ISO/TR 21934-1:2021

TECHNICAL ISO/TR
REPORT 21934-1
First edition
2021-06
Road vehicles — Prospective safety
performance assessment of pre-crash
technology by virtual simulation —
Part 1:
State-of-the-art and general method
overview
Véhicules routiers — Evaluation prospective de la performance
sécuritaire des systèmes de pré-accident par simulation numérique —
Partie 1: Etat de l’art et aperçu des méthodes générales
Reference number
©
ISO 2021
© ISO 2021
All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication may
be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting
on the internet or an intranet, without prior written permission. Permission can be requested from either ISO at the address
below or ISO’s member body in the country of the requester.
ISO copyright office
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: +41 22 749 01 11
Email: copyright@iso.org
Website: www.iso.org
Published in Switzerland
ii © ISO 2021 – All rights reserved

Contents Page
Foreword .iv
Introduction .v
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviated terms . 3
4.1 Symbols . 3
4.2 Abbreviated terms . 3
5 Evaluation objective and baseline of assessment . 4
5.1 Definition of the evaluation objective . 4
5.2 Establishment of baseline . 6
6 Input data . 7
6.1 General . 7
6.2 Active safety technology related data. 8
6.3 Accident data . 9
6.4 Data from naturalistic driving studies and field operation test.10
6.5 Infrastructure and traffic data .11
6.6 Data from tests in controlled environments.11
7 Implementation of virtual simulation .11
7.1 General .11
7.2 Simulation framework .11
7.3 Simulation tool .12
7.4 Simulation models .12
7.4.1 Vehicle model .12
7.4.2 Safety technology model .13
7.4.3 Environment model . . .14
7.4.4 Traffic situation model.14
7.4.5 Traffic model .15
7.4.6 Driver model .16
7.4.7 Collision model .16
7.5 Simulation control.17
8 Estimating safety technology safety performance .18
9 Validation and verification .20
10 Practical experience .23
10.1 General .23
10.2 Establishment of baseline .23
10.3 Simulation framework .24
10.4 Comparative study of different simulation tools .24
10.5 Estimating the safety performance .25
10.6 Validation and verification .25
11 Conclusions and limitations .25
12 Outlook .27
12.1 General .27
12.2 Automated driving .27
12.3 V2X technologies .28
Annex A (informative) List of tools .30
Annex B (informative) Input and output of simulation models .31
Bibliography .36
Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards
bodies (ISO member bodies). The work of preparing International Standards is normally carried out
through ISO technical committees. Each member body interested in a subject for which a technical
committee has been established has the right to be represented on that committee. International
organizations, governmental and non-governmental, in liaison with ISO, also take part in the work.
ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of
electrotechnical standardization.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular, the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www .iso .org/ directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of
patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of
any patent rights identified during the development of the document will be in the Introduction and/or
on the ISO list of patent declarations received (see www .iso .org/ patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation of the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO's adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT), see www .iso .org/
iso/ foreword .html.
This document was prepared by Technical Committee ISO/TC 22 Road vehicles, Subcommittee SC 36
Safety and impact testing.
A list of all parts in the ISO 21934 series can be found on the ISO website.
Any feedback or questions on this document should be directed to the user’s national standards body. A
complete listing of these bodies can be found at www .iso .org/ members .html.
iv © ISO 2021 – All rights reserved

Introduction
Different Active Safety and Advanced Driver Assistance Systems (ADAS), in the following both referred
to as active safety technology, have been developed and introduced into the market. The question that
goes along with the development and introduction is, what impact these technologies have on road
traffic and more specifically, to what extent these technologies prevent crashes and injuries. Such
questions are of relevance for different stakeholders, such as vehicle manufacturers and suppliers, road
authorities, research organisations and academia, politics, insurance companies as well as consumer
[1]
organisations.
