ISO/DIS 23985

ISO/TC 22/SC 33

Secretariat: DIN

Date: 2025-12-03

Passenger cars — Validation of vehicle dynamics simulation — Weave test for on-centre handling quantification

Voitures particulières — Validation de la simulation de la dynamique du véhicule — Essai en sinusoïde pour la quantification du comportement routier en ligne droite

DIS stage

Warning for WD’s and CD’s

This document is not an ISO International Standard. It is distributed for review and comment. It is subject to change without notice and may not be referred to as an International Standard.

Recipients of this draft are invited to submit, with their comments, notification of any relevant patent rights of which they are aware and to provide supporting documentation.

© ISO 2026

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

E-mail: copyright@iso.org

Website: www.iso.org

Published in Switzerland

Contents

Foreword iv

Introduction v

1 Scope 1

2 Normative references 1

3 Terms and definitions 1

4 Principle 2

5 Variables 3

6 Simulation tool requirements 4

6.1 General 4

6.2 Vehicle dynamic model structure 4

6.3 Vehicle dynamic model class recommendations 5

7 Physical testing 6

7.1 Test conditions 6

7.2 Test procedure 7

7.3 Filtering of measured data 9

7.4 Data evaluation and presentation of results 9

8 Simulation 11

8.1 General 11

8.2 Driver controls 11

8.3 Data recording and processing 11

8.4 Data evaluation and presentation of results 12

9 Validation criteria 12

10 Documentation 14

Annex A (informative) Example validation results 15

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.

International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.

The main task of technical committees is to prepare International Standards. Draft International Standards adopted by the technical committees are circulated to the member bodies for voting. Publication as an International Standard requires approval by at least 75 % of the member bodies casting a vote.

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 orall such patent rights.

ISO N132 was prepared by Technical Committee ISO/TC 22, Road vehicles, Subcommittee SC 33, Vehicle dynamics, chassis components and driving automation systems testing.

Introduction

The main purpose of this International Standard is to provide a repeatable and discriminatory method for comparing simulation results to measured test data from a physical vehicle for a specific type of test.

The dynamic behaviour of a road vehicle is a very important aspect of active vehicle safety. Any given vehicle, together with its driver and the prevailing environment, constitutes a closed-loop system that is unique. The task of evaluating the dynamic behaviour is therefore very difficult since the significant interactions of these driver–vehicle–environment elements are each complex in themselves. A complete and accurate description of the behaviour of the road vehicle must necessarily involve information obtained from a number of different tests.

Since this test method quantifies only one small part of the complete vehicle handling characteristics, the validation method associated with this test can only be considered significant for a correspondingly small part of the overall dynamic behaviour.

Passenger cars — Validation of vehicle dynamics simulation — Weave test for on-centre handling quantification

This document specifies methods for comparing computer simulation results, generated from a vehicle mathematical model, with measured test data obtained according to ISO 13674-1. The purpose of this comparison is to validate the simulation tool's accuracy in predicting vehicle behaviour for this specific test, enabling its application to one and more vehicle configuration. Extending this validation to further vehicle configurations and variants is outside the scope of this document.

It is applicable to passenger cars as defined in ISO 3833.

The following documents are referred to in the text in such a way that some or all of their content constitutes requirements of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

ISO 1176, Road vehicles — Masses — Vocabulary and codes

ISO 2416, Passenger cars — Mass distribution

ISO 3833, Road vehicles — Types — Terms and definitions

ISO 8855, Road vehicles — Vehicle dynamics and road-holding ability — Vocabulary

ISO 11010-1, Passenger cars — Simulation model classification — Part 1: Vehicle dynamics

ISO 13674-1, Road vehicles — Test method for the quantification of on-centre handling — Part 1: Weave test

ISO 15037-1, Road vehicles — Vehicle dynamics test methods — Part 1: General conditions for passenger cars

For the purposes of this document, the terms and definitions given in ISO 1176, ISO 2416, ISO 3833, ISO 8855 as well as the following apply.

