| 查看: 682 | 回复: 0 | |||
[交流]
2025年巴黎高科 - CSC合作公派读博项目 - 课题No.47
|
|
2025年巴黎高科 - CSC合作公派读博项目 - 课题No.47 2025 巴黎高科 - CSC公派博士项目 （12月8日截止） 网申通道：https://paristech.kosmopolead.com/phd 申请攻略：https://paristech.fr/fr/paristech-csc-phd-program-how-apply 课题47详情: TITLE: AUTOMATIC GUIDED CAD ASSEMBLY GENERATION WITH MECHANICAL CONSTRAINTS USING DEEP LEARNING Topic number : 2024_047 Field : Design, Industrialization - Mathematics and their applications Subfield: Geometric modeling ParisTech School: Arts et Métiers Research team : Research team website: Research lab: LISPEN - Laboratoire d'ingénierie des systèmes physiques et numériques Lab location: Aix-en-Provence Lab website: https://lispen.artsetmetiers.fr/ Contact point for this topic: Polette - Arnaud - arnaud.polette@ensam.eu Advisor 1: Jean-Philippe Pernot - jean-philippe.pernot@ensam.eu Advisor 2: Arnaud Polette - arnaud.polette@ensam.eu Advisor 3: Advisor 4: Short description of possible research topics for a PhD: The aim of this PhD is to implement methods for generating complex mechanical assemblies using existing databases of mechanical parts. The goal is to generate assemblies by controlling the level of coherence of the assembly according to the need (functionality, imposed interfaces, types of parts, etc.), while guiding the generation using functions to maximize certain objectives (encompassing shape, aesthetics, type of assembly, etc.). The idea would be to use databases of existing assemblies whose types and interfaces between parts are known, in order to learn the assembly logic according to the type of part and the consistency of functionality between the different geometries. Several solutions can be explored, such as reinforcement learning, in order to build an agent that can iteratively build these assemblies part by part. A second solution to explore would be to use auto-encoders and/or graphs (eg. GCN, Graph Convolutional Network). The first step will be to build a database (as the host laboratory has already carried out work on using this type of database, this part will be greatly facilitated by the existing databases), then to explore the automatic assembly methods, and to validate the operation of these methods on the databases built, by illustrating them with concrete uses in an industrial context. Required background of the student: Computer science, machine learning, geometric modeling, computer- aided design (assembly design) A list of (5 max.) representative publications of the group: (Related to the research topic) 1. Lucas Vergez, Arnaud Polette, Jean-Philippe Pernot. Multi-part kinematic constraint prediction for automatic generation of CAD model assemblies using graph convolutional networks. Computer-Aided Design, Volume 178, 2025, 103805, ISSN 0010-4485, https://doi.org/10.1016/j.cad.2024.103805. 2. Lucas Vergez, Arnaud Polette, Jean-Philippe Pernot. Interface-Based Search and Automatic Reassembly of CAD Models for Database Expansion and Model Reuse, Computer Aided Design, 2023 (online publication), Volume 167, 103630, https://doi.org/10.1016/j.cad.2023.103630 3. Lucas Vergez, Arnaud Polette, Jean-Philippe Pernot. Automatic CAD Assemblies Generation by Linkage Graph Overlay for Machine Learning Applications. Computer-Aided Design and Applications, 2022, 19(4), pp. 722–732, https://doi.org/10.14733/cadaps.2022.722-732 |
回复此楼» 猜你喜欢
各位大神，目前国内有哪些比较好用的逆合成软件？
已经有14人回复
-
国社科系统bug了，是不是要放榜了？
已经有10人回复
-
现代”学阀”该如何界定
已经有9人回复
-
课题组招2027级博士 上海工程技术大学 激光智能制造方向
已经有4人回复
-
上海工程技术大学 激光智能制造课题组 2027级博士研究生招生
已经有3人回复
-
上海工程技术大学激光智能制造课题组｜2027级博士研究生招生公告
已经有11人回复
-
申博发邮件
已经有11人回复
-
我的奶奶
已经有3人回复