• 学术动态

5月11日学术报告预告
作者:fanglei 发布日期:2017-05-09 浏览次数:

报告题目:计算机智能系统系列报告

报告时间:2017511日,1520

报告地点:郁文楼A102会议室

报告人:李强

报告人简介:

李强,博士,讲师,目前主要研究方向为无线传感器网络。20147月获得清华大学工学博士学位(导师:管晓宏教授),同年11月在浙江工业大学计算机科学与技术学院任教。参与浙江省科技厅公益技术项目1项;以第一作者或通信作者身份在《Peer-to-Peer Networking and Applications》、《IEEE International Conference on Communications》等国内外学术刊物与会议中发表多篇论文。

 

报告内容:An Enhanced MAC Protocol for Terahertz Communications through Relay Nodes with Multiple Antennas

Directional terahertz (THz) communication has been envisioned as one of the key technologies to support the high data rate for indoor wireless networks. However, on the one hand, THz communication distance is extremely limited by the severe path loss as a result of the high frequencies and the constrained transmission power at the transceivers. On the other hand, static obstacles or moving objects around the communicating THz nodes will significantly interrupt or block the Line-of-Sight (LoS) propagations, which necessitates new communication protocols. In this paper, an enhanced MAC protocol is presented for indoor THz networks to overcome the effect of obstacles and to improve the communication distance. The MAC protocol is designed to guarantee the communications by redirecting the antennas of transceivers to solve the facing problem, and utilizing the appropriate relay node to transfer the data across the obstacles. After that, a corresponding mathematical framework is developed to analyze the performance of the proposed MAC protocol in terms of packet delay, throughput and its theoretical upper bound. Numerical results are provided to evaluate the performance of the proposed protocol under different scenarios and are compared with the existing THz MAC protocols.


报告题目:计算机智能系统系列报告

报告时间:2017511日,1520

报告地点:郁文楼A102会议室

报告人:汪明梁

报告人简介:

汪明梁,博士,主要研究方向为数据建模与机器学习及其在锂电池中的应用。201711月获得香港城市大学哲学博士学位(导师:李涵雄教授),目前在计算机科学与技术学院任教。以第一作者身份在IEEE Transactions on Industrial Electronics》、《IEEE transactions on SMC: Systems》、《Journal of Power Sources》、《International Joint Conference on Neural networks》、《IEEE conference on SMC等国外学术刊物与会议中发表论文5篇。

 

报告内容:电池内部状态的自适应时空动态建模研究

电池内部状态的监测是维持电池健康与安全的重要基础。在复杂环境下,对电池内部状态的估计是该领域的前沿之一,其主要特点有以下三个方面:(1)非线性时空动态耦合;(2)内部状态的不可观测性;(3)由于电池老化,环境变化和噪音等产生的不确定性。本研究提出了基于时空分离的自适应时空动态建模方法,对电池内部状态进行实时地估计:兼顾模型的计算复杂度和精确度,使得模型能够实时的在电动汽车和航空航天交通工具中有效运行;将非线性时空动态的耦合通过时空分离解耦;结合物理机理和实验数据补偿未建模的时空动态,提高模型的精度;针对电池老化等不确定性,通过在线可测数据对模型参数和隐藏状态进行在线估计和更新,提高模型的鲁棒性;深入探讨深度学习的方法对各种非线性时空动态的低阶建模,以提高建模效率和精度。通过该研究实现了面向复杂环境的自适应时空估计方法,能广泛地应用于各类工业过程的时空分布预测、隐藏状态估计、以及故障分析与诊断等其他领域。