Cad agents build mechanisms that actually work as intended
Can It Actually Work? Synthesizing Functional CAD Mechanisms
Computer Vision and Pattern RecognitionComputation and LanguageGraphics
Summary
Designing computer-aided mechanisms that function correctly is difficult because many designs can look right but fail to perform their tasks in reality. The authors study how AI agents can create mechanisms that really work by testing them in simulations tailored to specific functions and improving designs based on feedback. They built a framework called FunCAD-Harness to guide this process and introduced a benchmark, FunCADBench, to measure performance on complex mechanism families. Their method finds subtle mechanical problems that visual checks miss, helping AI better turn human ideas into working machines.
What this means in practice
- •For mechanical engineers: Generate and verify physically functional mechanisms from vague specifications using AI and simulation-guided design improvements.
- •For industrial designers: Create prototypes of mechanisms that meet functional goals without costly physical iterations by simulating and revising designs automatically.
Authors
Wenqian Zhang, Yangyi Huang, Zigeng Xu, Weiyang Liu, Zhen Liu
Abstract
While LLM agents can generate visually plausible CAD assemblies, these designs often fail to fulfill their intended physical functions. We study functional CAD design, a setting where agents interpret underspecified requests to construct and iteratively revise mechanisms that satisfy functional requirements in physical simulation. Verifying these requirements is nontrivial: different candidate mechanisms may realize the same function through entirely different components, controls, and operating sequences. To bridge this gap, we propose FunCAD-Harness, a framework that translates abstract requirements into candidate-specific physical tests and uses simulation feedback to guide design revisions. To systematically evaluate this capability, we introduce FunCADBench, a benchmark spanning 46 mechanism families with interdependent and multi-step operations. Experiments across frontier LLM agents show that FunCAD-Harness exposes subtle mechanical failures invisible to visual inspection, substantially improving functional success rates over standard simulation-augmented baselines. Ultimately, our work marks a step toward CAD agents capable of translating human intent into working mechanisms.