Talk by Vijaya Yajnanarayana, Ericsson AI Research, Bangalore

Location: Auditorium, ECE MP Building

he Department of ECE welcomes you to a talk on April 8, 2019.

Title: AI meets 5G: Connectivity and Mobility for Next Generation Wireless Networks

Speaker: Vijaya Yajnanarayana, Ericsson AI Research, Bangalore

Date: Monday, 8th April 2019

Time: 2.30 PM (2:15 Tea/Coffee)

Venue: New Auditorium, MP Building, ECE


The next generation wireless networks have to co-exist with  machine type devices like drones, vehicles and IoT devices. This creates unique challenges for Radio Access Network (RAN). In this talk, I will discuss aspects related to providing cellular connectivity to drones.  One of the main challenge in providing connectivity to the low altitude unmanned aerial vehicles (UAVs) through existing cellular network arises due to the increased interference in the network. The increased altitude and favorable propagation condition cause UAVs to generate more interference to the neighboring cells, and at the same time experience more interference from the downlink transmissions of the neighboring base stations. I will discuss few interference mitigation strategies for hybrid cellular network providing connectivity to both aerial and terrestrial UEs. One can fly illegal drones to cause excessive interreference in the network thereby generating denial-of-service attacks. How to detect such drones is of interest and we will look at machine learning approach for the same. Next, we will shift the discussion towards mobility aspects in 5G. To support increased capacity requirement and to enable newer use cases,  5G networks will have a very dense deployment with nodes having advanced beam-forming capability. In such networks, providing a better mobility along with enhanced throughput performance requires an improved handover strategy. We will discuss a  novel method for handover in a 5G context using reinforcement learning (RL). In contrast to the conventional method, we propose to control the handovers between base-stations (BSs) using a centralized RL based machine learning (ML) agent. This agent handles the radio measurement reports from the UEs and choose appropriate handover actions in accordance with the RL framework to maximize a long-term utility.  We pose the handover mechanism as a contextual multi-armed bandit problem and  analyse the performance using different  propagation environments and compare the results with the traditional algorithms.

Speaker Bio: Vijaya Yajnanarayana received his M.S. degree in electrical engineering (EE) from the Illinois Institute of Technology, Chicago, USA and Ph.D. degree in EE from the KTH Royal Institute of Technology, Stockholm, Sweden. From 2001 to 2009, he worked in Motorola research and Motorola Signal Processing Departments, at Chicago and Bangalore. From 2008 to 2009, he served as a senior specialist in Nokia Siemens Networks (NSN). From 2009 to 2012, he served as a principal engineer at National Instruments (NI). In 2017, he joined Ericsson Radio Research Lab in Stockholm, Sweden. He is a recipient of the Program of Excellence Award from KTH Royal Institute of Technology, which carried an award of Kr. 1 million. He is also the recipient of best research paper award by Motorola technical journal and IEEE ICEMI conference in 2007 and 2011, respectively. He has contributed to more than 15 patents and more than 15 research articles in the area of communication and signal processing. Currently, he is working as a senior researcher in Ericsson AI Research Group in Bangalore, India.


Scroll Up