WorldSID, using multi-segment ribs, a more anatomically accurate spine, and high-resolution sensors to better detect rib deformation and analyze spine kinematics across a range of impact speeds. These developments demonstrate an upward trend of improving biofidelity and measurement techniques.

However, despite these improvements, biofidelity across different body parts and test subjects remains inconsistent, decreasing the reliability of the data we gather from ATDs. When comparing the test results between WorldSID and cadavers, the results were found to be different. For example, at high speeds (3m/s) WorldSID exhibited external chest deflections twice as high as at low speeds (1m/s), this pattern was not observed in cadavers. Furthermore, WorldSID’s chest deflections at high speed were comparable to cadavers impacted at low speeds. Additionally, the 5th percentile WorldSID female dummy had a non-biofidelic torso mass distribution and pelvic response during oblique impacts. Comparing child ATDs and pediatric volunteers also shows a discrepancy in data, where the ATD overestimates shoulder and cervical/thoracic spine stiffness.

It is difficult to make an ATD that is completely biofidelic as these devices need to be durable since they are costly to make, and the basis for biofidelity relies on validation against cadavers, which do not exhibit live tissue and organ responses of a living person. Additionally, most cadavers are elderly adults with lower muscle mass and bone density than the average person, and they cannot replicate bracing, muscle tensing, or other protective reactions someone might make during a vehicle crash. However, as much as possible, optimizing the biofidelity of ATDs should be a priority, as this challenge limits the accuracy of data collection. This can lead to inaccurate injury metrics/criteria, thereby impacting the design of injury prevention methods and safety regulation thresholds, and putting the population at risk.

Another limitation in side impact ATDs is the inability to measure internal soft-tissue strain or organ injuries. This limits occupant safety in real-life crashes because internal organ injury mechanisms can differ from the injury mechanisms ATDs are designed to undergo and measure. As such, safety systems may not be designed to properly prevent this type of injury. To supplement this, finite element modeling (FEM) has been used instead. However, FEM often relies on simplified material properties and boundary conditions, meaning it is unable to perfectly simulate soft-tissue strain. Additionally, cross-validation between FEM and cadaver data remains inconsistent.

Major Controversy

Due to population bias, most regulatory bodies only require the 50th percentile male dummy to be used in side impact testing, meaning that vehicle safety regulations are built around male-based injury risks and mechanisms. Although the 5th percentile WorldSID female dummy exists, it is usually not mandated for testing, despite differences in injury tolerance between males and females. Furthermore, this female dummy is merely a geometrically scaled-down version of the male dummy and was not developed through biomechanical characterization. This is a major controversy as efforts and resources could have been dedicated to developing a truly representative female dummy, but this has not be done.

Since males and females have different bone structures, bone densities, muscle masses, and seating postures, the biofidelic aspects of a female ATD are not accurately captured by a male ATD. Similar logic applies when comparing the 50th percentile male to people of different ages such as children or the elderly. This underrepresentation of women, children, and the elderly is a large concern as the ATD test results inform injury risks and mechanisms, which in turn impact vehicle safety regulations and safety system designs. Inaccurate data collected for these populations has led to higher risk of injury and lower protection during vehicle crashes. While the 50th percentile male dummy remains a valuable ATD due to undergoing extensive testing and validation, relying purely on this male-based ATD to predict injury risks and mechanisms, and to influence vehicle safety for other populations is highly problematic.

Future Research Priorities

Future research should address the previously mentioned limitations and controversy regarding side impact ATDs to improve the safety of the general population and reduce risk of injury. These research priorities include:

1. Biofidelity: better balancing of biofidelity with durability, including more external materials that mimic soft-tissue deformation, validating ATDs against younger cadavers, animals, or volunteers
2. FEM improvements: standardizing calibration for FEM based on cadaver testing, incorporating more complex materials to increase modeling accuracy
3. ATD diversity: develop more percentile dummies for women, children, and the elderly, focusing on biomechanical characterization instead of just scaling
4. Regulation diversity/implementation: including more diverse regulations that ensure female, child, and perhaps even elderly ATDs are required during testing