Deep Convolutional Neural Networks for Device Identification
- IEEE Cognitive Network Security SIG, Dec 2020
- Princeton University, Nov 2020
- Texas A&M University Kingsville, Oct 2020
Abstract: Network densification is poised to enable the massive throughout jump expected in the era of 5G and beyond. In the first part of the talk, we identify the challenges of verifying identify of a particular emitter in a large pool of similar devices based on unique distortions in the signal, or ‘RF fingerprints’, as it passes through a given transmitter chain. We show how deep convolutional neural networks can uniquely identify a radio in a large signal dataset composed of over a hundred WiFi radios with accuracy close to 99%. For this, we use tools from machine learning, namely, data augmentation, attention networks and deep architectures that have proven to be successful in image processing and modify these methods to work in the RF-domain. In the second part of the talk, we show how intentional injection of distortions and carefully crafted FIR filters applied to the transmitter-side can help in enhanced classification. Finally, we discuss how to detect new devices not previously seen during training using observed statistical patterns. We conclude by showing a glimpse of other applications of RF fingerprinting, like 5G waveform detection in large-scale experimental platforms and identifying a specific UAV in a swarm.
(Air)space is the Final Frontier: Experiments with Learning, Sensing and Communications in UAVs
- Rice University, Oct 2020
- Virginia Tech, Nov 2020
Abstract: The rapid growth in the deployment of unmanned aerial vehicles (UAVs) opens up new paradigms for spectrum sensing, RF actuation, and air-ground communication. In this three-part talk we will cover, from an experimental viewpoint, the challenges in creating intelligent and coordinated UAV swarms that are synchronized in their operation. In the first part, we will show the first of its kind implementation of distributed beamforming with UAVs, with tight constraints on timing and real-time feedback of channel state information. In the second part, we focus on millimeter (mmWave) radio mounted UAVs that can serve as mobile 5G base stations and aerial backhaul links. We will motivate the need for - and then design - new channel models for air-to-air and air-to-ground communication at 60GHz using Facebook’s Terragraph channel sounders that are custom-fitted on DJI M600 UAVs, considering practical limitations of hovering and airframe reflections. In the third part, we shall describe a method that ensures jamming resilient air-ground control channels using a combination of mmWave radar sensing and physical movement of the UAVs that form wireless signal constellations in the air. While concluding, we shall provide a glimpse of other ongoing work in the systems space, including a method for wireless charging of UAVs through magnetic resonant coupling and deploying UAVs within an urban region using an existing public bus transportation network.