Co-operating clusters based CoMP for high capacity dense small-cell networks and understanding mobile user outages: predictive analytics in wireless broadband networks

Date: 
14.11.2012 10:15
Place: 
SÄ110

 

Infotech Oulu Lecture Series

Co-operating clusters based CoMP for high capacity dense small-cell networks and understanding mobile user outages: predictive analytics in wireless broadband networks

Lecturer: Dr. Shirish Nagaraj, Nokia Siemens Networks

Date: November 14, 2012
Time: 10:15
Room: SÄ110

Abstract

PART I (30mins): Co-operating Clusters based CoMP for High Capacity Dense Small-Cell Networks

A key challenge to next generation cellular networks is to provide high capacity for data-intensive applications in hot-spots and events, which include stadiums, shopping malls, and dense urban areas. Cellular operators are looking at deploying dense networks of small cells in a selective manner to provide service in such high demand regions. Capacity of such deployments is inherently limited by inter-cell interference, making it challenging to scale capacity as a function of cell density in a cost-effective manner. Co-operating transmission and reception mechanisms, commonly referred to as co-ordinated multi-point (CoMP), is a critical technology to solve the capacity bottleneck arising in these dense radio access networks (D-RANs). In this talk, we propose techniques for high capacity uplink and downlink CoMP using the notion of co-operating clusters with multi-antenna processing, interference cancellation and interference alignment. We present architectures and algorithms allowing a different mix of centralized and distributed processing to enable co-operating cluster processing.  Extensive system simulation results shed key insights into the design of such deployments and algorithm performance, as well as show that such methods have the capability to provide high capacity and uniform user experience in these dense cell deployment areas.

PART II (15mins): Understanding Mobile User Outages: Predictive Analytics in Wireless Broadband Networks

In the second part of the talk, we will briefly discuss work in the area of machine learning for communications networks. Today's Telecom service providers are rapidly evolving their network, adding 4G/broadband capabilities to satisfy the ever increasing data demand from an exponentially growing smart phone user population. In this work, we focus on case study relating to service availability and continuity for a major 4G customer in a live mobile broadband network, using machine learning algorithms. We set up a supervised learning problem to predict customer outages utilizing field data in such a mobile network. Results using non-linear regression and ensemble techniques are shown to provide key variable importance measures and further suggest concrete recommendations for improving network performance and user experience.

Biography

Shirish Nagaraj, Ph.D.

Dr. Shirish Nagaraj is a research manager in the Technology and Strategy (CTO) department at Nokia Siemens Networks. He leads a group responsible for advanced technology innovations in the areas of statistical signal processing, cross-layer optimization algorithms, and machine learning/data analytics.

Dr. Nagaraj received his Ph.D. in Electrical Engineering from the University of Notre Dame in 2000. Prior to this, he did his B. Tech. in Electrical Engineering from the Indian Institute of Technology (IIT) Bombay in 1995. He also has a Graduate Certificate in Statistical Learning and Data Mining from the Statistics Department of Stanford University.

Dr. Nagaraj was with Motorola, Inc. from 2007-2011, and before that, with Bell Laboratories-Lucent Technologies (now Alcatel-Lucent) from 2000-2007, in their research and advanced technology organizations. He is a recipient of the “Best Paper Award” at the IEEE WCNC conference in 2012, the “Bell Labs Presidents’ Gold Award” at Lucent Technologies in 2005, and the “Best Researcher” award from the University of Notre Dame in 2000.

More information: Markku Juntti

Last updated: 13.11.2012