Collaborative Research: SWIFT: Collaborative Interference Cancellation for Radio Astronomy

This project seeks to enhance coexistence in shared frequency bands between wireless cellular networks and passive radio telescopes. This project helps address challenges created by growing congestion of the radiofrequency spectrum. If successful, the project will improve radio telescope capability through reducing interference problems.

It will reduce barriers to further deployment of wireless networks, thus facilitating the high bandwidth reliable and ubiquitous connectivity needed for continued growth in the benefits of telecommunications. Interference problems between communications and radio telescopes have existed for a long time. One approach studied to overcome them has been active interference cancellation, in which the telescope receiver estimates the interfering communications signal then subtracts it from the astronomical measurement in order to expose the desired information.

Most prior work on active cancellation assumes no cooperation from the cellular networks; this work studies how the telescope and cellular network could work together to improve cancellation performance. Most prior work measures the interfering signals at the telescope site, while this work measures the interfering signals at their transmission location; most prior work uses external spectrum monitoring devices, while this work characterizes the interference in the digital domain of the transmitter; these attributes of the approach give a clearer picture of the interfering signals at much less cost, but also create challenges that the project will seek to overcome.

The approach used focuses on active interference cancellation at the telescope supported by active bidirectional collaboration between the telescope and neighboring cellular networks. Key aspects include: characterize cellular system signals using software running inside the cellular base stations that analyzes the digital representation of the transmissions; decompose the cellular signals at the base stations into an orthogonal stochastic component representation (the Karhunen-Loeve Transformation); perform the decomposition using a neural network architecture with domain specific knowledge to combine high accuracy and computational efficiency; intelligently aggregate information about multiple emitters in the cellular network using topology aware distributed learning; perform the same decomposition on the astronomical signal at the telescope then combine it with the aggregated information about the cellular signals to cancel the interference; and use continuous quantitative feedback from the cancellation algorithms at the telescope to the cellular network to optimize the performance of the decomposition and aggregation algorithms.

The research will be validated with a prototype deployment of the interference measurement and cancellation apparatus at the Deep Synoptic Array telescope (DSA-110) at the Owens Valley Radio Observatory.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.