Planning under uncertainty in robotics: theory to practice, and serial to parallel
Speaker :
Time :
Location :
报告人:蔡盼盼博士
报告题目: Planning under uncertainty in robotics: theory to practice, and serial to parallel
报告时间:2018年5月17日18:00-19:00
地点:工商管理楼4楼会议厅
摘要:
Planning under uncertainty is critical for robust robot performance in uncertain, dynamic
environments, but it incurs high computational cost. In this talk, I will introduce a principled
approach to handle uncertainties, partially observable Markov decision process (POMDP),
and present our state-of-the-art online search algorithm, DESPOT, that have vastly improved
the computational efficiency of planning under uncertainty and made it a valuable tool for
robotics in practice. I will also introduce Hybrid Parallel DESPOT (HyP-DESPOT), a
massively parallel online planning algorithm that integrates CPU and GPU parallelism. In
HyP-DESPOT, we demonstrate how to achieve near real-time online planning performance
for challenging robotic tasks, by leveraging parallelization.
报告人介绍:
蔡盼盼博士2011年毕业于浙江大学数学系应用数学专业,2016年毕业于新加坡南洋理工学院(NTU)获博士学位,
2017年入职新加坡国立大学(NUS)计算机系,从事研究工作(Research Fellow)。她的研究主要集中在人工智能
和机器人领域。
联系人:王何宇(wangheyu@zju.edu.cn)
欢迎广大师生参加!