Exciting research in radar: Wind farms, commensal radar and whitespace radar

26 Sep 2014 - 10:45

This week, we were privileged to welcome a distinguished visitor from the United Kingdom to our southern shores: Professor Tony Brown of the University of Manchester, where he is Head of the School of Electrical and Electronic Engineering and holds the Chair in Communication Engineering.

Prof Brown is heavily involved in research in radar systems, antennas, propagation, wireless communications and radio astronomy instrumentation. His most recent research looks at the impact of wind farms on radar performance, and the design of extremely large multi-octave phased arrays for the Square Kilometer Array radio telescope.

The Impact of Wind Farms on Radar Performance

The impact of wind farms on many types of radar system has been a major factor in limiting the approval to deploy wind farms in the UK and elsewhere. Operational experience and field trials show that wind farms located within the line of sight of radar can cause interference with the radar performance, confusing trackers and affecting overall target detection.

Due the large size of a single turbine (which can have a blade length of up to 120m), prediction of the impact can be demanding. Moreover, when a full wind farm is considered, it may also be essential to predict the interactions between all the wind, and indeed potentially with the local terrain. Given that wind farms can be made up of several hundred wind turbines, the complexity is obvious.

Prof Brown’s lecture outlined the electromagnetic simulation of these large and complex structures, and gave an overview of the types of radars used in aviation and marine navigation. He also looked at the cause of the interference caused by wind turbines, identifying the main scattering characteristics of the turbine as an individual entity and the wind farm in totality. Lastly, he presented some of the available mitigation measures with their potential benefits and limitations.



NB: A recording of Prof Brown’s talk on The Impact of Wind Farms on Radar Performance is available here: http://meeting.uct.ac.za/p8nxmg432m5/.

The best HTML CheatSheet has its own visual editor where you can apply your own CSS code as well. Web developers and designers love this free online resource!

A Series of Lectures

In addition to Prof Brown’s talk, Prof Mike Inggs of the Radar Remote Sensing Group and Assoc Prof Daniel O’Hagan, the new Convener of the Radar Masters Programme, had arranged a series of short presentations in order to give Prof Brown an overview of the work that is being carried out in the RRSG.

Click on the links below to read more about the various topics:

UCT Radar Masters Programme

Assoc Prof Daniel O’Hagan introduced the UCT Radar Masters Programme, which was set up in 2011 to address the growing need for skilled engineers and scientists in the challenging fields of Radar and Electronic Defence/Electronic Warfare.

There is a skills deficit in South Africa, the United Kingdom, the United States and Australia, in, among others, the areas of gathering, analysing and interpreting complex radar information as well as developing, testing and maintaining new radars. Few academic institutes worldwide offer specific courses in radar and electronic defence, yet most appreciate the importance of these fields.

The aim of the UCT Radar Masters Programme is thus to develop highly skilled radar engineers fit for the demands of an intellectually challenging and stimulating career in radar and electronic defence, as well as to provide a launch-pad for top students wishing to pursue a PhD. Moreover, the programme equips graduates with transferable skills, for instance, in systems engineering, digital signal processing, mathematics, antenna theory, and principles and systems of radar and electronic defence/warfare.

Various streams are available, viz. a Professional Taught Masters Degree (MEng Radar), a Research Masters Degree with Coursework (MSc Eng specialising in Radar), and a Research Masters Degree by Dissertation (MSc Eng specialising in Radar). A wide range of courses is offered, some are taught by UCT-based teaching staff, whereas others are presented by guest lecturers who are often experts in their respective fields, from South Africa and overseas. These courses can also be done as part of the Continuing Professional Development (CPD) programme.

Research in the Radar Remote Sensing Group

Prof Mike Inggs looked back over 26 years of research in the Radar Remote Sensing Group (RRSG) since its early beginnings in 1988; interestingly, 2014 marks the 75th anniversary of Radar in South Africa, with the first emissions of a locally designed, integrated and tested system 75 years ago.

