Body Scan Processing, Generative Design, and Multiobjective Evaluation of Sports Bras

Sports bras are critical to the comfort and performance of female athletes, yet mechanical models of sports bras are generally not used to guide their design. Typically, assessing any sports bra's performance requires time-consuming and expensive biomechanical testing, which limits the number of designs considered. To more broadly advance knowledge on how different design properties of sports bras affect their performance, this paper presents a new design framework to explore and evaluate the sports bra design space. The framework incorporates methods for body scan analysis, fast simulation, design generation, and performance evaluation. Using these methods together enables the rapid exploration of hundreds, or thousands, of designs-each one having been evaluated on key metrics related to sports bra performance, namely, range of motion and average pressure. With this framework, designers can potentially discover a diverse set of new, high-performing sports bra concepts, as well as gain insights into how design decisions affect performance.