Text

theses

ETD doctoral

Floorplans are a standard representation of building layouts. Computer-Aided Design (CAD) applications and existing Building Information Modeling (BIM) tools rely on simple floorplan representations that are not amenable to automation (e.g., generative design). They do not account for how people inhabit and occupy the space. These are two central challenges that must be addressed for intelligent human-aware building design and this thesis's focus. This thesis addresses these challenges by exploring graph representation learning techniques to implicitly encode the latent state of floorplan configurations, which is more amenable to automation and related applications. Specifically, we use graphs as an intermediate representation of floorplans. Rooms are nodes, and edges indicate a connection between adjacent rooms, either through a door or passageway. The graphs are annotated with various attributes that characterize the semantic, geometric, and dynamic properties of the floorplan with respect to human-centered criteria. To address the variation in graphs' dimensionality, we utilize an intermediate sequential representation (generated by random walks) to encode the graphical structure in a fixed-dimensional representation. We propose the use of RNN-based vanilla/variational autoencoder architectures to embed attributed floorplans. We enhance graph-based representations of floorplans with human occupancy attributes extracted by statically analyzing the floorplan geometry and running simulations on large datasets of real and procedurally generated synthetic floorplans. We explore the potential of our proposed methods and floorplan representations on various tasks, including finding semantically similar floorplans, floorplan optimization, and generative design. Our approach and techniques are extensively evaluated through a series of quantitative experiments and user studies with expert architects to validate our findings. The qualitative, quantitative, and user-study evaluations show that our embedding framework produces meaningful and accurate vector representations for floorplans. Our models and associated datasets have been made publicly available to encourage adoption and spark future research in the burgeoning research area of intelligent human-aware building design.

ETD_11401

Ph.D.

Includes bibliographical references

doi:10.7282/t3-jdxm-rm08

The author owns the copyright to this work.