Human Response to Electric Vehicle Noise: Laboratory Findings on Localization, Annoyance, and Non-Auditory Effects

Paper in proceeding, 2026

Electric vehicles (EVs) are increasingly present in urban environments and typically emit little noise at low driving speeds, requiring artificial warning sounds via acoustic vehicle alerting systems (AVAS). While these sounds are intended to improve pedestrian traffic safety, they also become part of the urban acoustic environment. This contribution summarizes a series of laboratory studies on human responses to electric vehicle noise, focusing on auditory localization, subjective appraisal, and physiological and cognitive responses. The first study addresses the localization of electric vehicles in a static parking lot scenario. Using a concealed array of 24 loudspeakers and auralized stimuli derived from vehicle recordings, localization accuracy and response time were examined while simulating EVs with different AVAS signal types and configurations. Results from 52 participants show that AVAS design can strongly affect localization performance. In particular, highly tonal signals that may be effective in single-vehicle detection tasks showed poor localizability, especially when multiple vehicles with similar AVAS are present. The second experiment focused on residents exposed to indoor traffic noise from EVs driving slowly on a nearby street. In a living-room-like laboratory environment, 60 participants performed a combined Eriksen Flanker and spatial Stroop attention test while being exposed to low-level EV noise at realistic closed-window levels (LA,eq ≤ 21.5 dB). In addition to subjective workload and annoyance ratings, electrodermal activity was recorded as an indicator of physiological arousal. While attention performance was not measurably affected, highly tonal warning sounds elicited higher annoyance ratings, larger perceived workload, and greater physiological arousal than other signals.  Taken together, these findings demonstrate that human responses to electric vehicle noise extend beyond detectability and depend on signal characteristics and context. The results suggest that AVAS design choices can influence multiple aspects of human response, underlining the usefulness of quantitative human-response measurements in AVAS development and evaluation.