Our Keynote Speakers

Abstract:

Understanding the spectrum characteristics of an environment has been central to several areas ranging from network planning and deployment to coexistence between heterogeneous technologies. The advances in computing and AI has led to sophisticated models for such spectrum characterization under the broader umbrella of “digital spectrum twins”(DSTs). In this talk, I will highlight the importance and benefits of going beyond radio frequency (RF) in the creation of such DSTs through two case studies relating to the “seen” and the “unseen” dimensions.

​In the “seen” dimension, I will present the case of an “aerial” mobile network that needs to be set-up on-demand to enable public-safety operations. Having a digital spectrum twin for the environment is paramount to optimized aerial base station deployment and connectivity. Yet, one of the biggest challenges in generating an accurate DST is the associated expensive and time-consuming effort of RF data collection in such new environments. I will highlight how the use of a complementary modality such as vision sensors (widely available on most aerial platforms) coupled with intelligent, compact models that can reside on drones, be instrumented to learn RF features from vision images, thereby informing and adapting the RF data collection process. The result is a significant reduction in time and overhead spent in the creation of DST without compromising on its accuracy.

​In the “unseen” dimension, I will present the case of a 5G mobile network coexisting with another heterogeneous network (e.g. fixed satellited, WiFi) in the same spectrum. In contrast to the conventional approach of analyzing RF signals and accordingly adapting communications, I will present a novel paradigm called “interference tomography” – where client devices are transformed to serve as distributed “interference” sensors, albeit at the access layer. By analyzing the 5G access statistics of the clients collectively in a scalable manner, we can transform the existing 5G communication network into a dual interference-sensing network that can “blue-print” and localize the unseen interference sources in the environment, thereby leading to efficient coexistence decisions.