用Transformer自动生成能精准画出指定轨迹的机械结构
MechaFormer: Sequence Learning for Kinematic Mechanism Design Automation
- 将机械设计转为条件序列生成，一次产出拓扑与参数
- 路径匹配准确率超现有方法，还能生成多样新结构
- 输出可作优化起点，大幅提升传统设计效率
设计能够精确追踪特定轨迹的机械机构是经典但极具挑战性的工程问题，其搜索空间庞大且复杂，包含离散的拓扑结构和连续的几何参数。我们提出MechaFormer，一种基于Transformer的模型，将机制设计视为条件序列生成任务。该模型学习将目标曲线转化为领域专用语言（DSL）字符串，一次性确定机构的拓扑结构和几何参数。MechaFormer显著优于现有基线，达到最先进的路径匹配准确率，并生成大量新颖且有效的设计方案。我们展示了一系列采样策略，可显著提升解的质量并为设计者提供灵活选择。此外，MechaFormer生成的高质量输出可作为传统优化器的良好初始点，形成混合方法，在极短时间内找到更优解。
原文摘要 · Abstract (English)
Designing mechanical mechanisms to trace specific paths is a classic yet notoriously difficult engineering problem, characterized by a vast and complex search space of discrete topologies and continuous parameters. We introduce MechaFormer, a Transformer-based model that tackles this challenge by treating mechanism design as a conditional sequence generation task. Our model learns to translate a target curve into a domain-specific language (DSL) string, simultaneously determining the mechanism's topology and geometric parameters in a single, unified process. MechaFormer significantly outperforms existing baselines, achieving state-of-the-art path-matching accuracy and generating a wide diversity of novel and valid designs. We demonstrate a suite of sampling strategies that can dramatically improve solution quality and offer designers valuable flexibility. Furthermore, we show that the high-quality outputs from MechaFormer serve as excellent starting points for traditional optimizers, creating a hybrid approach that finds superior solutions with remarkable efficiency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。