Exploring Human-Virtual agent Collaboration for layout design in VR

This thesis explores human-virtual agent collaboration for layout design in a virtual reality environment. Specifically, we developed a human-in-the-loop optimization-based method that allows the decision-making of the virtual agent. Our algorithm accounts for spatial constraints in furniture placement by evaluating wall proximity, collision costs, and relationships between furniture items. It also considers the current configuration of the living room, whether modified by the virtual agent or the user during the co-design process, to guide the virtual agent’s furniture items placement decisions in the virtual living room environment. We compared our method (i.e., optimization-based) against two other co-design strategies (i.e., template-based and random-based) following a within-group (N = 24) study design. We found the optimization-based approach significantly enhanced perceived collaboration compared to template-based and random-based methods. Moreover, our participants attributed higher private and public awareness to the optimization-based virtual agent. Our analyses of user activity logged data revealed that participants placed more furniture items and made fewer corrections to virtual agent placements in the optimization-based condition. Our findings demonstrate that adaptive virtual agent behavior that dynamically responds to user actions while maintaining spatial coherence created more effective collaborative experiences in immersive design environments.