Commensal Radar
Commensal Radar Introduction Commensal Radar: an ongoing collaborative project between Peralex, UCT and CSIR using the latest techniques and technologies to make passive radar viable Why Commensal Radar? Other terms: Passive Radar Passive Bistatic Radar (PBR) Passive Coherent Location (PCL) Passive Covert Radar (PCR) Commensal Radar is a project aimed at exploring multiple techniques and different approaches to finding a practical solution for an optimised result The collaboration: UCT The academic research partner CSIR The application and industry research partner Peralex The industry partner
UCT Radar Remote Sensing Group RRSG Initiated in 1988 by Prof. Michael Inggs with the aim of developing advanced sensors utilising radar technology for the user community Valuable expertise in research, algorithm and software development Provides vital function of producing young engineers in the field of EW Work specific to PCL first published in 2007 Since then many projects have focussed on the subject: Propogation modelling research Tracking filter research Antenna Design Investigations on DVB-T Sebastiaan Heunis's MSc "Passive Coherent Location Radar using Software-Defined Radio techniques in 2010 the first to practically detect a target using FM Craig Tong took over as current team leader whilst writing his PhD on the subject
CSIR - DPSS Expertise: Understanding of the Radar & EW environment and support the DoD and SANDF in this knowledge field. Thus, being the bridge between the academic world and industry Research capability in Radar & EW Application knowledge in Radar & EW Technical knowledge Modelling and simulation capability Background to Commensal Radar: Investigations to identify technology that could aid as a gap filling radar technology. This was based on the Awarenet concept with a primary radar, and supplemental to this radar, PCL was identified as a technology that could aid as gap filling radar technology for: Border Safeguarding Anti-piracy Study and simulations on coverage over South Africa Hardware requirements study Joined Commensal Radar collaboration in 2011, represented by Francois Maasdorp
Peralex Founded in 1987 Approximately 50 employees Based in Cape Town Strong and lasting relationship with GEW Technologies Close relationship with UCT Electrical Engineering department led to formation of Commensal Radar collaboration in 2010
Peralex: Core Competencies System Product and system level design and integration Hardware Mixed signal board level design Complex PCB design VHDL for mainly Xilinx FPGAs Mechanical design RF Software Low level DSP Device drivers High performance software Signal processing algorithm development and implementation Performance optimisation (GPU programming) Man machine interfaces Domain knowledge Direction Finding Signal Detection Signal demodulation and decoding
The Passive Radar Concept Tx 1 Tx 2 ref surv ref surv Rx 2 Rx 1
Pros and Cons Pros: Lower frequencies than traditional radar Effective against stealth aircraft Minimised effects from weather conditions Lower cost (capital outlay, operation and maintenance) Covert No transmitter, therefore radar stations are invisible in RF spectrum No spectrum license required Cons: Complexity of deployment. Geometries are a critical performance factor Reliance on third party transmitters of opportunity Signal content inconsistent Range resolution depends on signal bandwidth Long integration times required leading to slow update rates Computationally intensive
History of Passive Radar Watson-Watt s earlier experiments in 1935 demonstrated the concept First successful use of a passive bi-static radar system by Germany in World War II The Klein Heidelberg which exploited the British Chain Home radar transmitter Bi-static radar made way for conventional mono-static radar Resurgence in the 1980s with the emergence of cheap computing power Lockheed-Martin s Silent Sentry initially released in 1998 is the first commercial product Thales develop Homeland Alerter 100 (2005-2007) There has been a noticeable resurgence in passive radar recently. A large number of papers on the subject being presented at the most recent Radar 2012 IET conference in Glasgow
Current Notable Products/Demonstrators Homeland Alerter 100 Thales Released in 2007 Range : 100km Azimuth : 360 Elevation : 90 Transmitters of opportunity : FM radio, possible extension to DAB, AVB, DVB-T Cassidian Technology Demonstrator unveiled in 2012 Transmitters of opportunity : FM,DAB,DVB Range : test results reported at 250km bistatic Small ground based targets (cyclist) detected at short range CELLDAR BAE & Roke Exploiting GSM transmitters for medium range coverage
Current Notable Demonstrators/Research Warsaw University of Technology Developed PaRaDe (Passive Radar Demonstrator) Reported detections of 750 km bi-static range using 60kW FM transmitter Fraunhofer research institute Wide interest in the subject Supports many projects/research including Cassidian University of Pisa Demonstrator of Passive Inverse Synthetic Aperture Radar (P-ISAR) to image ships using DVB-T signals University College London Demonstrator system exploiting DVB-T
