Nilanjan Banerjee photo photoUMAMAHESWARI C DEVI photo
Vinay Kolar photoAPURVA KUMAR photoRamasuri Narayanam photo

Research Areas

Group Name

Telecom Research

Tab navigation

Telecom Research
The increasing digitization of almost everything around us and the proliferation of mobile and ubiquitous devices connected to the Internet offers significant opportunities for telecommunications providers. But their ability to capitalize on this potential has been a point of contention and debate. Telecom providers clearly need upgraded networks and innovative technology platforms to handle more sophisticated content in order to effectively address their target markets. But equally important, they must begin delivering value beyond just access – providing a step change in consumer experience, smarter network optimization and data monetization. To cater to the emerging trends in the Telco domain and help our clients, while advancing state-of-the-art in relevant research areas, the Telecom Research group explores problems in several directions below:

  1. Telecom and Network Analytics

    The focus here is on deriving insights from the vast amounts of network and consumer data traversing the cellular networks for various purposes such as improving customer experience and customer relationship, network optimization and administration, and creating new revenue streams and differentiation from competitors.  The group has been conducting research on and developing advanced algorithms for network optimization and custmer experience management including (1) targeted base station upgrades based on mobile user trajectories and persistent failure analysis, (2) choice of femto-cell or WiFi offload vs. macro-cell upgrades, (3) spatio-temporal load skew analysis and cell-on-wheels provisioning at different times of day, and (4) inter-base station handoff analysis and narrowing-down coverage hole locations. On the new revenue streams creation front, the group has been working on mining mobility patterns and data usage profiles of people from CDR data, and identifying interest profiles associated to user hangouts, which can be used to enrich customer profiles.
  2. Telco Big Data Platforms and Architecture

    The data that is at the disposal of telecom providers for drawing actionable insights is from varied sources (e.g., XDR, BSS, OSS, Network, Operational) and requires a combination of batch, real-time streaming, and interactive analytics. A large amount of data movement, pre-processing such as mediation, enrichment, and de-duplication is common to any telecom analytics effort, for which a variety of big data tools and technologies such as Hadoop, Flume, and Spark can be utilized. The focus of the Telco Big Data Platforms and Architecture group is to establish a scalable, reference architecture that can maximize reuse across multiple telecom analytics efforts by (1) factorizing the data-movement and pre-processing aspects so that each of the analytics modules can focus more exclusively on the specific analysis they aim to deliver and (2) accommodating multiple suppliers of analytics in a seamless manner.
  3. Internet of Things (IoT)

    IoT enabled smart dwelling places are expected to proliferate significantly in the near future, particularly in the context of wellness living, patient healthcare and elderly care. Being able to quickly make sense of this data (by detecting high level events from sparse/noisy low level sensor event streams) and transforming these insights into measurable actions is critical to RoI of such deployments.  With growing aging population in several countries, it has become essential to mine and detect activities of daily living with the goal of offering digital care services to the elderly. The work in this stream is currently focussed on the following: (i) mining personalized signatures that capture the heterogeneous nature of activities of daily living in a large population, (ii) real-time detection of activities from a stream of sensor events, (iii) Automated techniques for sensor fault detection and intelligent maintenance scheduling.


    Past Projects (Selected)

  1. Hybrid Network Architectures

    Leveraging computation and storage “at the edge” on small cells, WiFi offload points, micr-cells and macro-cells, our research solutions enable deploying contextual services, analytics, and intelligent delivery mechanisms at the edge itself. For example, for contextual services, a consumer walking into a cricket stadium could be specifically targeted with player statistics or replays relevant to the scheduled match based on association with the small cell deployed within the stadium. Similarly, a consumer waiting at a movie hall can be sent movie trailers of of the same genre that the user is going to view.
  2. M2M Infrastructure

    Device and traffic management techniques for handling the enormous growth of M2M devices on Telco networks in the future. On the device management side, our focus has been on specification languages for device on-boarding, rules and methodologies for semantic enrichment of data, and novel ways of authorized access across devices without man-in-the-middle. On traffic management side, our focus has been on flexible scheduling and signaling mechanisms to ensure that M2M data can be carried on Telco networks with least overhead.
  3. Asynchronous Time-Shifted Delivery over Cellular Networks (Async)

    Providing asynchronous delivery of heavy content such as videos smooths network traffic. Typically, network traffic fluctuates significantly with phases of overload and underload. At base stations, these fluctuations occur at fine timescales. Async is a solution to enable delayed delivery of certain kinds of network traffic. Async results in (1) smoothened network traffic resulting in delayed network capacity upgrades, (2) user selected delivery of transfer mode for graceful degradation of quality, and (3) enablement of new business for Telcos in markets with price-conscious users.