The answers to these questions are derived from assessment of such technologies in terms of road
traffic safety. Different assessment methodologies have been developed in the past and are being
[2]
used today. In general, the utilized methodologies can be divided in two types of assessment. The
first type determines the technology’s safety effect after its market introduction. Typically, in this
assessment type accident statistics are analysed in order to determine the difference between the
[1]
accident situation with the technology compared to a control group without the technology. These
methods are called retrospective assessment methods. A precondition for these methods is that the
technology under assessment has reached a sufficient penetration rate in the market and that sufficient
accident cases with and without the technology are recorded for a comparison. The penetration rate
does not necessarily need to be related to the whole vehicle fleet, but can also be related to a certain
[3]–[5]
vehicle subgroup or class. On the other hand, there are methods that predict the technology's
[6][7]
effect on traffic in relation to traffic safety before its market introduction. These methods are
called prospective methods using different approaches and tools.
This document focuses on the prospective assessment of traffic safety for vehicle-integrated
technologies acting in the pre-crash phase by means of virtual simulation.
The safety performance of a technology is determined by means of comparing data from the baseline
and treatment simulations based on a certain metric. The baseline for the assessment is the situation
without the vehicle-integrate technology under assessment present. The virtual simulation with the
technology is called treatment simulation.
The described assessment is limited to “vehicle-integrated” technology and does not consider
technologies operating off-board. The virtual simulation method per se is not limited to a certain vehicle
type. Although the main focus is often on passenger cars, the method is also applicable to motorised
two-wheelers as well as heavy goods vehicles. Furthermore, the assessment approach discussed in this
document focuses rather on accident avoidance and the technology’s contribution to the mitigation of
the consequences. Safety technologies that act in the in-crash or the post-crash phase are not explicitly
addressed by the method, although the output from prospective assessments of crash avoidance
technologies can be considered as an important input to determine the consequences. The extension
of the method to technologies, such as automated driving and V2X based technologies, are discussed in
the outlook at the end of this document.
In general, the assessment of active safety technologies requires the consideration of interaction with
surrounding traffic as well as the host vehicle driver. These interactions increase the complexity
of the assessment due to the high number of resulting variables. Consequently, for a comprehensive
assessment, the technology’s safety performance is analysed in a high number of test scenarios, in order
to cover all relevant circumstances that affect the critical situation and crashes. The virtual simulation
approach allows for running large numbers of test scenarios while offering a promising combination
of safety performance, flexibility, reproducibility, and experimental control. The need for using virtual
simulations in the prospective assessment of safety technologies is generally recognized. However,
standardized terminology and processes of methodological aspects to perform such assessments are
[1]
not available to date, which makes results hardly comparable. For this reason, automotive industry,
1)
research institutes, and academia joined in the P.E.A.R.S. (Prospective Effectiveness Assessment
for Road Safety) initiative with the objective to develop a comprehensible, reliable, transparent, and
accepted methodology for quantitative assessment of crash avoidance technology by virtual simulation.
[1]
This document aims to provide an overview on the state-of-the-art in the prospective assessment
of road safety for vehicle-integrated (active) safety technologies by means of virtual simulation, see
Figure 1.
After the introductive Clauses 1 to 4, the general method for a prospective assessment study is
described in Clause 5, where special attention is given to the definition of the traffic safety evaluation
scope and the establishment of the baseline. Clause 6 describes various data that can be used as input
for different tasks within the assessment procedure. Then a general virtual simulation framework and
various simulation models needed for conducting the simulation are presented in Clause 7, followed
by a description of the approaches to quantify the derived safety effect in Clause 8. A description of
validation and verification aspects as well as an overview on tools are given in Clause 9. Clause 10 of
the document provides a practical example of a comparative study of different simulation tools and
discusses the lessons learned. Clause 11 provides conclusions as well as describes limitations for the
state-of-the-art methods. Clause 12 provides an outlook towards the prospective safety performance
assessment for automated driving as well as the follow up to the current document.
Figure 1 — Overview of the process of prospective assessment of traffic safety for vehicle-
integrated safety technologies by means of virtual simulation and the structure of this
document
1) P.E.A.R.S. is an open consortium (established in 2012) in which engineers and researchers from the automotive
industry, research institutes and academia join with the objective to develop a comprehensible, reliable, transparent
and accepted methodology for quantitative assessment of crash avoidance technology by virtual simulation. Partners
of P.E.A.R.S. are (status Sep. 2020): Automotive Safety Technologies, AZT Automotive, BMW Group, Federal Highway
Research Institute (BASt), Chalmers University of Technology, Continental, Denso, Fraunhofer IVI, Generali, RWTH
Aachen University (ika), LAB, Swiss Re, TH Ingolstadt, Technical University Dresden, Technical University Graz, TNO,
Toyota, Technical University Dresden, TÜV Süd, University Leeds, UTAC CERAM, Virtual Vehicle, Volkswagen, Volvo
Cars, VUFO, ZF. More information at https:// pearsinitiative .com/ .
vi © ISO 2021 – All rights reserved

TECHNICAL REPORT ISO/TR 21934-1:2021(E)
Road vehicles — Prospective safety performance
assessment of pre-crash technology by virtual
simulation —
Part 1:
State-of-the-art and general method overview
1 Scope
This document describes the state-of-the-art of prospective methods for assessing the safety
performance of vehicle-integrated active safety technologies by virtual simulation. The document
describes how prospective assessment of vehicle-integrated technologies provides a prediction on
how advanced vehicle safety technology will perform on the roads in real traffic. The focus is on the
assessment of the technology as whole and not of single components of the technology (e.g. sensors).
The described assessment approach is limited to “vehicle-integrated” technology and does not consider
technologies operating off-board. The virtual simulation method per se is not limited to a certain
vehicle type. The assessment approach discussed in this document focuses accident avoidance and the
technology’s contribution to the mitigation of the consequences. Safety technologies that act in the in-
crash or the post-crash phase are not explicitly addressed by the method, although the output from
prospective assessments of crash avoidance technologies can be considered as an important input to
determine the overall consequences of a crash.
The method is intended as an overall reference for safety performance assessment studies of pre-crash
technologies by virtual simulation. The method can be applied at all stages of technology development
and in assessment after the market introduction, in which a wide range of stakeholders (manufactures,
insurer, governmental organisation, consumer rating organisation) could apply the method.
2 Normative references
The following documents, in whole or in part, are normatively referenced in this document and are
indispensable for its application. For dated references, only the edition cited applies. For undated
references, the latest edition of the referenced document (including any amendments) applies.
ISO 12353-1, Road vehicles — Traffic accident analysis — Part 1: Vocabulary
3 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO 12353-1 and the following
apply.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— ISO Online browsing platform: available at https:// www .iso .org/ obp
— IEC Electropedia: available at https:// www .electropedia .org/
3.1
levels of automation
levels that primarily identify how the “dynamic driving task” is divided between human and machine
Note 1 to entry: See Reference [8].
3.2
baseline
initial set of data to which the performance of the technology under study is compared when performing
prospective assessments (3.7) of the technologies' performance
Note 1 to entry: This concept also complements treatment (3.13).
3.3
cooperative
applications based on vehicle-to-vehicle, vehicle-to-VRU and vehicle-to-infrastructure communication
3.4
host vehicle
vehicle, which is subject for assessment, i.e. is equipped with the technology in the treatment simulation
3.5
injury risk function
description of the probability of an injury in relation to crash attributes
Note 1 to entry: The most frequently used injury risk functions describe the probability of an injury occurrence
in relation to crash severity, e.g. impact speed or change of velocity.
3.6
projection
indicates what the future changes in a population would be if the assumptions (often based on patterns
of change which have previously occurred) about future trends actually occur
Note 1 to entry: Population projections – in the sense of Reference [9] - are estimates of total size or composition
of populations in the future, see Reference [10].
