Abstract
This research explores the influence of frequency combinations for use in Acoustic Vehicle Alerting Systems (AVAS) on annoyance perception, focusing on UN-R138-compliant sounds synthesised from four frequency components. With the exponential increase of Electric Vehicles (EVs), effective AVAS solutions are required by law to mitigate risks to vulnerable road users due to the vehicles' quietness. Other than detectability, these systems should minimise contributions to urban noise annoyance which is associated with negative health effects. In parallel, sonic identity of an EV is a crucial design focus for manufacturers, not only for making quiet vehicles perceptible but also for reflecting their unique character, type, or communicating vehicle dynamics.The research question under investigation was ``How do different combinations of frequency components in UN-R138-compliant AVAS sounds influence noise annoyance ?''. This question aims to explore frequency component combinations in AVAS sounds and how they contribute to annoyance in order to provide a better understanding of which sounds are optimal, in annoyance terms, for AVAS use. The sound synthesis model used for AVAS in this study was of a physical model of a rolling sphere, as it was considered to better communicate the perception of an EV.
Zwicker's psychoacoustic annoyance model was used to objectively evaluate all possible fourfrequency component UN-R138-compliant sounds. This method allowed for selecting a representative subset of the sounds for subjective evaluation through an experimental study. This study used a virtual world simulation method to evaluate the sounds in audio and audiovisual domains. Hierarchical clustering was utilised to categorise sounds and investigate Inter-Frequency Distance (IFD) patterns. The experimental study revealed that base frequency and IFD in sounds comprising of four frequency components influences annoyance perception depending on specific frequency ranges. A significant negative correlation between the psychoacoustic annoyance model and subjective annoyance ratings was observed indicating complexities in annoyance perception of AVAS.
| Date of Award | 17 Jan 2025 |
|---|---|
| Original language | English |
| Sponsors | SYN-ENOSIS |
| Supervisor | Hyunkook Lee (Main Supervisor) & Minsi Chen (Co-Supervisor) |