The early work of the RRSG (in the 1990s) focused on the following areas: ship and aircraft recognition from both low and high resolution imaging; Inverse Synthetic Aperture Radar (ISAR) – in close collaboration with the CSIR in Pretoria; the use of Ground Penetrating Radar (GPR) for landmine detection and geotechnical work, using Stepped Frequency Continuous Wave (SFCW) radar; airborne SAR – fully polarimetric – in the VHF band; cluster computing studies for machine vision and SAR processing; and the investigation of high speed signal sources for GPR systems.

In the 2000s, the research focus changed to the following areas: the development of networked radar in collaboration with University College London; the use of Commensal Radar in the FM band, based on a single transmitter site in conjunction with multiple receivers; the study of high speed digital hardware (specifically looking at Field Programmable Gate Arrays [FPGAs]) for use in radio astronomy (this was in the early days of the Karoo Array Telescope project, the precursor to the Square Kilometre Array South Africa project); and lastly, heterogeneous computing for Radio Astronomy, making use of FPGAs and Graphics Processing Units (GPUs).

From 2010 onwards, postgraduate students in the RRSG have been looking at a new set of areas: integrated FM Band Commensal Radar with Doppler only tracking, using the Gauss Newton filter developed by Dr Norman Morrison; the development of software for heterogeneous computing; the design of novel microwave filter technology (Assoc Prof Riana Geschke focused on this in her presentation); White Space Radar and Orthogonal Frequency-Division Multiplexing (OFDM) research (Dr Amit Mishra spoke about this); antennas and high power duplex/limiter (work being done by Prof Barry Downing); and precision timing for radar and radio astronomy.

The latest project, known as NeXtRAD, is a collaboration with University College London to develop a novel multi-node bistatic/multistatic radar system in the L and X Bands, the successor to the previous NetRAD system. Trials are planned for early 2015, with a prototype system to be deployed in the Western Cape to track ship targets within sea clutter. It is envisaged that further trials will be done in Europe and the UK from 2016.

Commensal Radar Project

PhD student Craig Tong spoke about the Commensal Radar project that he and several fellow students are engaged in. We previously wrote about this here (the research work being done by fellow PhD student Francois Maasdorp) and here (the field tests of the prototype multistatic commensal radar system).

His presentation gave an overview of the aspects that the RRSG has been concentrating on for the past decade as well as indicating the direction for the future. The envisioned system makes use of commercial FM broadcast band transmitters for the purpose of air traffic control. Furthermore, given the typical broadcasting schemes used in Africa, which consist of sparsely located, high powered transmitters, the commensal radar system is envisioned to consist of only a handful of these transmitters (often only one) and many receiver sites placed optimally to create multiple transmitter/receiver pairs, which are required to detect and position an aircraft. This is contrary to what might typically be implemented in continental Europe, where the transmitter density is high and a single radar receiver site could then be deployed to create the multiple transmitter/receiver pairs with many available transmitters.

Some of the research demonstrated in his presentation included the work on propagation modelling in order to select optimal sites to place the radar receivers. This aims to maximise the signal to interference ratio of the system as well as the overall geometry to in turn minimise geometric dilution of precision and in so doing, to facilitate the accurate positioning of aircraft. A real-time processing chain was developed on graphics processor units to perform interference suppression as well as the very long correlations required to detect aircraft of interest. The “separated reference” configuration is a technique for recording reference and surveillance signals of the radar at separate locations in order to receive cleaner versions of these signals. This is advantageous, as the locations required to do these recordings typically have conflicting requirements.

The separated reference receivers make use of GPS disciplined oscillators to maintain sampling coherence. The same technique is applied in the NetRad radar system and will also be applied in NetRad’s successor NeXtRad. An initial proof of concept for target positioning using multilateration is demonstrated as well as a review of a recent measurement campaign where data was collected for study into a new tracking scheme to position aircraft accurately and robustly. Results are also presented from several previous field tests showing long range detections of large commercial airliners, medium range detections of small aircraft as well as measurements of blade modulation speed caused by the propeller of the small aircraft. Finally, it was demonstrated that combining several FM broadcast channels together in the same bistatic triangle provides improved robustness against multipath and drops in modulation bandwidth of the FM signal.

