Timothy Bateman, UNSW Canberra Space; Edwin G. W. Peters, University of New South Wales Canberra; Rabbia Saleem, UNSW Canberra Space; Andrew Lambert, UNSW Canberra Space; Melrose Brown, UNSW Canberra Space
Keywords: passive RF, wide band, space domain awareness, space situational awareness
Abstract:
The number of satellites in low Earth orbit (LEO) is expanding at the fastest rate in human history. Current and planned LEO constellations look set to introduce up to 50,000 satellites in close proximity to Earth and identifying, tracking and cataloguing this proliferation of objects becomes more important than ever. Traditional Space Domain Awareness (SDA) approaches combine optical observations and narrow band RF observations of actively transmitting satellites. Radar observations of the LEO belt, both bistatically and monostatically add additional information on passive debris and non cooperative non-RF emitting satellites, but this comes with the complications of frequency compatible RF emitters and spectrum licensing. The current state of passive RF SDA requires a-priori information in the form of the timing of transit windows, transmitting frequency and modulation schemes and due to data processing limitations is inherently offline, and data intensive. To keep the data collected in a typical LEO satellite pass to practical storage limits, narrow band captures with software defined radio (SDR) receivers is commonly performed. In this work, we detail a wide band approach using low cost commercially available COTS SDR receivers. Instead of narrow field of view RF sensors we employ wide field of view omni-directional antennas, reducing the need for troublesome and unreliable antenna rotators and accurate orbit information for pointing. The low cost and ever more capable computational resources of consumer grade graphics processing units (GPUs) are leveraged. This allows large RF bandwidths to be processed in real time. Open modular network based message protocols allow block based software units to connect to network sockets and receive both decimated and full rate RF data. Data products generated by the software pipeline include visual frequency vs time waterfall plots, time and frequency estimation via digital phased locked loops (PLLs) implemented to run in real-time utilizing GPUs and traditional IQ capture for Doppler analysis. This work offers several advantages over a traditional SDA processing pipeline. Wide band frequency capture allows for the possibility of object detection, negating the need for a-priori PSRA intelligence. Continuous capture means that patterns of life studies are enabled for particular frequency ranges. Anomaly detection is enabled. We present a system developed to capture and analyse wide band UHF spectra of up to 60 MHz, which is the limit of the SDR hardware at hand. The system will use omni-directional antennas to capture the entire sky. The proposed processing pipeline produces data in real-time, and reduces a 60 MHz IQ data stream to information of mere mega bytes per minute. The information includes visual aggregated spectrograms for visual inspection and time frequency estimations that can be used for orbit determination (OD) and identification. Future work will focus on the addition of accurate time tagging information to allow increased accuracy for OD. Additionally, other bands can be studied through the consideration of the novel antenna structures such as aperture phased arrays and switched antenna networks. Once the quality of this pipeline is proven, data products can further be analysed by machine learning algorithms and ingested into mission systems.
Date of Conference: September 19-22, 2023
Track: Space Domain Awareness