Elnaz Tafrihi on Insight XR
We're thrilled to support our students in their summer research. Read about Elnaz Tafrihi's research into InsightXR below!
In the summer of 2023, with the help of the BCNM summer research grant, I worked on advancing my dissertation and writing a paper related to my PhD research. In this paper, a case study is proposed to test the capabilities of InsightXR with respect to applying Operative Generative Design in augmented reality during early-stage massing studies. In addition, the proposed case study was used to assess a k-means clustering method for visualizing selected design options. This case study explored using the operative generative design using the NSGA-II algorithm by generating design solutions for early-stage massing studies for a mixed-use building on a site located at 74 Mission Rock St, San Francisco, CA 94158. In this case study, the objectives related to rule-of-thumb sustainable design and early-stage design are used.
Design operations explored in this case study are Carving and Expansion. In each generation, a selection of design alternatives (solutions) is presented to the users in augmented reality. At the end of the process, the Pareto Front can be visualized for further investigation of the solution space. In order to reduce user fatigue, three design alternatives are selected per generation using k-means clustering to be shown to the users.
The applied algorithm was able to successfully cluster each generation’s solutions and select a representative case among the solutions to be viewed by the users in InsightXR. In addition, InsightXR’s application was able to successfully view each generation’s selected design solutions in real-time. The proposed generative design process can also be non-realtime to allow users to visualize design alternatives.
Future research can investigate the inclusion of user feedback in the optimization process using an interactive NSGA-II algorithm in InsightXR. In addition, research needs to be conducted to understand the perception of users regarding InsightXR and understand the opportunities and pain points of using InsightXR. In addition, this paper selected five generations as the termination criterion for the algorithm, which can also be tuned or defined with respect to other criteria in future research.