Guest Editorial: Selected Papers From Radar 2023 (Sydney, Australia)

It is our great pleasure to introduce a series of extended papers from the 2023 IEEE International Radar Conference (RADAR 2023), 6–10 November 2023, held in Sydney, Australia. The conference theme was Dreaming the Radar Future. Befitting this theme, the conference received a diverse range of contributions from leading international researchers. A total of 200 full papers were published in the proceedings of RADAR 2023. A selection of authors from the best papers, including the winners and finalists of the best paper and best student paper competitions, were invited to submit extended journal papers for this special issue. The extended papers were required to contain at least 40% new material compared with the conference submissions and were subjected to a separate peer-review process, conducted with the same rigour as regular issues of IET Radar, Sonar & Navigation.

This special issue contains 8 papers, across a wide range of application areas. These include space domain awareness, distributed radar sensor networks, flying target detection from SAR images, clutter processing, drone characterisation, track confirmation, polarimetry for target imaging and over the horizon radar.

Achieving synchronisation presents a major challenge for the deployment of distributed network of radars. Addressing this problem can unleash the vast potential of the sensor network. Kenney et al. [1] present a decentralised technique for attaining frequency, time and phase synchronisation across a distributed network. The presented approach builds on a previously published method for correcting phase and clock bias, and has the advantage of not requiring RF hardware upgrades. A comprehensive theoretical analysis is presented in this paper, along with simulations, that show the proposed technique approaches the theoretical performance limit. A beamforming scenario is used to illustrate how the proposed technique can be implemented to solve a practical problem.

Tracking evasive targets present a great challenge to modern radar systems, particularly for radar resource management. This problem is highly complex, with multiple agents affecting many scenarios. Dolinger et al. [2] present an approach to this problem with reinforcement learning set within a game theoretic context. Three game theory strategies are implemented and tested in a simulated radar environment, with the performance compared against heuristic methods. The results show that collaborative methods achieve greater performance, and that finer nuances in the performance that point towards future research directions.

Howard and Nguyen [3] present a collection of techniques for manipulating the radar ambiguity function for over the horizon radar. They present a novel characterisation of the ambiguity function in terms of twisted convolutions and show how it can be transformed by an area preserving linear transformation of the delay-Doppler plane. The findings have implications on waveforms used for over the horizon radar and the subsequent signal processing. The results show promise of greater robustness against spread Doppler clutter, leading to improved ability to detect small targets.

Radars are a crucial part of counter-drone systems, as they both detect drones and provide features for their classification. The work of Pappas et al. [4] uses three multitask learning (MTL) neural network approaches on micro-Doppler spectrograms to estimate several attributes simultaneously. The three networks share intrinsic features, which improves generalisability. The approach is validated in a variety of classification and regression tasks, with the MTL models offering improvements over single-task learning models.

Synthetic aperture radar (SAR) imaging is a well-established technique with numerous applications. The rich information contained in SAR images offers the potential to extend their utility into new areas. Hansen et al. [5] present the idea of focusing the blurred signatures of flying objects to improve their detectability. This approach is attractive because it repurposes collected radar data for enhanced capability. Key aspects of the work include estimating the target motion parameters and suppressing ground clutter. It was shown that significant gains can be achieved in detection and tracking performance.

Multi-polarisation measurements provide additional information on the scattering properties of targets imaged using inverse synthetic aperture radars (ISAR), which can enhance target recognition. A novel approach to multi-polarisation ISAR is presented by Kumar et al. [6], using a circular transmit and linear receive configuration. This hybrid-polarisation approach captures sufficient target information without resorting to the complexities of a full-polarimetric setup. Extensive experimental results, obtained on real ISAR systems, show the hybrid-polarisation system closely matching the performance of the full-polarisation system.

Maritime radar detection schemes commonly require knowledge of the covariance matrix of the underlying clutter and noise, which usually needs to be estimated. In practical scenarios, the available data may be insufficient to estimate the full matrix, thus requiring reduced rank approaches. Gray et al. [7] use the multistage Weiner filter and propose an algorithm for estimating the number of stages, based on the minimum description length method. The algorithm has computational advantages over rival techniques and its effectiveness in estimating rank and enhancing target detection are demonstrated with simulations.

With the proliferation of space objects in near space, effective sensing of such objects has gained currency in recent years. Objects in low-earth orbit (LEO) have sufficiently low altitude for passive radars to detect them using terrestrial illuminators, leading to cost effective solutions. Jędrzejewski et al. [8] use such an approach with the LOFAR (LOw-Frequency ARray) radio telescope as a surveillance receiver. Multiple reference signals are captured with inexpensive software-defined radios, resulting in multi-bistatic measurements of LEO objects. These measurements are combined to achieve three-dimensional localisation. Real measurements confirm the feasibility of the approach, with additional simulations providing estimates of the position accuracy.

The papers in this special issue are an exciting showcase of radar applications. We trust this is a useful collection for experienced and new researchers alike, which not only offers a description of the state-of-the-art but may also inspire excellent new research directions for radar of the future. It has been an honour to guest edit this special issue. We would like to thank all the contributors and the IET RSN staff for their patient support.

The authors declare no conflicts of interest.