Exploring non-verbal methods in voice interaction systems for autonomous driving applications

Abstract

As autonomous driving technology continues to evolve, in-vehicle voice interaction has become more natural and personalized. However, these systems often face limitations when managing tasks that require rapid responses or precise control. Traditional voice input may not be suitable for all scenarios due to challenges with processing speed and response latency. To address these constraints this study explored the combination of non-verbal sounds and voice input in autonomous driving, with a focus on system activation and continuous non-driving-related tasks. In Experiment 1, participants used non-verbal sounds to wake-up the system and compared this method with traditional wake-up words and wake-up free approach. Results showed that many users still preferred traditional wake-up methods, although snapping fingers did not show a significant disadvantage in terms of interaction duration. In Experiment 2, non-verbal sound input was further developed for continuous task control and was tested alongside multiple voice commands and the Stop input for continuous non-driving-related tasks. While the combination of non verbal input methods was innovative, the Stop command was highly favored by participants, likely due to its higher accuracy and lower subjective workload, which may have been influenced by task design. Overall, this study introduces a novel approach to non-verbal sound input, offering new insights into voice input design and future interactions in autonomous vehicles.

Date of Award	15 Jul 2025
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Xu Sun (Supervisor) & Qingfeng Wang (Supervisor)