3.7
prospective assessment
assessment of the performance of technologies in a predictive way
Note 1 to entry: The assessment can be done, for example, before their deployment into a vehicle population.
3.8
target population
all situations or accidents that are addressed by the function under assessment
3.9
real-world data
data collected in a non-experimental, non-virtual situation
3.10
retrospective assessment
assessment of the performance of technologies after their deployment into a vehicle population
3.11
time series
series of data points indexed (or listed or graphed) in time order
3.12
traffic situation
crash-, near-crash or normal driving situation whose description can be considered for the
establishment of the baseline (3.2)
2 © ISO 2021 – All rights reserved

3.13
treatment
use of a specific technology to affect the course of an event in a traffic situation (3.12) in order to avoid
or mitigate crashes
Note 1 to entry: Treatment simulations provide data on the performance of the technology under assessment
to compare with the baseline (3.2) data when performing prospective assessments (3.7) of performance of
technologies.
Note 2 to entry: This concept also complements baseline (3.2).
3.14
test scenario
detailed description of trajectories, geometrical relations, speeds, etc. of a traffic situation (3.12)
Note 1 to entry: See References [11]–[13].
3.15
vehicle-integrated
technology under assessment operating on-board of the vehicle
4 Symbols and abbreviated terms
4.1 Symbols
E Effectiveness / safety performance
Weighted frequency of the metric (e.g. percentage of crashes) in the simulation without
N
the technology under assessment
Weighted frequency of the metric (e.g. percentage of crashes) in the simulation with the
N’
technology under assessment
v Velocity
4.2 Abbreviated terms
ACC Adaptive Cruise Control
ADAS Advance Driver Assistance Systems
AEB Autonomous Emergency Braking
BAAC Analysis report of road accidents involving physical injury (France)
BASt Federal Highway Research Institute (Bundesanstalt für Straßenwesen)
CEDATU Central Database for In-Depth Accident Studies (Austria)
CIDAS China In-Depth Accident Study
EES Energy Equivalent Speed
ETAC European Truck Accident Causation
FESTA Field opErational teSts supporT Action
FOT Field Operation Test
GIDAS German In-Depth Accident Study
HIL Hardware-in-the-loop
IEC International Electrotechnical Commission
IIHS Insurance Institute for Highway Safety
IGLAD Initiative of Global Harmonisation of Accident Databases
ISO International Organization for Standardization
ITARDA Institute for Traffic Accent Research and Data Analysis
J-TAD Japan Traffic Accidents Databases
KBA German Federal Motor Transport Authority (Kraftfahrtbundesamt)
LDW Lane Departure Warning System
LIDAR Light detection and ranging
MIL Model-in-the-loop
NASS National Automotive Sampling System
NDS Naturalistic Driving Studies
RAIDS Road Accident In Depth Studies
P.E.A.R.S. Prospective Effectiveness Assessment for Road Safety
PTW Powered Two Wheelers
RASSI Road Accident Sampling System - India
SCP (cr/cl) Straight Crossing Paths (cyclist from the right / cyclist from the left)
SIL Software–in-the-loop
TTC Time to collision
V2X Vehicle to X (Vehicle and / or Infrastructure) Communication
VIN Vehicle identification number
VRU Vulnerable Road User
V&V Validation and Verification
5 Evaluation objective and baseline of assessment
5.1 Definition of the evaluation objective
Since there are numerous objectives to conduct prospective safety performance assessments, it
is important that a precise research question for the assessment is formulated. Then by identifying
relevant traffic situations – the target population - to address the research question, a more precise
[14]
specification and application for a virtual simulation study is provided. Figure 2 shows the place in
the process overview.
4 © ISO 2021 – All rights reserved

Figure 2 — Overview of the process — Definition of the evaluation objective
[1]
Various objectives to conduct safety performance assessments have been identified, the main ones
are:
— quantification of effects (positive and negative) of a certain technology in terms of traffic safety;
— prioritization and optimization of safety technologies during research and development;
— identification of business opportunities and anticipation of regulations and consumer testing.