ISO and IEC maintain terminological databases for use in standardization at the following addresses:

• IEC Electropedia: available at http://www.electropedia.org/
• ISO Online browsing platform: available at http://www.iso.org/obp

simulation

calculation of motion variables of a vehicle from equations in a mathematical model of the vehicle system

simulation tool

simulation environment including software, model, input data, and hardware in the case of hardware in the loop simulation

vehicle under testbehavior

VUT

vehicle tested according to ISO 13674-1

characteristic value

specific numerical result calculated from test or simulation data according to a defined procedure, representing a particular aspect of vehicle behaviour or system performance in the weave test

The weave test for quantification of on-centre handling represents straight-line directional stability characteristics of a vehicle subject to low level lateral acceleration, typically below 1 m/s². On-centre handling is concerned primarily with features that directly influence the driver's steering input, such as steering system and tyre characteristics. Thus, test method for the evaluation of on-centre handling behaviour seek to minimize other factors that influence the wider aspects of straight-line directional stability, such as disturbance inputs due to ambient winds and road irregularities.

NOTE While the analysis focuses on low level lateral acceleration, typically below 1 m/s², the actual test maneuver (per Clause 7) targets a higher peak acceleration to get good data through this low-g range.

Weave test for on-centre handling defines a maneuver that involves driving the vehicle in a nominally straight line at a constant forward speed. During the test, driver inputs and vehicle responses are measured and recorded. From the recorded signals, characteristic values are calculated. For simulation purpose, the recorded driver inputs, steering wheel angle and longitudinal velocity, from physical test are applied to simulation model as inputs and corresponding vehicle responses are computed, and characteristic values are calculated.

The recorded variables from physical test and simulations are taken in pairs and plotted one against the other on Cartesian coordinates as shown in Figure 1. For each pair of variables, this produces a series of hysteresis loops laid one over another, the number of loops corresponding to the number of data cycles analyzed.

From the polynomial curve fits to the combined hysteresis loops, the following parameters are evaluated that are directly related to on-centre handling characteristics:

• ordinate dead band;
• abscissa dead band;
• gradient.

This International Standard aims to verify the predictive capability of a vehicle simulation tool by comparing its simulated vehicle behaviour to physical test results. The simulation tool is used to simulate a specific VUT undergoing the open-loop tests defined in ISO 13674-1. To establish a reliable comparison baseline, the VUT is subjected to at least four consistent test cycles:

Key

A range of polynomial curve fit

1 ordinate dead band

2 abscissa dead band

  gradient = y/2x

Figure 1 — Definition of parameters

The following variables shall be measured from physical testing. Applying the measured steering-wheel angle and longitudinal velocity as simulation input, all other variables are computed.

• Steering-wheel angle, δ_H;
• Steering-wheel angular velocity, dδ_H/dt;
• Steering-wheel torque, M_H;
• Yaw velocity, dψ/dt;
• Longitudinal velocity, v_X.
• Lateral acceleration, a_Y.

Measuring equipment including transducers and their measuring range and error tolerances, transducer installation, and data processing procedures shall be in accordance with ISO 13674-1, Clause 6.

The simulation tool used to predict behaviour of VUT shall include a mathematical model capable of calculating variables of interest (see Clause 5) for the test procedures being simulated. In this International Standard, the mathematical model is used to simulate a weave test as specified in ISO 13674-1 and provide calculated values of variables of interests.

The procedure for obtaining input data from experiments may differ for simulation tools, however, the input data shall not be manipulated for better correlation. However, adjusting input data parameters to reflect measured test conditions such as ambient temperature, road surface friction coefficient determined, etc., is permitted. Manipulation of model parameters solely to improve correlation without physical justification is not permitted.

Minor deviations between the physical test environment and the simulation, such as road cross fall, should be accounted for. These deviations can be handled through methods like offset calculations or by averaging the results of tests conducted in opposite directions.

Overall vehicle dynamic model for  weave test simulation is shown in Figure 2 (ISO 11010-1, Clause 5), where all the subsystems of vehicle and their interdependences are defined.  The vehicle dynamic model is structured in the following model classes: vehicle body (VH), powertrain (PT), brake (BR), steering (ST), suspension (SU), aerodynamics (AE) and tyre (TY) and the road with road surface (RS) and road wind (RA). Some model classes consist only of a physical system model XXM, others as combination of a physical system XXM and a control system model XXC.

