01 Sep

IEEE Distinguished Lecture Tour

The IEEE joint VT/COM/IT Sweden Chapter Board, and the ACCESS Linnaeus Center, are delighted to invite you to an IEEE Distinguished Lecture by Prof. Urbashi Mitra, University of Southern California, USA. Prof. Mitra will be visiting KTH, Ericsson in Kista, Uppsala University, Chalmers University of Technology, and Linköping University according to the agenda below.

The lectures will be held on:

Wednesday September 2nd in Stockholm, at 13.30 hosted by the ACCESS Linnaeus Center,

Location: Lecture Theatre K53, Teknikringen 28, floor 5, KTH City Campus; https://www.kth.se/places/room/A43:15/585

Title: Biological Communication Systems:  Engineered and Natural

Abstract: See below.


Friday September 4rd in Uppsala, at 13.15 hosted by Signal and Systems

Location: Room 1211 (floor 2, Building 1) ITC, Polacksbacken, Lägerhyddsvägen 2, Uppsala

Title: Biological Communication Systems:  Engineered and Natural

Abstract: See below


Monday September 7th in Gothenburg, at 13.00 hosted by the Communications Systems Division

Location: Lecture hall ED, EDIT building, Chalmers campus Johanneberg http://maps.chalmers.se/#3bc54704-a690-4476-b886-3a39429f9612

Title: Biological Communication Systems:  Engineered and Natural

Abstract: See below


Tuesday September 8th in Linköping, at 10.15 hosted by Communication Systems

Location: Algoritmen

Title: Wireless channel Estimation: Opportunities for Exploiting Structure and Sparsity

Abstract: See below



Biological Communication Systems:  Engineered and Natural

The prospect of biological communication systems is nearing a reality.  Such systems can be viewed from two perspectives:  the construction of engineered biological communication systems exploiting recent advances in nano-machines or the investigation of existing natural systems wherein a communication framework can enable design, control and enhanced understanding.  In this talk, both perspectives will be examined.  In molecular communication, information is conveyed via molecules versus electromagnetic waves.  In particular, diffusion-based molecular communication most closely remembers conventional wireless communication.  Diffusion-based systems do not require prior communication link infrastructure, but do depend on the presence of transmitting and receiving nano-machines. Several modulation and transceiver designs for molecular communications will be reviewed and the capacity of such links examined.  Properties of diffusion and the inherent inter-symbol interference are taken into consideration resulting in novel channel models that necessitate new designs and analysis.  Such systems have application in health-care and in-body drug delivery systems.  From the second perspective, we examine the modeling of microbial communities.  In particular, we examine the electron transfer mechanism in living cells and its role in cell-to-cell interaction.  We describe a new queueing theoretic model for the internal workings of a bacterium as well as methods based on statistical physics to scale up the queuing models. Preliminary experimental comparisons show a good fit for the model. Our goal is to couple modeling with experiment to optimize the design of microbial fuel cells which are a very promising renewable energy source.   Furthermore, we adapt our initial model in order to develop capacity results for bacterial cables which are akin to multi-hop networks in wireless communications.  The ultimate goal is to develop methods for three-dimensional biofilms.


Wireless channel Estimation: Opportunities for Exploiting Structure and Sparsity

Wireless communication systems typically require some form of channel state information in order to provide high performance.  Traditionally, wireless channels are modeled as linear systems, which could be time-varying depending on the communication scenario.  With the advent of very wideband communication and high speed applications, channel estimation becomes more challenging.  Herein we show how exploiting structure inherent in many wireless communication channels can overcome challenges introduced by modern wireless applications. In many wideband signaling scenarios, channels can be modeled as sparse.  To be truly practical, one must consider the effects of practical channels (not purely sparse) and transceiver characteristics such as bandlimited pulse shapes in order to design highly accurate channel estimation.  We propose hybrid channel models suitable for ultrawideband radio and underwater acoustic systems based on both sparse and diffuse components and provide asymptotic performance analyses.  These hybrid models can be extended to time-varying channels. We show that vehicle-to-vehicle channels, in particular, show further structure in the form of both sparse specular components and groups of diffuse components.  A new channel estimation based on a nested thresholding algorithm is shown to be optimal for this channel structure and offers strong performance improvement over previous methods.  Finally, we examine truly wideband channels where mobility induces Doppler scaling (versus Doppler shifts as approximated in narrowband systems).  We show that OFDM signaling after passing through such a multi-scale, multi-path channel has a low rank representation.  This feature can be employed to improve robustness; we eventually pose the channel estimation problem as a structured spectral estimation where sparsity can be exploited with classical spectral techniques.  Strong performance gains are achieved over previously proposed methods.