Furthermore, two types of processes are used to formulate the target for this kind of studies.
— A technology-driven process in which a request is put forward to estimate the safety benefit of a
safety technology. This technology can be more or less defined at the time of the study; it can be an
idea, a concept, a product under development or a product that already has been implemented but
not introduced into the market (also often called a bottom-up approach).
— A traffic safety-driven process in which existing or expected safety problems or certain relevant
traffic situations are identified. In this case, the target for the study is not linked to a particular
safety technology but to a targeted lack of safety (also often called top-down approach).
Hence, it is important to note that if results between different studies are compared the research
question needs to be a) accessible and b) precisely formulated. This requires to rephrase the question
asking additional information such as: “What type of safety technology will be evaluated?”, “What data
segments will be addressed (pre-impact situation, traffic participants, type of road, etc.)?”, “What time
horizon is being considered?”, “Should the installation rate of an optionally equipped safety technology
in the vehicle fleet be considered?”, “What metric is suggested for the safety effect?”, “What is the
expected accuracy of the result?”, “What could change the consequence on the road, if the cars were
equipped with new safety technologies?”.
An adequate example of a properly formulated research question is: What is the relative change in car-
to-cyclist crashes due to an autonomous emergency braking (AEB) system with 100 % penetration rate
in a specific car in urban car-to-cyclist situations in Germany in two years from now?
Once the research question is set, relevant traffic situations for virtual simulation can be identified, for
example the definition of a target population for the study. Relevant traffic situations can be derived
by, e.g. analysis of retrospective crash data, naturalistic driving studies, and knowledge gathered
during technology development. The outcome of the identification process is an overall description and
quantification of the traffic situations and the involved traffic participants of the simulation. When it
comes to analysis of real-world crash or near-crash data, various types of classification schemes can
be used to set boundaries for the study. Especially important aspect is pre-impact relative movement
of involved traffic participants before a crash or near-crash. One example is Straight Crossing Path
scenarios (from right: SCPcr / from left: SCPcl), where the car was moving forward, and the cyclist was
crossing the path either from left or right, see Figure 3. The pre-impact situation is often accompanied
by pre-crash-factors that include parameters that may have influenced the course of events before the
crash. Examples are speed-related measures, driver status, and traffic environment related factors such
as light condition, road layout, and road status. In addition, the crash configuration can be of interest,
e.g. the impact point and direction.
[15][16]
Figure 3 — Example from a pre-impact situation classification scheme
To summarize, for a fictitious version of an AEB system addressing the example research question
above, the target population could be expressed as; SCPcr and SCPcl situations during daylight on roads
with lane markings and where the driver is visually distracted.
At this stage it important to mention that the metrics to be used for estimating technology’s safety
performance consider the potential impacts and the required input data. After establishing a baseline
according to the target population (see 5.2), the outcome of simulations with and without the safety
technology will be compared by this certain metric. Details with respect to the topic “metric” are
presented in Clause 8.
5.2 Establishment of baseline
When target traffic situations are identified which address the research question, a detailed,
measurable definition of these situations for the upcoming virtual simulations is provided, i.e. the
baseline. In general, the prospective safety performance assessment conducts a comparison between
traffic situations without and with the technology under assessment. Thus, the baseline refers to the
situation without the technology under assessment present. This includes traffic situations that are
needed to evaluate both positive and negative performance, according to the evaluation objective. The
establishment of the baseline defines the reference to be used in the upcoming simulations and a real-
world reference is essential. Figure 4 shows the place in the process overview.
Figure 4 — Overview of the process — Establishment of the baseline
Three main approaches are distinguished where the cases in the baseline are generated in different
ways:
— baseline with original cases of real-world traffic situations,
— baseline with modified cases of real-world traffic situations,
— baseline with synthetic cases based on relevant characteristics of real-world traffic situations.
Below are explanation of each respective baseline.
