2016 Joint Telematics Group/IEEE Information Theory Society Summer School

on Signal Processing, Communications and Networks.

IISc Bangalore, June 27 - July 01, 2016.

** Venue: Golden Jubilee Hall, Department of ECE**

Random quantities that `arise in practice' are not `violently random'; they remain pretty close to their mean value with overwhelming probability. Quantitative versions of this statement are `concentration inequalities'. Such inequalities can be used --- to great advantage --- in understanding, analysing and controlling these quantities. These inequalities have applications in diverse areas: graph theory, computer science, statistics, biology, spin glasses and so on. Starting at basic level, this course is aimed at understanding some of these techniques, results and their use.

mmWave represents the next frontier in wireless communication, providing "effectively unlimited" spectrum for short to medium range networks, due to the huge amounts of available spectrum and the aggressive spatial reuse enabled by highly directive links. Similarly, short-range mmWave radar is a key enabler for sensing applications such as gesture recognition and vehicular situational awareness. mmWave systems differ fundamentally from existing wireless systems because of the order of magnitude smaller carrier wavelength, and the order of magnitude higher available bandwidth. In this mini-course, we discuss some of these key differences and their design consequences, ranging from hardware/signal processing co-design to network protocols.

Polar coding is a method that can achieve the Shannon limits in a wide range of source and channel coding scenarios. The goal of this mini-course is to give an in depth coverage of the fundamentals of polar coding, as well as discussing a range of selected theoretical and practical topics. In particular, a comparison of polar coding with other codes for various application scenarios will be presented.

- Overview and Tchebycheff
- Cramer, Chernoff, Hoeffding-Azuma
- Cramer, Chernoff, Hoeffding-Azuma continued
- Effron-Stein and McDiarmid (bounded differences)
- Stein and Chatterjee
- Entropy, log-Sobolev, and Herbst
- Entropy, log-Sobolev, and Herbst continued
- Talagrand

- Overview: mmWave characteristics, concept systems, technical challenges
- The mm wave channel: diversity, multiplexing, blockage
- Steering large arrays: compressive and scan-based approaches
- Fundamentals of compressive estimation: theory and algorithms
- Networking with highly directional links
- Networking (contd.)
- Signal Processing at high bandwidths
- Short-range mmWave radar

- Overview of polarization and polar coding
- Encoding, decoding and code construction
- Polarization details: Martingale analysis, rate of polarization
- Density evolution and polar code construction
- Polar codes in practice: performance, complexity of implementation, comparisons with other codes
- Generalizations: Non-binary codes, non-uniform inputs, larger kernels, universal constructions
- Polar coding in multi-user settings: multiple-access, degraded broadcast, Slepian-Wolf, Wyner-Ziv, Gelfand-Pinsker, etc.
- Open problems, discussion, and topics for future research