Machine Learning Case Study: SQ Metric Development for Electric Vehicles

The trend towards vehicle powertrain electrification leads to an indispensable need for quantification of tonal content in vehicle interior noise. With current progress made on hearing models for tonality calculation, reliable acoustic predictors are meanwhile available for instrumental evaluation of tonal noise. Using Machine Learning approaches, Sound Quality Metrics for EV powertrain whine noise can be designed to predict perceived acoustic quality.

This online-seminar covers a typical Machine Learning case study, presenting current progress in the development of Sound Quality Metrics for EV powertrain whine noise.

Topics to be covered:

• Pre-processing of a measurement dataset using OTPA
• (Psycho-)acoustic description of EV powertrain whine noise
• Jury testing as a key methodology for initial dataset labelling
• SQ Metric Development based on regression
• Cross-Validation of regression models to ensure robust predictions for unknown data
• Consulting Services and Cooperation Opportunities