Note that multiple vehicle dynamic modeling approaches can be employed for simulating weave tests for on-centre handling, depending on the specific simulation objectives and applications. Consequently, Figure 2 provides a general modeling framework for reference purposes only.

NOTE Figure 2 in this document is a corrected version of Figure 2 found in ISO 11010-1. The figure in ISO 11010-1 will be replaced with this corrected version at the next revision.

Key

signalbus (bidirectional)

actor

sensor

forces

kinematics

Figure 2 — Top-level model architecture

Recommendations for vehicle dynamic model of each system to perform  weave test simulation are summarized in Table 1. Model class numbers in Table 1 that are designated according to modeling fidelity recommendations and associated mathematical method are adopted from ISO 11010-1.

Model classes given in Table 1 specifies recommendations for models, thus more complex model class can be used depending on the purpose of simulation. 

All model classes used in the simulation shall be reported.

Table 1 — Recommended model classes for vehicle dynamic model

VUT shall be tested using test conditions specified in ISO 13674-1.

NOTE This International Standard does not define all of the details of the testing procedure. This Clause does describe the parts of the test procedure that are typically simulated.

General test conditions shall be in accordance with ISO 15037-1, Clause 6.

The test track requirements shall be in accordance with those of ISO 15037-1, Clause 6.2. In addition, the gradient of the test surface should not exceed 1 %, and the test track shall follow a straight-line path.

During a test, the ambient wind velocity shall not exceed 5 m/s when measured at a height above ground of not less than 1 m. Ideally, the maximum ambient wind velocity should not exceed 1,5 m/s. If this cannot be achieved, then conditions of significant “gusting” should be avoided, i.e. testing should be avoided in conditions where changes in wind velocity exceed a range of 1,5 m/s. In the event that the ambient velocity exceeds 1,5 m/s or the range of “gusting” exceeds 1,5 m/s, or both, the vehicle should be tested in a direction such that the ambient wind is a tail wind. For each test, the climatic conditions shall be recorded in the test report (see ISO 15037-1, Annex B).

General

See ISO 15037-1, Clause 6.4.1.

Tyres

For general information regarding tyres used for test purposes, see ISO 15037-1, Clause 6.4.2. In addition, the following recommendations are given.

Since tyre characteristics have significant effects upon the vehicle behaviour, tyres with known characteristics should be used wherever possible. If unavailable, original equipment rather than replacement market tyres should be used.

For similar reasons, caution should be exercised if worn tyres are to be used. It is known that some tyre characteristics affecting on-centre handling behaviour continue to change throughout the tyre lifespan, especially during the early tyre wear stage (up to several thousand kilometers). To ensure reliable test results, tires with unknown histories should be avoided.

All wheel/tyre assemblies should be balanced before use. Assemblies with excessive run-out or imbalance (detectable as vibration at road wheel rotational frequency) should be avoided.

Operation components

See ISO 15037-1, Clause 6.4.3.

Loading conditions of vehicle

See ISO 15037-1, Clause 6.4.4.

General test conditions shall be in accordance with ISO 15037-1, Clause 6.

See ISO 15037-1, Clause 7.1.

This test maneuver does not mandate any specific initial driving condition. However, the user may find a steady-state, straight-ahead run useful for establishing the vehicle at the required longitudinal test velocity and obtaining transducer reference values. In such cases, the procedure in accordance with ISO 15037-1, Clause 7.2.2 and Figure 2, should be used. The user could find it useful to end the steer input by re-establishing the steady-state, straight-ahead condition at the end of the test run.

See ISO 15037-1, Clause 7.2, for guidance on selection of the appropriate transmission gear for performing the test.

The weave test is an open-loop procedure conducted on a test track that follows a straight-line path. The vehicle is driven at a nominally constant longitudinal velocity.

The standard test velocity is 100 km/h.

