查看: 377|回复: 0

[招生信息] 瑞典于默奥大学全奖博士生(机器学习与嵌入式控制系统) [复制链接]

Rank: 2

声望
50
寄托币
62
注册时间
2014-4-21
精华
0
帖子
18
发表于 2019-2-5 20:47:28 |显示全部楼层
于默奥大学(Umea university)是瑞典北方的综合性旗舰(flagship)大学,学生总数为33000。2018年软科世界大学排名为瑞典第10名。博士研究课题为机器学习,控制论与实时嵌入式系统的交叉研究,以及在自动驾驶领域的应用。不需要gre/托福/雅思成绩。申请截止期为2019-02-28。此职位仅针对研究生同学,因为瑞典的博士入学前需要有硕士学位。(明年上半年毕业的硕士研究生也可以申请。) 欢迎大家在以下网址提交申请。如有其它问题请联系导师邮箱 zonghua.gu@umu.se

https://www.umu.se/en/work-with- ... rol-systems_249606/
we offer a phd student position in applied electronics with focus on machine-learning-enabled embedded control systems. apply by 2019-02-28 at the latest.

description of the phd-project
an embedded control system feature closed-loop, real-time interactions between an embedded computing system and a physical environment, and tight integration of computing, communication and control. embedded control systems are often safety-critical systems that are subject to safety certification, e.g., iso 26262 certification for passenger cars. while traditional control theoretic techniques, including pid control, model-predictive control, optimal control, have mature, rigorous design and analysis techniques that are capable of achieving high assurance, a novel class of control systems that include machine-learning algorithms in the loop are increasingly prevalent, e.g., deep reinforcement learning, with its many and diverse applications ranging from alphago to self-driving cars. in these and many other important applications, the control algorithms need to be implemented on embedded computing platforms with limited hardware resources due to cost and power constraints. therefore, the system designer needs to make careful tradeoffs among multiple design objectives, incl. performance, cost, power, security, etc. this ph.d. project aims to tackle the problem of how to achieve safety certification with machine learning in the loop, with main applications in self-driving cars and other potential applications of mobile robotic systems. this project addresses both design-time assurance with formal verification techniques, and runtime assurance with monitoring and enforcement of safety constraints.

highly desirable knowledge and skills include: machine learning algorithms, incl. reinforcement learning and deep learning, as applied in embedded control systems; formal verification techniques, incl. model-checking, sat modulo theories, milp; real-time embedded systems, incl. real-time scheduling algorithms and design optimization techniques. knowledge of control theory or hands-on experience working with robotic platforms, is an advantage but not required.

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

关闭

站长推荐

寄托飞跃大礼包
寄托飞跃大礼包——回馈在寄托上报19fall offer的同学们 内含【寄托独家纪念品】X 5 、 【所去地区的手机卡】、 【全球电源转换插头】还有很多......自己开箱看吧!

查看 »

报offer 祈福 爆照
回顶部 我要纠错