— Baseline with original cases of real-world traffic situations
In a straightforward application, the baseline corresponds to real-world traffic situations that have been
reconstructed from crash data or other sources such as NDS/FOT datasets. The cases are represented
6 © ISO 2021 – All rights reserved

according to parameters found in the corresponding database (e.g. collision speed and collision angle).
[17]-[21]
Furthermore, the crash database parameters can be used in a model to perform a reconstruction of
the cases, thus simulation is used to recreate real accidents in order to have a detailed, numerical time
series description of the cases in the baseline. An example of this approach is the German In-Depth
[22]
Accident Study (GIDAS) based Pre-Crash-Matrix (PCM). Typical parameters needed in the PCM
database are vehicle trajectories and speed related measures, crash configurations, sight obstructions,
information on the traffic environment and driver behaviour.
— Baseline with modified cases of real-world traffic situations
As crashes reported in the database reflect the actual crash, with possibly rather old vehicles,
replications of the traffic situations with a modern vehicle can be performed, i.e. a re-simulation to
[20][23][24]
establish a baseline with more recent properties of the vehicles involved.
Another challenge in crash databases is the limited information on pre-crash parameters, for example
vehicle trajectories and driver behaviour such as inattention or drowsiness that can be influenced by a
safety technology. The use of recorded crashes such as in naturalistic driving studies or usage of event
data recorder data can enable more qualitative estimations when available.
If the crash sample does not provide a sufficient representation of the traffic situation identified based
on the research question, sampling techniques can be used to create random, synthetic cases based
[25]
on marginal distributions of event related variables. However, in contrast to the next approach,
presented below, the synthetic created cases still reflect the original traffic situation.
— Baseline with synthetic cases based on relevant characteristics of real-world traffic situations
Cases for a baseline can also be generated based on the understanding of contributing factors involved
[26]–[28]
in the targeted traffic situations; the crash mechanisms.
Once these mechanisms are revealed, the situation is modelled using distributions of selected
parameters. Sampling methods, for example Monte Carlo simulations, can be used to vary the
characteristics of the cases in the baseline, such as driver reaction/response as well as vehicle
[29]
properties, vehicle trajectories, and traffic and environmental variables. When the simulations
for generating situations are performed, only a portion of the cases in the baseline might end up in a
collision. The baseline then consists, besides cases where a collision occurs, also of cases without a
collision or risk of a collision. These cases can be used to investigate situations, where an activation
might not be desired or required.
The baseline is to be used in virtual simulations, with and without the safety technology present. The
complexity and the level of detail depend on the way the baseline has been represented and to which
degree the safety technology interferes, e.g. to the way that the driver, the vehicle, the surrounding
traffic etc. are modelled. The virtual simulation framework and the various models needed are
described in Clause 7.
6 Input data
6.1 General
Input data are required for different tasks within the process of assessing a technology’s safety
performance by means of virtual simulation. These tasks are:
— establishing the baseline of the simulation (see Clause 5.2);
— development, training and parametrisation of models used in the simulation tool - in particular
traffic participant (e.g. driver) behaviour models and injury risk function (see Clause 7);
— performing subsample weighting analysis and projection of simulation output (see Clause 8);
— validation and verification of the simulation as well as its models (see Clause 9).
In relation to these different tasks and with regard to the research question, the quality and
representativeness of the data sample are important and relevant aspects throughout the process.
In general, a wide range of data is necessary for prospective safety performance assessment. Although
in most cases, data from real world are used, the input data do not necessarily need to be gathered
in the real world. Verified data from previous simulations or data collected in specific tests may be
used as input data for the assessment as well. In the following, the most common relevant data sources
are presented and discussed. These sources are (details on the different sources are provided in the
sections below):
— safety technology related data;
— accident data (general and/or in-depth data);
— data from naturalistic driving studies (NDS) or field operation tests (FOT);
— infrastructure and traffic data;
— test data gained in a controlled environment, such as test track or driving simulators.
