| dc.description.abstract | The aim of this PhD thesis is to facilitate the integration of human emotional dimensions in engineering, and engineering design in particular. Grounded and guided by the fundamentals in affective neuroscience and complementary to the human-computer interaction paradigm of affective computing, we call for the emergence of Affective Engineering as a key contributor to future product and system design and development.
Affect is a critical dimension in human-machine interactions, whether it be on a ship bridge, in a fully autonomous car, or in the operator station of a nuclear power plant, because of potentially irrational and risky behavior of the human operator. For example high-stress/high-arousal situations include high risk of human failure, such as stress-induced unfavorable decision-making in close-to-collision situations. In contrast, low-stress/low-arousal situations may involve irrational behavior induced by drowsiness, e.g., falling asleep and leaving a vehicle unattended. In an engineering context, we aim to quantitatively measure situational variables that allow us to control for those affective behavior responses; that is, we seek to use primarily physiology sensors to “read” the human in human-machine interactions. Based on an accurate reading of a human’s affective state, the objective is to design and engineer interactions/interfaces that adapt dynamically in order to support the human in the best way. Ultimately, we seek to engineer new interaction designs that incorporate the dimension of human affect; we would like to call this approach New Affect Integrated Interaction Design.
This thesis addresses the academic fundamentals of New Affect Integrated Interaction Design by advancing (1) theory from psychology, neuroscience, and physiology; (2) the instrument of applying psychological, physiological, and behavioral measurement tools to quantify affect; and (3) designing and building in situ experiments on critical human-machine interactions. Specifically, this thesis presents measurement tools based on the principles of the physiological response to affect, and highlights the challenges of in situ experiments especially regarding confounding parameters. We present two empirical studies – stress measurements of ship captains with salivary cortisol, and functional near-infrared spectroscopy (fNIRS) cortical activity measurements in simulated autonomous driving. We also validate salivary cortisol as a new measurement tool for human-computer/humanmachine interactions. Based on the empirical results, we argue for the methodological triangulation of measurement tools. We offer considerations for designing and conducting reliable Affective Engineering experiments, including the sharing of data. Finally, we present our human sensory system-based concept which may be useful for quantifying design parameters in New Affect Integrated Interaction Design.
This thesis illustrates that we, as a collective Affective Engineering research community, have achieved the first milestone – namely, the ability to measure affect in critical human-machine interactions. Also, first endeavors towards reaching milestone 2 – that is, the testing and understanding of the impact of design parameter manipulation on human affect and related behaviors as foundation for New Affect Integrated Interaction Design – are appearing. Milestone 3 will witness the translation of lessons and insights from research to normative recommendations, tools, and processes that will allow industry to manufacture New Affect Integrated Interaction Designs on a bigger scale. Following the examples of Design for X or Design for Manufacturing, we thus predict the emergence of Design for Affect Integration. | nb_NO |