Other longitudinal velocities may be used: they should be decremented or incremented by 20 km/h from the standard velocity. Details shall be recorded in the test report (see ISO 15037-1, Annex B, under Test method specific data).

The transducer signals shall be recorded throughout the initial driving condition, if applicable, and for the duration of the test. In order to ensure that the required data are not affected by the instrumentation system, recording should be continued for a further 1 second after the test runs.

The weave test procedure requires the steering-wheel to be subjected to an oscillatory input. The preferred steering input waveform is nominally sinusoidal, but other steering input waveforms (e.g. triangular) may be used. The frequency of the steering input shall be 0,2 Hz ± 10 %. Variations in the steering input frequency within the specified tolerance in the physical test are accounted for in the simulation validation process by using the measured steering-wheel angle as the input for the simulation.

The amplitude of the steering input in the physical test conducted according to ISO 13674-1 shall be sufficient to produce the required peak value of vehicle lateral acceleration. As specified in ISO 13674-1, while analysis for this test is focused on lateral acceleration levels around ±1 m/s² ± 10%, the standard peak lateral acceleration value for the test maneuver is 2 m/s², to ensure good and adequate test data and that the vehicle and its subsystems are working outside of hysteresis effects. Lower peak values and values up to 4 m/s² may also be used for the maneuver as per ISO 13674-1.

Details of the steering input waveform and the lateral acceleration amplitude shall be recorded in the test report (see ISO 15037-1, Annex B, under Test method specific data).

Throughout the test, both the peak amplitude of the steering angle and the angular velocity of the steering- wheel through the centre position shall be as near constant as possible. In addition, variation in the position of the accelerator pedal shall be kept to a minimum, consistent with maintaining vehicle longitudinal velocity within required limits. The longitudinal velocity during the test sequence or sequences to be used for data analysis shall not vary from the nominal value by more than ± 3 %.

Choice of data used for analysis is based upon consistency of steering input and of vehicle longitudinal velocity relevant to that data. A minimum of four consistent cycles of steering input and of vehicle response are required for subsequent data analysis.

The steering input for this test may be made manually or with a steering robot machine. Where manual input is used, in order to obtain the minimum number of cycles required for analysis, the test should be performed for a minimum continuous duration of 40 s, sufficient to capture at least eight cycles of data. Where proving-ground space limitations prohibit a run of this duration, and consequently the reliable capture of a sufficiently long sequence of consistent data, it is permissible to perform a series of shorter runs and to combine the data for analysis. In this event, a nominal 20 cycles of data should be captured and statistical methods should be used during analysis. The use of a steering robot offers the possibility of enhanced consistency of steering input and hence of test data, allowing fewer test cycles.

Raw measurements of steering-wheel angle, yaw velocity, lateral acceleration, longitudinal velocity, and other optional variables shall be filtered and conditioned as specified in ISO 15037-1.

General data shall be presented in the test report in accordance with ISO 15037-1, Annexes A and B. For every change in vehicle loading or configuration, the general data shall be documented again.

At the present level of knowledge, it is not yet known which variables best represent the subjective feeling of the driver and which variables — i.e. which characteristic values — best describe the dynamic reactions of vehicles. Therefore, the following specified variables represent only examples for the evaluation of results.

NOTE Variables evaluated from runs performed at different nominal peak amplitudes of lateral acceleration, or from runs performed using different waveforms of steering input, might not be comparable.

Time histories serve to monitor correct test performance and functioning of the transducers. In particular, the time histories of steering angle amplitude, steering-wheel angular velocity, vehicle longitudinal velocity and vehicle lateral acceleration are examined to identify valid data for evaluation. A minimum of four consistent cycles, for which the control criteria are best met, shall be selected for data analysis. Time histories of the variables listed in Clause 5 shall be presented for the data selected for analysis.

General

The recorded variables are taken in pairs (see Clauses 7.4.3.3 to 7.4.3.7) and plotted one against the other on Cartesian coordinates. For each pair of variables, this produces a series of hysteresis loops laid one over another, the number of loops corresponding to the number of data cycles analysed.

