Iterative Generative Design via Virtual Reality Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing generative design processes in computer-aided design (CAD) software are inefficient due to the manual determination of design goals and features during the early prototyping stages, requiring a predetermined objective function that is often unavailable, leading to time-consuming and tedious design generation and evaluation.
Innovation Solution
A computer-implemented method that iteratively generates designs by performing layout operations on virtual objects based on initial design constraints, modifying these constraints in a virtual reality environment based on feedback from mid-air representations, and generating new designs that achieve specified design goals without relying on a predetermined objective function.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If generative design applications execute multi-objective optimization algorithms to automatically synthesize designs, then the number of designs generated increases and design evaluation becomes more comprehensive, but the time required to define design goals and set up the objective function increases significantly
Solution Approach 1:
The system performs preliminary manual prototyping and goal determination before executing automated generative design. Designers manually create initial prototypes and determine design goals during early stages, then use these pre-established goals to guide the automated multi-objective optimization process, avoiding the need to define complex objective functions from scratch
Solution Approach 2:
The design process is segmented into distinct phases: manual prototyping and goal determination phase, followed by automated generative design phase. This segmentation allows designers to focus on high-level goal setting initially, then delegate the detailed design generation to automated algorithms, reducing overall time investment
2Adaptability or versatility
If designers manually position and orient objects to generate multiple design variations, then design flexibility and creativity are maintained, but the time required to generate and evaluate designs increases prohibitively
Solution Approach 1:
The system creates multiple design variations by copying and modifying base design elements. Instead of manually positioning each object from scratch for every design variation, the system generates copies of objects and systematically varies their positions and orientations based on the established objective function, maintaining design flexibility while dramatically reducing time investment
Solution Approach 2:
The system transitions from static manual design to dynamic automated design generation. The multi-objective optimization algorithms dynamically generate and evaluate numerous design variations based on the objective function, allowing the design process to adapt and evolve automatically while maintaining the flexibility to explore diverse design solutions
3Measurement precision
If designers manually scrutinize prototypes to identify desirable features, then accurate design goal determination is achieved, but the process becomes time-consuming and tedious
Solution Approach 1:
The system implements feedback loops where designers evaluate a limited set of representative prototypes and provide feedback on desirable features. This feedback is then used to refine and adjust the objective function, which subsequently guides the automated generation of additional design variations. This iterative feedback process maintains accuracy in goal determination while reducing overall time investment compared to manual evaluation of numerous prototypes
Data Source
AI summary
In various embodiments, a generative design application iteratively generates designs via a generative design process. In operation, the generative design application performs one or more layout operations on virtual objects based on a first set of design constraints to generate a first design. The generative design application then modifies the first set of design constraints based on feedback associated with a mid-air representation of the first design displayed in a virtual reality environment to generate a second set of design constraints. Subsequently, the generative design application performs one or more layout operations on the virtual objects based on the second set of design constraints to generate a second design that achieves design goal(s). Advantageously, enabling a designer to incrementally indicate design goal(s) as constraints via a virtual reality environment instead of as a predetermined objective function reduces both the time and effort required to generate designs.