Guidelines for methods studying human perception of sound and vibration in passenger cars
Doctoral thesis, 2024
As automotive technology advances, particularly in combustion-engine (CV) and electric vehicles (EV), ride comfort has become a critical attribute for future car development. A multitude of factors, including seat, sound, and vibration, significantly influence the perceived ride comfort in passenger cars. Despite numerous studies on human responses to sound and vibration, there is a noticeable gap in research investigating real occupants’ experiences under various real-world driving scenarios. Additionally, there is a lack of clear guidelines for utilizing advanced technologies in the study of ride comfort.
This thesis aims to bridge this gap by examining human experiences of sound and vibration in conventional passenger cars, developing methodologies to assess their impact on perceived ride comfort. The primary purposes of this thesis are as follows: (1) to define ride comfort from the occupant’s perspective, identifying factors that influence it, (2) to investigate how sound and vibration specifically affect ride comfort, and (3) to propose guidelines and a framework for using advanced technologies in studying ride comfort.
The research methodology encompasses a literature review, a field study on sound and vibration experienced during various driving scenarios, an interview study on the use of driving simulators, the development of a machine learning framework, and a focus group study to evaluate the proposed framework.
The literature review reveals that while significant findings are available from laboratory settings, studies integrating all parameters affecting overall ride comfort in real-world contexts are limited. Furthermore, there is a need to delve deeper into how sound and vibration influence occupants' overall ride comfort.
To address this, the field study was conducted using eight typical driving scenarios with ten participants in both a CV and an EV. Results indicated similarities in initial comfort aspects such as seat adjustment and body room but differences in dynamic discomfort, with body movements being a concern in the CV, and sound annoyance more prominent in the EV. Moreover, induced body movements dominated vibration discomfort, while sound annoyance consistently compounded over time, making relaxation difficult for occupants.
In addition to field studies, this thesis also explores the role of driving simulators in user performance, experience, and ride comfort studies. Through an interview study involving 14 participants, guidelines for using high-level driving simulators were proposed. The research acknowledges the advantages of simulators, such as improved safety, repeatability, controllability. Furthermore, it emphasizes their capability to isolate variables and conduct experiments with fewer physical constraints, along with enabling rapid transitions between components, structures, and vehicle models. However, the research also addresses limitations, including space constraints and communication difficulties.
To tackle the challenges of traditional ride comfort evaluations, this thesis proposes a machine learning framework to overcome limitations such as data quality and quantity, cross-study comparison, and model interpretability. This framework aims to augment existing data, propose suitable performance metrics, and improve the accuracy and reliability of ride comfort prediction models. Additionally, a focus group study evaluates the feasibility of these machine learning methods, identifying their advantages of enhancing prediction performance and refinement methods that could be integrated.
In conclusion, this thesis provides a set of guidelines derived from field studies, driving simulator research, and innovative machine learning approaches to address the multifaceted nature of ride comfort in automotive design.
This thesis aims to bridge this gap by examining human experiences of sound and vibration in conventional passenger cars, developing methodologies to assess their impact on perceived ride comfort. The primary purposes of this thesis are as follows: (1) to define ride comfort from the occupant’s perspective, identifying factors that influence it, (2) to investigate how sound and vibration specifically affect ride comfort, and (3) to propose guidelines and a framework for using advanced technologies in studying ride comfort.
The research methodology encompasses a literature review, a field study on sound and vibration experienced during various driving scenarios, an interview study on the use of driving simulators, the development of a machine learning framework, and a focus group study to evaluate the proposed framework.
The literature review reveals that while significant findings are available from laboratory settings, studies integrating all parameters affecting overall ride comfort in real-world contexts are limited. Furthermore, there is a need to delve deeper into how sound and vibration influence occupants' overall ride comfort.
To address this, the field study was conducted using eight typical driving scenarios with ten participants in both a CV and an EV. Results indicated similarities in initial comfort aspects such as seat adjustment and body room but differences in dynamic discomfort, with body movements being a concern in the CV, and sound annoyance more prominent in the EV. Moreover, induced body movements dominated vibration discomfort, while sound annoyance consistently compounded over time, making relaxation difficult for occupants.
In addition to field studies, this thesis also explores the role of driving simulators in user performance, experience, and ride comfort studies. Through an interview study involving 14 participants, guidelines for using high-level driving simulators were proposed. The research acknowledges the advantages of simulators, such as improved safety, repeatability, controllability. Furthermore, it emphasizes their capability to isolate variables and conduct experiments with fewer physical constraints, along with enabling rapid transitions between components, structures, and vehicle models. However, the research also addresses limitations, including space constraints and communication difficulties.