Data processing and method for computing characteristic values

The hysteresis loops should be “averaged” by some suitable method. The recommended method is to make a polynomial curve fit to the combined data over the range “A” for each of the upper and lower sides of the combined hysteresis loops, where “A” is an appropriate percentage of the peak-to-peak abscissa range of the data (see Figure 1). The recommended order for the polynomial curve fit is 3, and the recommended value for “A” is between 50 % and 70 %. The value chosen for “A” should be sufficiently large to adequately cover the range of data of interest, but sufficiently restrictive to avoid end effects from the limits of the loops.

The recommended method for evaluating a gradient is to make a best straight line fit to the polynomial in the region of interest. For an average gradient, this would be over a specified range (as detailed below), and for an “instantaneous” gradient, this would be over a small interval around the point of interest; typically, the size of the interval would correspond to that defined by a lateral acceleration of ± 0,1 m/s².

Other methods may be employed for analysing and averaging the data. For example, straight line rather than polynomial curve fitting may be used. Each hysteresis loop may be analysed individually and the characteristic parameters, yielded from all the loops, averaged to obtain overall results.

The actual procedures and details used will depend upon the analysis software package employed and the nature of the data, and shall be stated in the test report (see ISO 15037-1:2006, Annex B, under Test method specific data).

From the polynomial curve fits to the combined hysteresis loops, the following parameters are evaluated:

• ordinate dead band;
• abscissa dead band;
• gradient.

The pairs of variables to be plotted (ordinate given first) and the characteristics that can be evaluated are given in Clauses 7.4.3.3 to 7.4.3.7.

NOTE All the following characteristic values may not be computed. Depending on the purpose of the test, only relevant characteristic values may be considered.

Steering-wheel torque versus steering-wheel angle (M_H vs. δ_H)

Steering-wheel torque versus steering-wheel angle, as shown in Figure 1, shall be plotted and the following characteristic values shall be computed and recorded in Table 2.

Yaw velocity versus steering-wheel angle (ψ vs. δ_H)

Yaw velocity versus steering-wheel angle, as shown in Figure 1, shall be plotted and the following characteristic values shall be computed and recorded in Table 2.

Yaw velocity versus steering-wheel torque (ψ vs. M_H)

Yaw velocity versus steering-wheel torque, as shown in Figure 1, shall be plotted and the following characteristic value shall be computed and recorded in Table 2.

Lateral acceleration versus steering-wheel angle (a_Y vs. δ_H)

Lateral acceleration versus steering-wheel angle, as shown in Figure 1, shall be plotted and the following characteristic values shall be computed and recorded in Table 2.

Steering-wheel torque versus lateral acceleration (M_H vs. a_Y)

Steering-wheel torque versus lateral acceleration, as shown in Figure 1, shall be plotted and the following characteristic values shall be computed and recorded in Table 2.

The simulation tool shall be configured to replicate the physical test methods outlined in Clause 7. Because the computer simulations may not be sensitive to all physical test factors, not all details specified in ISO 13674-1 and ISO 15037-1 are relevant. Requirements from the physical testrelated to repeated tests, warm-up procedure, transducer properties can be disregarded since they do not influence simulation results. Factors that do not affect the simulations may be neglected to reduce the number of simulations and simplify data processing.

Steering-wheel and longitudinal velocity for simulation shall be from the measured values in the physical testing.

The purpose of each test is to obtain time history of values. These values shall be extracted from the simulation tool and written to file such that they can be used to calculate characteristic values.

Extracted variables shall be filtered with the same method used for measured data as in Clause 7.3.

The time histories of steering angle amplitude, steering-wheel angular velocity, vehicle longitudinal velocity and vehicle lateral acceleration are examined to identify valid data for evaluation. A minimum of four consistent cycles that correspond to those from physical testing shall be selected for data analysis. Time histories of the variables listed in Clause 5 shall be presented for the data selected for analysis.

General

The computed variables from simulations are taken in pairs (see Clauses 7.4.3.3 to 7.4.3.7) and plotted one against the other on Cartesian co-ordinates. For each pair of variables, this produces a series of hysteresis loops laid one over another, the number of loops corresponding to the number of data cycles analyzed.

NOTE All the following characteristic values may not be computed. Depending on the purpose of the test, only relevant characteristic values may be considered.

Data processing and method for computing characteristic values

The same method used for data processing and computing characteristic values for physical testing, as stated in Clause 7.4.3.2, shall be applied to those from computed data from simulation model.

Steering-wheel torque versus steering-wheel angle (M_H vs. δ_H)

A plot of steering-wheel torque versus steering-wheel angle, as shown in Figure 1, shall be generated. The characteristic values defined in Clause 7.4.3.3 shall be calculated and recorded in Table 2.

Yaw velocity versus steering-wheel angle (ψ vs. δ_H)

A plot of yaw velocity versus steering-wheel angle, as shown in Figure 1, shall be generated. The characteristic values defined in Clause 7.4.3.4 shall be calculated and recorded in Table 2.

Yaw velocity versus steering-wheel torque (ψ vs. M_H)

A plot of yaw velocity versus steering-wheel torque, as shown in Figure 1, shall be generated. The characteristic value defined in Clause 7.4.3.5 shall be calculated and recorded in Table 2.

Lateral acceleration versus steering-wheel angle (a_Y vs. δ_H)

A plot of lateral acceleration versus steering-wheel angle, as shown in Figure 1, shall be generated. The characteristic value defined in Clause 7.4.3.6 shall be calculated and recorded in Table 2.

Steering-wheel torque versus lateral acceleration (M_H vs. a_Y)

A plot of steering-wheel torque versus lateral acceleration, as shown in Figure 1, shall be generated. The characteristic value defined in Clause 7.4.3.7 shall be calculated and recorded in Table 2.

As a preliminary step, qualitative visual comparison of the time histories and cross-plots shall be performed to check for major discrepancies in overall shape and trends. The simulation tool is considered formally valid if the relative percentage errors for the characteristic values, calculated as defined in Table 2, are within the specified ranges.

Informative examples illustrating the application of these validation criteria are provided in Annex A.

NOTE The percent relative error is defined as ∥X_S − X_T∥ / ∥X_S∥ × 100, where X_S is the value from the simulation and X_T is the value from the physical test. While the conventional definition of relative error uses the measured or 'true' value (X_T) in the denominator, this document uses the simulated value (X_S). This approach is chosen to provide a more stable basis for the calculation, as the simulation results (X_S) are not subject to the measurement noise and run-to-run variability inherent in physical test data (X_T). Using the more consistent simulated value as the denominator provides a more stable error metric when evaluating results from multiple tests.

Table 2 — Simulation validation criteria and tolerances

The simulation shall be documented to the extent needed to reproduce the simulated tests. This shall include names of software tools, including version numbers, and internal model names. A list of files used to run the simulation shall be provided, and copies of the files shall be archived.

1. 
   (informative)
   
   Example validation results

This annex provides two examples of validation results to illustrate the application of the criteria defined in Table 2. The following tables illustrate two distinct outcomes of the validation process. Table A.1 represents Case 1, an example of a simulation model that demonstrates full compliance with the tolerance values for relative for percent errors. In this case, all calculated percent relative errors for the characteristic values fall within the 15,0% tolerance limit specified in Table 2. Based on these results, the simulation tool would be considered validated for this test procedure.

On the other hand, Table A.2 represents Case 2, where the simulation shows partial compliance with the tolerance values for relative percent errors. Several key characteristic values, such as "Response dead band" (21,3%) and "Lateral acceleration dead band" (27,8%), have percent relative errors that exceed the 15,0% tolerance. Therefore, the simulation in Case 2 is not considered validated, prompting investigations into the source of the errors (e.g., model architecture, parameters, test conditions). Based on this analysis and the intended use of the model, the resulting action can be either to refine the model or to justify a relaxation of the tolerance values.

Table A.1 — Example validation results: Case 1 (Full compliance)

Table A.2 — Example validation results: Case 2 (Partial compliance)