【GAMES Webinar 2025-384期】
软体机器人仿真与可持续制造
· 1 ·
报告题目
Design for Sustainable Fabrication
报告嘉宾
Yuxuan Mei
Wesleyan University
报告时间
2025年11月6号 晚上20:00-20:30（北京时间）
报告方式
GAMES直播间:
https://live.bilibili.com/h5/24617282
报告摘要
Design tools play a crucial role in how effectively the advances in manufacturing can be utilized to create products of unprecedented complexity. The usability of these tools directly impacts who can design, customize, and ultimately enhance the performance of products. Meanwhile, the revolution in manufacturing is fundamentally changing what can be made and by whom, enabling on-demand production of complex and functional objects. Despite its ever-increasing potential, manufacturing continues to consume considerable natural resources and produce significant amounts of waste and carbon emissions. This raises a critical research question: how can we foster sustainable fabrication practices in this rapidly evolving landscape? In this talk, we address the challenge by developing computational design systems that make it easier for users to create sustainable designs, thereby reducing material usage and waste at the source. Our approach centers on creating domain-specific abstractions for design systems that enable more effective exploration of the design space considering sustainability and expression of users’ design intent. We briefly show how this approach can be applied to various fabrication domains, including electronics, home accessories, and textiles.
嘉宾简介
Yuxuan Mei is an Assistant Professor of Computer Science at Wesleyan University. She received her B.S. degree in computer science from Columbia University and M.S. and Ph.D. degrees in computer science & engineering from the University of Washington. Her research interests are computational design, fabrication, and simulation, with the goal of creating more sustainable designs that take less material and machine cost to fabricate.
个人主页
https://merlinyx.github.io/
· 2 ·
报告题目
Sim2Real of Soft Robot Simulation
报告嘉宾
高俊朋
山东大学
报告时间
2025年11月6号 晚上20:30-21:00（北京时间）
报告方式
GAMES直播间:
https://live.bilibili.com/h5/24617282
报告摘要
软体机器人在仿真环境中难以精确建模，导致“仿真到现实”(sim-to-real)的迁移存在障碍。为应对这一挑战，我们提出了一种基于学习的修正策略。该策略的核心是利用神经网络对仿真模型与现实世界之间的残余物理(residual physics)进行补偿，进而提升仿真模型的精度。我们在软体悬梁臂和气动软体机械臂上进行了实验验证，仿真误差减小了超过30%。
嘉宾简介
高俊朋，本科毕业于上海科技大学，硕士毕业于苏黎世联邦理工学院（ETH Zürich），即将前往卡耐基梅隆大学，师从 Prof. Andrew Spielberg 攻读博士学位，研究方向为 Computational Design。目前在山东大学 IRC 交叉研究中心作为 Research Intern。
个人主页
https://junpenggao233.github.io/
主持人简介
赵海森
山东大学
赵海森，教授，博士生导师，山东大学杰出中青年人才，国家级青年人才。主要研究方向为计算机图形学、智能设计与制造、数字孪生。曾任美国华盛顿大学博士后，奥地利科学技术研究所博士后。主持海外优青项目、重点研发计划课题、基金委面上项目等国家级科研项目。作为项目骨干，参与科技部新一代人工智能 (2030) 项目，基金委区域联合重点基金项目等项目。曾在ACM TOG、ACM SIGGRAPH/Asia、IEEE TVCG等顶级期刊或会议上发表论文十余篇，申请十余项国家发明专利和一项美国发明专利，出版专著1部。获山东省自然科学奖一等奖、“CCF优秀博士学位论文奖”等荣誉。
长按二维码，关注公众号
GAMES网站: http://games-cn.org