The session will cover database analysis including review of documents and discussion points from pilots view, AVAS term dL Dynamic factor x, review of decisions and work for the next meeting.

This presentation analyzes RD-ASEP model performance using test run data from Annex_9_test_runs where both left and right sound were measured. Current model shows 1.3% of measurements above L_exp; without the 2 dB margin this increases to 8.4%. Model overestimates at low speeds and underestimates at higher accelerations. Tyre rolling sound parameters have most influence on results and dominate most of the data set. Optimizing all 13 parameters reduces mean absolute error from 2.1 to 1.1 dB. More high acceleration measurements are needed to reliably calibrate the dynamic component.

Task Force Data Analysis (Sub Group to IWG RD-ASEP) held its 13th meeting on Wednesday, 3rd June 2026. The group reviewed database analysis including statistical analysis of differences between expected and measured sound levels, sub-model construction, and optimization calculations for parameter tables. Discussion addressed model construction concerns, vehicle failure differentiation, and data collection requirements including approval values according to UN R117. Next meetings are scheduled for Tuesday, 7th July and Monday, 24th August 2026.

The agenda includes a review of the schedule with next meetings on 7th July and 24th August, continuation of database analysis presentation from RDW, review of decisions and work for the next meeting, report to IWG RD-ASEP.

The Task Force Data Analysis held 14 meetings, with 12 used to eliminate database errors and 2 to begin data analysis. The analysis revealed the need to better understand the concept and history of RD-ASEP development. Questions were raised regarding whether RD-ASEP should follow envelope functions framing typical sound behaviour of vehicle components or average behaviour with acceptable tolerance. Comparing databases from 2017/2020 and 2023/2024 showed different pass/fail rates across vehicle types. The group seeks clarification on RD-ASEP model optimization, accuracy expectations, and documentation explaining historical background and parameter selection.

The agenda includes welcoming participants, reviewing and adopting the agenda, scheduling next meetings on 7th July and 24th August 2026, analyzing database distance in dB from the borderline, discussing proposals to TF-Editing based on draft UN R51.04 from the June Gothenburg session, preparing wording proposals for items 13–19 from the issues list, and reviewing decisions and work for the next meeting before reporting to IWG RD-ASEP.

This presentation, intended for discussion within TF Data Analysis RD-ASEP and not for conclusions, analyzes model performance using L_exp data from Annex_9_test_runs where both left and right sound were measured. Model performance varies across speed bins, with higher mean absolute error at low and high speeds. Adjusting theta_tr values improves mean absolute error across all speed bins. Testing without an envelope function results in more measurements exceeding expected values, requiring extra margins.

Proposal to add para. 4.4.1. through 4.4.8. to restructure target operation conditions provisions, add para. 4.4.7. to permit acceptance of runs outside the control range under specified provisions, modify para. 3.6.1. to change reference acceleration determination from maximum to selected acceleration for a selected reference v-t-e ratio, modify para. 4.5.4.1. to set negative calculated acceleration values to zero, add heading “Carry-over parameters from Annex 3” to para. 2.2.1.2., add para. 2.2.1.2.1. and 2.2.1.2.2., and modify constant speed test parameters to report vehicle speed vBB’_CRS_ANCHOR rather than reference speed vTEST of 50 km/h.

Task Force Data Analysis held an online meeting on 7 July 2026. The group discussed database analysis, including insufficient data for higher performance and acceleration, and agreed that more data are needed. When new data are generated, operation conditions should focus on higher performances covering the ASEP range, higher accelerations up to and exceeding 5 m/s, pass-by runs with gear shifting, and low speed measurements in the UN R117 AVAS control range. The group reviewed proposals to the TF-Editing based on draft UN R51.04 and agreed some changes, with review results to be uploaded as Rev.1. A question regarding the phrase “but no more than 2.0 m/s” in paragraph 3.1.2.1.4.3 regarding vehicles with only one gear ratio was noted as issue #23 pending further review.

The Task Force Data Analysis meeting scheduled for 15 September 2026 will discuss database pilot documents covering amendments to factor x, reference acceleration, AVAS term considerations, simplified EV models, and RD-ASEP model slopes. The task force will prepare proposals based on draft UN R51.04 for items 13–19 from the Editing TF issues list and address a question regarding the phrase “but no more than 2.0 m/s” in paragraph 3.1.2.1.4.3 concerning acceleration values for vehicles with single gear ratios.

Document presents considerations on amendments to the RD-ASEP Sound Expectation Model SEM. A simplified approach for electric vehicles is proposed based on Supplement 11 to UN R51.03, using a tyre rolling sound curve with slopes identical to internal combustion engine and hybrid electric vehicles, plus an acceleration term from UN R51.03 Supplement 11 as an additional term, and the AVAS term from Option 1, without other parts of the SEM.

Document presents suggested revisions to test procedures in Annex 7 and Annex 7 Appendix 1:

• Renumber para. 4.4. into paras. 4.4.1. through 4.4.8. to structure target operation conditions,
• Insert new para. 4.4.7. to permit acceptance of measurement runs outside the control range when agreed by manufacturer and authority,
• Modify para. 3.6.1.1. to define maximum reference acceleration for a selected reference v-t-e ratio and clarify determination procedures,
• Add provision that in case calculated acceleration is negative, the value is set to zero,
• Restructure para. 2.2.1.2. to introduce subsections 2.2.1.2.1. and 2.2.1.2.2. for parameters from acceleration and constant speed tests respectively, and modify parameter definitions including use of vPP,CRS instead of fixed 50 km/h reference speed.