Design and EM modelling of Passive Devices

Assoc Prof Riana Geschke‘s presentation looked at Activities in the Design and EM modelling of Passive Devices at UCT.

Microwave systems are used for various purposes within a frequency range of 1 to 30 GHz. Three notable microwave applications are communications links (personal communications, wireless internet), radar systems and receivers for radar astronomy, such as the Square Kilometre Array Project. Many systems now operate over multiple frequency bands and may also include operation over wide bandwidths.

While digital components are increasingly used in microwave systems, the front-ends require microwave components such as antennas and microwave filters, which cannot be replaced by digital components. The presentation focused on filters: ways to provide operation over multiple bands and include electronic reconfigurability. Results from a number of projects were presented.

Whitespace Radar Project

Dr Amit Mishra introduced the Whitespace Radar project.

The idea of making communication and radar systems coexist has been investigated since World War II. The last few years have witnessed a surge of activities in the field of utilising communication broadcast signals for target detection and tracking using receiver-only sensors. These have been called passive and sometimes parasitic radars, but we prefer to use the phrase Commensal Radar.

Recently, there have been papers in the open literature on Commensal Radar using FM signals, digital TV signals and even satellite signals, all centered around developing independent receivers. The work in RRSG on whitespace radar is on a novel system based on the concept of extending the commensal principle to the whole system, such that the Radar can become an integrated commensal function of the communication system.

The recently coined term ‘whitespace’ refers to any band of spectrum that is sparsely occupied. For instance, we chose to work in the analog TV band based whitespace because there are very few transmitters active in those bands currently. This has many advantages.

Digital backend board developed in-house to cater to both Whitespace communication and radar usages


Firstly, it gives us access to a wide bandwidth, which in turn means better resolution for the coexisting radar system. Secondly, the standardization process for Whitespace communication is over IEEE 802.22. And lastly, this will be a communication system of choice for sparsely populated areas, which are also sites suitable for radar surveillance. A survey of unused spectrum in Southern Africa has shown immense potential for whitespace devices in the rural areas, where the spectrum is mostly empty.

The TV band signals, having a broad bandwidth as well as long distance propagation capabilities without much signal loss compared to standard wireless communication signals, which operate at much higher frequencies, give us an opportunity to tap into their potential for networked radar. This can be used in remote areas, where there may not be any access to wired networks. There have been studies regarding the utilization of communication signals for radar functionality as well.

This is ongoing work; we have already filed a UCT owned patent for this work. This is an image of the digital backend board that was developed in-house to cater for both Whitespace communication and radar usages.

Hybrid High Performance Backend Data Processing and Skills Development

Lastly, Dr Simon Winberg, spoke about Hybrid High Performance Backend Data Processing and Skills Development.

Hybrid systems are composed of multiple types of processing nodes, for example, clusters of computers that integrate both CPUs and GPUs for performing parallel data processing. The emphasis is on backend processing – i.e. processing that happens after the analogue circuitry, sampling and initial filtering has been done. Frequently, the general structure of a hybrid processing system involves either one, or a cluster of, CPU-based computer systems, the processing nodes, which connect to a plugin or networked accelerator board. The choice of processing node design is, to a large extent, dependent on the application and type of processing approach that suits the application.

For example, GPUs usually like data presented as blocks, which are transferred to and back between the CPU’s main memory and the GPU card’s memory (with GPU processing happening between these instances of data transfer), rather than processing data streams, which are often better suited to DSP or FPGA accelerator cards. There are several activities within the RRSG related to this topic, particularly with regard to student projects supervised by Dr Winberg. The focus is on skills development, specifically for parallel hybrid systems and FPGA-based data processing.