Transmitters of Opportunity FM Pros: Typically high power transmissions Wide coverage Drawbacks: Inconsistent signal content Narrow bandwidth (160kHz) resulting in poor range resolution although multiple channels can be combined to improve range resolution Analog TV Drawbacks: Not ideal for range resolution because of cyclic nature of waveform Is being replaced by DVB DVB-T Pros: Wide bandwidths (8MHz) Consistent signal content Typically high power transmissions Cons: Coverage although this is changing quickly Often deployed in a Single Frequency Network (SFN) which can result in ambiguities that are difficult to resolve GSM Exploited by CELLDAR system from BAE/Roke Typically short range applications: traffic monitoring, maritime traffic surveillance WiFi Satellite TV Weather RADAR
Broadcast FM Transmitters: Southern Africa Source: BR IFIC 2726 (Aug 2012)
FM coverage in South Africa Source: CSIR
Commensal Radar: System Overview Reference Ch Surveillance Ch Signal Acquisition Adaptive Cancellation ARD plot Cross Correlation CFAR Detection Line Tracking Line tracks from other Rx sites Association, geo positioning and target tracking Target track
Commensal Radar: Key Features Receiver hardware based on existing product developed for GEW technologies High Dynamic Range receivers and A/D technology Aim to exploit FM Strong transmitters Direct sampling is possible Maximise spurious free dynamic range (SFDR) Reduce cost by avoiding a super heterodyne receiver front-end Real-time processing achieved using GPUs Mobile Receiver, GPU processor, and user interface on laptop run off a single inverter off a vehicle battery
Commensal Radar: 2 different approaches 2 different approaches are considered: Co-located architecture The reference and surveillance channels are in the same location 2 channel receiver is used common clock, therefore output channels are time and frequency aligned Directional antennas are used to reduce direct signal interference on the surveillance channel, but remains a challenge as directional FM antennas are large Separate reference architecture The reference and surveillance channels are captured in different locations Receivers are frequency and time aligned using a GPS disciplined oscillator (GPSDO) Can use relief of the land to hide surveillance channel from direct path interference Challenge: How to transfer large amounts of data to a central location for data processing
Approach 1: Co-located architecture Tx 2 Data Processing ref surv Rx 1
Approach 1: Separated reference architecture Minimised Direct path Interference surv Rx 2 Tx 2 ref Data link Data Processing Data link Rx 1
Measurement Setup Environment Cape Town is ideally suited for test measurements due to the shielding from mountain and availability of at least 2 strong FM transmitters Approach Use UCT s propagation modelling tool to find optimal receiver locations Real-time measurements using receiver confirm suitability of theoretical sites Truth data recorded using SBS receiver which captures transmissions from commercial airliner ADS-B transponders Flight name GPS location, altitude and velocity recorded vs time Data is correlated with measured results In-field real-time ARD processing vital to confirming test setup All raw captured data recorded for offline processing
Measurement Setup: Map
Measurement Setup `
Measurement Setup
Results: Early Detections Transmitter: FM radio from Tygerberg Tx Rx Location: Tygerberg Rx site 2 Targets: Boeing 737-400 Airbus A340-500 Bi-static range: >130km
Results: Long range detections Transmitter: FM broadcast from Tygerberg Tx Rx Location: Malmesbury Rx site Target: Airbus A330-200 Bi-static range: >270km Addition of CFAR implementation to software
Results: Separated Reference Rx Location: Reference at Peralex Surveillance at Boyes Driver Transmitter: Broadcast FM from Constantiaberg Tx Range: ~65km bi-static This needs further investigation Strong interference signal at 40km, which is a suspected multipath problem that affects the cancellation algorithm Further investigation is required
Results: Multi-static measurements Transmitter: FM broadcast from Tygerberg Tx Rx Locations: Tygerberg site Malmesbury site Backsberg site Positive detections at all 3 sites Successful correlation of results with truth data
Results: Multi-static measurements Result 2012-08-07.kml
Current Challenges Direct path interference Multipath Long integration times (4s) High bandwidth data links required for separated reference architecture
Future Work Integration of UCT s latest tracking algorithms Combining multiple FM channels Exploit transmit diversity Addition of direction finding (DF) to system Multichannel antenna Electronic steered null and beam Increased sensitivity Tracking can be done using range, Doppler and bearing leading to increased robustness Further investigation of separated reference architecture Investigate other locations for measurements Further small target measurements
Questions?