Considerations by TF-Data Analysis regarding tyre slopes, power train slopes, and part load model for the RD-ASEP Sound Expectation Model. Tyre slopes α = 40 and β = 20 are confirmed as appropriate for UN R117 approved tyres. Power train slopes are confirmed as well-framed to normal vehicle behaviour and consistent with Annex 7 ASEP Slope-Assessment principles. For part load model, the function is confirmed for ICE vehicles and HEV in ICE mode, while for EVs and HEVs in EV mode the part load area may require reworking, but the part load model should be applied uniformly to all vehicle techniques.

Document presents considerations on revising the AVAS term in the RD-ASEP Sound Expectation Model to align with UN R138.02 specifications for electric and hybrid electric vehicles. Two options are discussed: Option 1 implements only constant speed specifications from UN R138.02, while Option 2 extends this with an additional term to reflect interpretation of paragraph 6.2.8 for accelerated conditions. A revised AVAS term formula applicable for vehicle speeds up to 50 km/h is proposed.

Proposal to amend the determination of reference acceleration aREF by introducing a theoretical calculation formula based on PMR, rated engine speed S, and vehicle mass mro, with a correction function accounting for transmission losses, virtual mass increase, tyre slip, and turbo delay, and to allow manufacturers to alternatively determine reference acceleration by discrete measurements.

Document presents considerations on factor x used in the Sound Expectation Model to split constant speed test results into tyre rolling sound and power train mechanic references. Factor x values are concentrated between 80 and 99, indicating tyre dominance. Small variations in measurements cause disproportional variability in the minor sound source when one source dominates. Two options are proposed: Option 1 requires factor x determined during type approval to be reported and used for subsequent assessments, with any redetermination not exceeding the type approval value; Option 2 requires the power train reference from type approval to be used for all further assessments regardless of measurement variations.

Agenda for the Task Force Data Analysis meeting of 29 September 2026. Discussion topics include amendments to factor x for improved model robustness, consideration on LDYN with amendments on factor x, considerations on the AVAS term to match UN R138, amendments to reference acceleration, temperature correction, indoor testing, simplified model for electric vehicles with integrated AVAS approach and amendments from Supplement 11, and considerations on slopes in the RD-ASEP model. The Task Force will prepare proposals for wording on issues identified in the editing task force list and address a clarification request regarding the phrase ‘but no more than 2.0 m/s’ in para. 3.1.2.1.4.3. concerning vehicles with only one gear ratio.

Proposal to amend Annex 7, para. 2.3.1 to constrain the factor x determined by discrete coast-down measurement to a minimum of 10 and a maximum of 90 for internal combustion engine and hybrid electric vehicles, or 95 for electric vehicles, ensuring the Sound Expectation Model remains robust against measurement variations in tyre rolling sound and powertrain mechanic sound splitting.

Document presents considerations on amendments to the RD-ASEP Sound Expectation Model SEM concerning calculation of the dynamic base and dynamic load addon. The dynamic sound increase of a vehicle is covered by a dynamic term aggregating all dynamics. A minimum threshold is proposed to be specified as max(0.25; ΔL_dyn) to ensure the function does not become zero or negative due to measurement uncertainties. The value 0.25 is selected to deliver comparable results as the current 10 dB minimum dynamic specification. In combination with a restriction of factor x to an area of 10 to 90 for ICE, the model becomes very robust with fluent transients and no unsteady steps.

Considerations by TF Data Analysis on revising the AVAS term in the RD-ASEP Sound Expectation Model to align with UN R138.02 specifications for electric and hybrid electric vehicles. The document presents two options: Option 1 implements UN R138.02 specifications for constant speed conditions; Option 2 extends Option 1 with an additional term to reflect paragraph 6.2.8 requirements for accelerated conditions. A revised AVAS term formula is proposed applicable only for vehicle speeds up to 50 km/h.

Proposal to restructure para. 4.4. by dividing it into paras. 4.4.1. through 4.4.8., add a new para. 4.4.7. permitting acceptance of measurements of RD-ASEP leading to operation conditions outside the control range when agreed by the manufacturer and authority, modify the heading of para. 3.6.1. to reference a selected reference v-t-e ratio, update references to v-t-e ratio terminology throughout para. 3.6.1.1., add a provision that negative acceleration values be set to zero in para. 4.5.4.1., restructure para. 2.2.1.2. by renumbering into paras. 2.2.1.2.1. and 2.2.1.2.2., and modify the reference vehicle speed parameter designation from vTEST to vPP,CRS.

Agenda for the 18th meeting of the Task Force Data Analysis, a sub-group to IWG RD-ASEP. The meeting will review and adopt the agenda, discuss the meeting schedule for November and December 2026, continue discussion on necessary amendments to RD-ASEP including LDYN definition, AVAS term alignment with UN R138, temperature correction, indoor testing, simplified EV model with integrated AVAS approach and Supplement 11 amendments, and slopes and parameter table settings. The meeting will review proposals to the TF-Editing and decisions for the next meeting.

Task Force Data Analysis provides input to the Editing Task Force on three items. For vehicles with single gear ratios, remove brackets from text restricting acceleration to no more than 2.0 m/s and continue with adopted text for UN R51.04. Restructure Annex 7 paragraph 4.4. into subparagraphs for clarity. Rework paragraph 3.6.1. to define reference acceleration with v-t-e ratio and specify procedures for determining maximum reference acceleration. Add specification to factor x determination to prevent extreme results and harmonize its representation as either percentage or decimal value.

Proposal to replace the current logarithmic function with a segmented function that linearizes the curve in the unstable area near the clipping boundary. The new formulation uses a tangent line to the logarithmic curve below the tangency point and clips to a minimum value of 10 dB. Two parameters define the tangency and clipping conditions: clipping value of 10 dB and slope of 5 dB per unit. The detailed formulation provides the mathematical definitions of the logarithmic function, the linearized segment, helper variables, and constraints including tangency and asymptote conditions.