To tackle the challenges of traditional ride comfort evaluations, this thesis proposes a machine learning framework to overcome limitations such as data quality and quantity, cross-study comparison, and model interpretability. This framework aims to augment existing data, propose suitable performance metrics, and improve the accuracy and reliability of ride comfort prediction models. Additionally, a focus group study evaluates the feasibility of these machine learning methods, identifying their advantages of enhancing prediction performance and refinement methods that could be integrated.
In conclusion, this thesis provides a set of guidelines derived from field studies, driving simulator research, and innovative machine learning approaches to address the multifaceted nature of ride comfort in automotive design.
Human perception
Vibration
Ride comfort
Sound
IMS Room Studio 1
Opponent: Arne Nykanen
Author
Xiaojuan Wang
Chalmers, Industrial and Materials Science, Design & Human Factors
Human Response to Vibrations and Its Contribution to the Overall Ride Comfort in Automotive Vehicles - A Literature Review
SAE Technical Papers,;Vol. 2020-April(2020)
Journal article
Sound and vibration influence overall ride comfort in a combustion passenger car under different driving scenarios
Influence of sound and vibration on perceived overall ride comfort – A comparison between an electric vehicle and a combustion engine vehicle
Guidelines for using high-level driving simulators in user studies – an interview study regarding user performance, experience, and ride comfort
An approach for predicting vibration annoyance via machine learning: Integrating objective measurements of sound and vibration with passengers’ individual characteristics
Enhancing Ride Comfort: Bridging Human Experience and Advanced Automotive Technology
Picture yourself inside a modern vehicle—be it a combustion-engine or electric vehicle—where the journey is not just about the destination but also about the experience. Ride comfort has emerged as a pivotal aspect of automotive design, playing a crucial role in our daily commutes. Traditionally, comfort has been attributed to factors like seating, noise, and vibration. However, while these elements have been extensively studied, there remains a significant gap in understanding how real passengers perceive comfort in diverse driving conditions.
This thesis addresses the crucial yet underexplored aspect of passenger-perceived ride comfort in modern vehicles, focusing on how sounds and vibrations influence this perception in various driving conditions. The research aims to: (1) define ride comfort from a passenger's perspective, (2) analyze the impact of sound and vibration, and (3) establish advanced technology guidelines for comfort studies.
Our methodology includes a literature review, field studies in diverse driving scenarios, driving simulator analyses, and a machine learning framework evaluation through focus groups. Findings indicate that, while seating and space are universal comfort factors, body movements and sound annoyance create distinct discomfort in combustion and electric vehicles, respectively. Vibrations cause immediate discomfort, whereas sound disturbances accumulate, affecting relaxation.
We also explored driving simulators' potential, establishing guidelines to enhance study safety, repeatability and efficiency despite spatial and communication constraints. The novel machine learning framework aims to improve data quality and ride comfort prediction accuracy.
In summary, this thesis provides guidelines from field studies, simulation research, and machine learning to enhance ride comfort understanding and prediction, ultimately improving passenger experiences.
Picture yourself inside a modern vehicle—be it a combustion-engine or electric vehicle—where the journey is not just about the destination but also about the experience. Ride comfort has emerged as a pivotal aspect of automotive design, playing a crucial role in our daily commutes. Traditionally, comfort has been attributed to factors like seating, noise, and vibration. However, while these elements have been extensively studied, there remains a significant gap in understanding how real passengers perceive comfort in diverse driving conditions.
This thesis addresses the crucial yet underexplored aspect of passenger-perceived ride comfort in modern vehicles, focusing on how sounds and vibrations influence this perception in various driving conditions. The research aims to: (1) define ride comfort from a passenger's perspective, (2) analyze the impact of sound and vibration, and (3) establish advanced technology guidelines for comfort studies.
Our methodology includes a literature review, field studies in diverse driving scenarios, driving simulator analyses, and a machine learning framework evaluation through focus groups. Findings indicate that, while seating and space are universal comfort factors, body movements and sound annoyance create distinct discomfort in combustion and electric vehicles, respectively. Vibrations cause immediate discomfort, whereas sound disturbances accumulate, affecting relaxation.
We also explored driving simulators' potential, establishing guidelines to enhance study safety, repeatability and efficiency despite spatial and communication constraints. The novel machine learning framework aims to improve data quality and ride comfort prediction accuracy.
In summary, this thesis provides guidelines from field studies, simulation research, and machine learning to enhance ride comfort understanding and prediction, ultimately improving passenger experiences.
Subject Categories
Mechanical Engineering
Other Engineering and Technologies
Areas of Advance
Production
ISBN
978-91-8103-145-4
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5603
Publisher
Chalmers
IMS Room Studio 1
Opponent: Arne Nykanen