Toyota Research Institute (TRI) has introduced an innovative generative artificial intelligence (AI) tool designed to boost the creative process of vehicle designers. The tool enables designers to generate design sketches by providing text prompts that incorporate specific stylistic attributes, such as “sleek,” “SUV-like,” and “modern.” Additionally, designers can optimize quantitative performance metrics to create initial prototype sketches.
This advancement empowers designers to explore their creativity while ensuring efficient and effective design development. The technique developed by TRI researchers combines optimization theory principles from computer-aided engineering with text-to-image-based generative AI models. As a result, designers can optimize engineering constraints while preserving their text-based stylistic prompts, allowing for the generation of images that adhere to both aesthetic and engineering requirements.
The integration of vehicle constraints, such as drag, chassis dimensions, and more, enhances the image generation process. By explicitly incorporating and optimizing over-engineering constraints, designers can strike a balance between aesthetics and engineering considerations more quickly and efficiently.
The goal of this project is to expedite the creation of new designs by streamlining the trade-offs between aesthetics and engineering. The addition of constraints to generative AI tools provides guide rails for the AI-generated designs, ensuring that the resulting designs meet engineering requirements and safety measures.
The development of this technique was motivated by the challenges faced by vehicle designers, who struggle to find inspiration for new designs and face prolonged iteration processes between design and engineering. The collaboration between designers and engineers, each with their own perspectives and modes of thinking, often leads to designs that initially fail to meet engineering requirements. The integration of generative AI into the design process aims to reduce iteration time and foster collaboration between design and engineering teams.
Toyota plans to integrate this tool into its vehicle design and development process, working alongside various teams within the company. While the technology is currently in the research phase, TRI continues to explore ways to enhance quality of life through its research endeavors.
By leveraging generative AI technology, the hope is that designers worldwide can expand the power of design ideas while accelerating the speed of design development. TRI is committed to responsibly harnessing the potential of generative AI to amplify human capabilities across multiple research areas.