We demonstrate a novel architecture enabling the correlation in real-time of broadband RF signals (> GHz). Contrary to conventional digital correlators, our technique is analog: no digitization nor digital signal processing is required. The correlation is performed in the optical domain, enabling the processing of multi GHz signals. Moreover, the proposed architecture calculates in real-time the correlation function for more than 200 values of the delay simultaneously. Applications of the technique range from radio-astronomy, to transmitter localization by Time Difference of Arrival.

This paper investigates the properties and application of structured radio beams using phase-controlled radio-frequency (RF) antenna arrays. First, is an investigation of a unique, structured, mixed-frequency radio beam carrying orbital angular momentum (OAM). EM beams with OAM exhibit self-healing properties in the optical realm, which extends to radio communications, sensing, and combined radar-communications. Airy beams also self-heal, and have free-accelerating properties. Reproducing these properties via small phased antenna arrays enables remote sensing applications, including phase-controlled leveraging of the parabolic trajectory. Results from laboratory experiments over 0.7 to 6.0 GHz as well as results from numerical simulations are presented.

Drone recognition has become a topic of increasing concern for defense applications. Due to the high speed of rotation of the drone blades, however, accurate drone recognition relies on sufficient time-frequency resolution of the drone radar micro-Doppler signature. Although one of the more commonly used time-frequency transforms is the spectrogram, such classical estimators embody a sub-optimal trade-off in temporal resolution versus frequency resolution. In this work, we evaluate the efficacy of various time-frequency transformations based on the latent space of deep neural networks (DNNs). In particular, we consider alternatives to the short-time Fourier transform, such as the wavelet transform, Wigner-Ville distribution, Choi-Williams distribution, and and super-resolution techniques, which have been recently shown to be effective on non-radar datasets, such as superlets. Transforms are compared for various millimeter wave radar systems for DNN-based classification.

Accurately counting numbers people is useful in many applications. Currently, camera-based systems assisted by computer vision and machine learning algorithms represent the state-of-the-art. However, they have limited coverage areas and are prone to blind spots, obscuration by walls, shadowing of individuals in crowds, and rely on optimal positioning and lighting conditions. Moreover, their ability to image people raises ethical and privacy concerns. In this paper we propose a distributed multistatic passive WiFi radar (PWR) consisting of 1 reference and 3 surveillance receivers, that can accurately count up to six test subjects using Doppler frequency shifts and intensity data from measured micro-Doppler (µ-Doppler) spectrograms. To build the person-counting processing model, we employ a multi-input convolutional neural network (MI-CNN). The results demonstrate a 96% counting accuracy for six subjects when data from all three surveillance channels are utilised.

Automatic target recognition is an important technology that has numerous important applications. This technology is presented by examining the current state of the art and then reviewing the past from a statistical pattern recognition perspective to follow the evolution into the deep learning approaches popular today. However, automatic target recognition has additional challenges that are outside the sweet spot of current deep learning approaches. To look to the future, key technical challenges are outlined and a way forward is proposed to accelerate automatic target recognition technology development.