Developing Affective Computing for Smart Indoor Environments: A VR-Based Multimodal Framework on Window View Access

Understanding how visual environmental parameters influence human emotions is essential for designing intelligent and affective indoor spaces. This study presents a VR-based multimodal framework for assessing psychological and physiological responses to window view access parameters in indoor environments. The framework integrates virtual scene generation, real-time physiological signals acquisition, and interactive subjective evaluation to quantify user responses under controlled conditions. Two virtual experiments were conducted: the first examined the effects of window-to-wall ratio (WWR) and space type (café vs. workspace), while the second explored the influence of viewing distance (close, medium, and far). Emotional valence, arousal, and satisfaction were collected alongside electrodermal activity (EDA) and heart-rate variability (HRV). Results showed that, under the visually pleasant and constant external landscape condition used in this study, larger windows and shorter viewing distances significantly enhanced positive affect and satisfaction, producing a high-valence yet low-physiological-arousal emotional pattern. EDA proved most sensitive to changes in visual openness, while HRV showed limited variation. These findings demonstrate the feasibility of the proposed VR-based multimodal framework for capturing affective responses to architectural features and provide methodological and empirical support for emotion-aware design in smart indoor environments.