In Table 1 the typical data sources are mapped to the tasks of prospective safety performance
assessment.
Table 1 — Overview on often used data types for the different tasks within the prospective
safety performance assessment
Active safe- Accident Accident data NDS/FOT Infra-struc- Test data
ty technol- data (gen- (in-depth data ture and (test track,
ogy related eral data) data) traffic data simulator)
data
Establishing the
baseline – direct input X X X (X)
(see 5.2)
Establishing the base-
line – modified input X X X (X) (X)
(see 5.2)
Establishing the
baseline – stochasti-
X (X) X X X (X)
cally generated input
(see 5.2)
Development of models
(X) (X) X X X X
(see 7.4)
Data projection (see
X (X) (X) X
Clause 8)
Validation and verifica-
(X) (X) X X X X
tion (see Clause 9)
NOTE ‘X’ marks commonly used data sources, ‘(X)’ marks rarely used data sources.
6.2 Active safety technology related data
The purpose of the prospective safety performance assessment is to determine the safety effect of
a certain technology. To perform the assessment, specific information about the technology under
assessment is required. The information describes under which conditions (e.g. speed range and
environmental conditions) the technology operates, which conditions lead to deactivation as well as
[30]
how the technology performs its function – sensing, controlling, actuating. For an active safety
technology, the intended situation is typically a critical driving situation, such as a potential collision
with another object or an unintended road departure. The relevant information can further be split
8 © ISO 2021 – All rights reserved

into information related to the activation of the technology and information related to the behaviour of
[31]
the technology once activated (e.g. type and strength of technology intervention).
The required data are provided by a description, by a model or is derived by means of separate tests
[see further information in the subclause on data from tests in controlled environments (6.6)].
6.3 Accident data
One of the most important input data sources for the prospective safety performance assessment is
data that describe accident situations. In general, two types of accident data are available: general
accident data and in-depth data.
The general accident data describe the accident situation on macroscopic level – often on national or
international representative level. Typically, the data of such databases provide parameters like the
total number of accidents, or the number of accidents with a certain level of injury as reported by the
police. Thus, most of these databases contain the exhaustiveness of the road accidents but with very few
details. A classification of the pre-impact situation, the road type, at which the accident occurs, and/or
[31]–[33]
the involved vehicle type is mainly available but not necessarily reported. In-depth information
such as intrusions or reconstruction parameters are not reported in these databases. An overview of a
few selected databases that provide general accident data is given in Table 2.
Table 2 — Overview on selected general accident databases according to Reference [35]
Database Collected information (examples)
UNO / WHO Traffic fatalities and injuries
Nationally-representative sample of police-reported motor
GES
vehicle crashes of all types, from minor to fatal
Combining different national European statistics including
CARE parameters, e.g. person class, gender, age group, vehicle group,
collision type, lighting and weather conditions, day of the week
Crash data (e.g. fatalities injury crashes by road type, road user,
IRTAD age), exposure data (e.g. vehicle kilometres driven) and other
safety data (e.g. seatbelt waring rates)
National statistics (e.g. in Germany Federal Sta- Among other parameters traffic fatalities and injuries, the type
tistical Office of Germany or BAAC in France) of accident and VIN
Statistics on regional level (e.g. statistical offices Among other parameters traffic fatalities and injuries, more
of the German states) detailed type of accident
In-depth accident databases provide detailed information about the accident and the sequence of events
but for a limited number of road accidents. Such databases exist in different countries as indicated by
Table 3. These databases either cover specific regions of a country, the entire country, accidents for
a specific car brand or accidents with different accident severity (e.g. with material damage, injuries,
fatalities). In case only specific regions are covered by the database, the representativeness of the data
[36][37]
for the country needs to be checked. For single accidents many parameters are collected that
describe the accident sequences, the condition of the involved vehicles as well as the environmental
[38] [34]
condition (see e.g. GIDAS Codebook ). The data can either be logged by accident event recorders
[39]
or are determined b
...