SBIR Phase I: Artificial Intelligence Enhanced Design Automation for General Engineering Systems

小型企业创新研究(SBIR)第一阶段项目的更广泛影响/商业潜力是改进需要多物理模拟的先进半导体产品的原型和验证流水线。这将影响多芯片模块、新兴半导体器件和新型模拟电路的设计，所有这些都可能构成下一代电子产品的基础。目前半导体公司使用的电子设计自动化(EDA)软件往往无法用于设计这些技术，因为EDA往往缺乏足够的模块化和详细的模型，如热物理和机械物理。该项目将开发一种新的EDA软件工具集来满足这些需求，使设计师能够大幅降低开发成本/时间，并为目前设计难度太大的新产品打开大门。这个小型企业创新研究(SBIR)第一阶段项目将开发物理建模算法，将新物理连接到电子设计自动化(EDA)管道中。现有的元件建模工具，如最先进技术中的晶体管，要么使用缓慢的有限元分析，要么使用高度专业化的紧凑模型。通过开发新的算法将复杂的物理转化为高效的模块模型，这些限制将被克服。这个项目将产生一个基于新算法的EDA工具，(1)无需手动调整即可推广到广泛的组件，(2)具有高精度，(3)执行速度快。该EDA工具将集成到EDA工具链中，保持与现有设计和优化软件的互操作性。集成EDA工具的性能将根据运算放大器的电路模拟和多芯片模块的有限元分析模拟进行验证，以演示快速和准确的系统评估。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to improve the prototyping and verification pipeline for advanced semiconductor products that require multiphysics simulations. This will impact the design of multi-chip modules, emerging semiconductor devices, and novel analog circuits, all of which are likely to form the basis of next-generation electronics. Current electronic design automation (EDA) software used by semiconductor companies often cannot be used to design these technologies as EDA often lacks sufficient modular and detailed models, such as thermal and mechanical physics. This project will develop a new EDA software toolset to address these needs, allowing designers to dramatically reduce development cost/time and open the door to new products that are currently too challenging to design.This Small Business Innovation Research (SBIR) Phase I project will develop physical modeling algorithms to link new physics into the electronic design automation (EDA) pipeline. Existing modeling tools for components, such as transistors in state-of-the-art technologies, employ either slow finite element analysis or highly specialized compact models. These limitations will be overcome by developing new algorithms to convert complex physics into efficient modular models. This project will produce an EDA tool based on the new algorithms that (1) are generalizable to a wide range of components without manual tuning, (2) have high accuracy, and (3) execute fast. This EDA tool will be integrated into the EDA toolchain, retaining interoperability with existing design and optimization software. The performance of the integrated EDA tool will then be validated against circuit simulations of op-amps and finite element analysis simulations of multi-chip modules in order to demonstrate fast and accurate system evaluation.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.