Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.687173
Title: Understanding and improving methods for exterior sound quality evaluation of hybrid and electric vehicles
Author: Singh, Sneha
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2016
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
Electric and Hybrid Electric Vehicles [(H)EVs] are harder for pedestrians to hear when moving at speeds below 20 kph. Laws require (H)EVs to emit additional exterior sounds to alert pedestrians of the vehicles’ approach to prevent potential collisions. These sounds will also influence pedestrians’ impression of the vehicle brand. Current methods for evaluating (H)EV exterior sounds focus on “pedestrians’ safety” but overlook its influence on “vehicle brand”, and do not balance experimental control, correct context along with external and ecological validity. This research addresses the question: “How should (H)EV exterior sounds be evaluated?” The research proposes an experimental methodology for evaluating (H)EV exterior sounds that assesses pedestrians’ safety and influence on the vehicle brand by measuring a listener’s detection rate and sound quality evaluation of the (H)EV in a Virtual Environment (VE). This methodology was tested, improved and validated through three experimental studies based on their findings. Study 1 examined the fidelity of the VE setup used for experiments. The VE was immersive with sufficient degree of involvement/control, naturalness, resolution, and interface quality. It also explored a new interactive way of evaluating (H)EV sounds where participants freely navigate the VE and interact with vehicles more naturally. This interactivity increased the experiments’ ecological validity but reduced reliability and quadrupled the experiment duration compared to using a predefined scenario (non-interactive mode). Thus, a predefined scenario is preferred. Study 2 tested the non-interactive mode of the proposed methodology. Manipulating the target vehicle’s manoeuvre by varying factors, namely the vehicle’s “arrival time”, “approach direction” and “distance of travel”, across the experiment conditions increased ecological validity. This allowed participants to think, respond and pay similar attention as a real world pedestrian. These factors are neglected by existing methodologies, but were found to affect the participants’ detection rate and impression of the vehicle brand. Participants detected the vehicle more than once due to confusion with real world ambient sounds. In the real world, pedestrians continuously detect a vehicle in presence of non-vehicular ambient sounds. Therefore, recommendations to improve the representation of the real-world processes in the vehicle detection during listening experiments include an option to re-detect a vehicle and subjective evaluation of ‘detectability’ of the vehicle sounds. The improved methodology adds ‘detectability’ and ‘recognisability’ of (H)EV sounds as measures and (H)EV’s arrival time as an independent variable. External validity of VEs is a highly debated yet unanswered topic. Study 3 tested external validity of the improved methodology. The methodology accurately predicted participants’ real world evaluations of the detectability of (H)EV sounds, ranked order of the recognisability of (H)EV sounds and their impressions about the vehicle brand. The vehicle’s arrival time affected participants’ detection rate and was reaffirmed as a key element in the methodologies for vehicle sounds’ detection. The final methodological guidelines can help transportation researchers, automotive engineers and legislators determine how pedestrians will respond to the new (H)EV sounds.
Supervisor: Not available Sponsor: Warwick Manufacturing Group
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.687173  DOI: Not available
Keywords: TL Motor vehicles. Aeronautics. Astronautics
Share: