Page 26 - Winter 2020

ALERTING PEDESTRIANS BY SOUND
Figure 2 illustrates the psychoacoustic effects caused by the superposition of alert signals based on a few tones in a virtual road traffic scenario according to Genuit (2016). If the vehicles have slightly different speeds (indicated in Figure 2, right), the alert signals based on a few tones slightly shifted in frequency produce an overall disharmonic sound composed of multiple modulations (see Figure 2, bottom right). To avoid those disharmonic modulations and to achieve a reasonable detectability, the use of amplitude modulations is frequently proposed (cf Robart et al., 2013).
Audible amplitude modulations seem to be beneficial for increased detectability and localizability. Regulations specifying minimum sound requirements do not require any form of modulation, although it is not excluded.
If traffic noise resulting from introduction of alert signals increases by 0.5 dB, there is no reason to expect a significant increase in (highly) annoyed people. However, it is well-known that certain noise properties, such as prominent tones, can lead to an increase in noise annoyance (Schäffer et al., 2016). To account for such annoyance-relevant properties beyond the sound pressure level, penalties in decibels are frequently added to a measured overall sound pressure level to reflect increased noise annoyance. For example, the German standard DIN 45681 (2005) proposes that tones penalties of 3 dB (or even 6 dB) be added to the measured overall sound pressure level. Also, dissonant noise patterns caused by several “untuned” superposed alert signals can increase noise annoyance in addition to the absolute level. Thus, we need to apply a penalty of a few decibels to the new road traffic noise to quantify its “perceived” level increase due to annoyance. If we follow this logic, the impact on the human environment might be more than only “negligible.” Consequently, a study of the acceptance of AVAS showed a significant increase in the emotional arousal level, measured with the self-assessment manikin method (SAM), as a result of an AVAS sound based on prominent tonal components (Fagerlönn et al., 2018).
The Future
Although much has already been learned, it is clear that more study is needed to determine the actual impact of alert signals on pedestrians and bicyclists with respect to both accident rates and noise annoyance. This can
become critical as the number of EVs in road traffic increases and all new EVs are required to emit alert sounds in compliance with the regulations of their nation. The study of benefits and drawbacks of the nationwide alert sounds must focus on road traffic noise effects on the public as well as on the accident rates associated with them. Of course, it is also necessary to consider the well- being of visually impaired persons because these sounds are particularly important for that group. Any adjustment in the regulations and specifications, up to a waiver of alert sound requirements, must be well grounded in the observation of the real impact of the alert signals.
It goes without saying that diverse alternative technological solutions are conceivable for supporting,
Figure 2. Simulations of a single alert signal based on a few tones emitted by one car (left) and of a microtraffic scenario with five vehicles all with slightly different speeds (10, 11, 12, 13, and 14 km/h) applying pitch shift (right).Top, spectra over time (fast Fourier transform vs. time) displayed for both scenarios; color, sound pressure level for each frequency. Bottom, modulations occurring in both scenarios (modulation spectrum vs. band [degree of modulation] for the different simulations above), with information about the modulation rates (x-axis) and the carrier frequencies (y-axis); color, strength of modulation in terms of the degree (percentage) of modulation. From Genuit, 2016.
26 Acoustics Today • Winter 2020