Compressive Imaging in Radio Interferometry

The project "Compressive Imaging in Radio Interferometry" (CIRI) aims to bring new advances for interferometric imaging with next-generation radio telescopes, together with theoretical and algorithmic evolutions in generic compressive imaging.

Radio Interferometry (RI) allows observations of the sky at otherwise inaccessible angular resolutions and sensitivities, providing unique information for astrophysics and cosmology. New telescopes are being designed, such as the Square Kilometer Array (SKA), whose science goals range from astrobiology and strong field gravity, to the probe of early epochs in the Universe when the first stars formed. These instruments will target orders of magnitudes of improvement in resolution and sensitivity. In this context, they will have to cope with extremely large data sets. Associated imaging techniques thus literally need to be re-invented over the next few years.

The emerging theory of compressive sampling (CS) represents a significant evolution in sampling theory. It demonstrates that signals with sparse representations may be recovered from sub-Nyquist sampling through adequate iterative algorithms. CIRI will build on the theoretical and algorithmic versatility of CS and leverage new advanced sparsity and sampling concepts to define, from acquisition to reconstruction, next-generation CS techniques for ultra-high resolution wide-band RI imaging and calibration techniques. The new techniques, and the associated fast algorithms capable of handling extremely large data sets on multi-core computing architectures, will be validated on simulated and real data.

Astronomical imaging is not only a target, but also an essential means to trigger novel generic developments in signal processing. CIRI indeed aims to provide significant advances for compressive imaging thereby reinforcing the CS revolution, which finds applications all over science and technology, in particular in biomedical imaging.

CIRI is thus expected to impact science, economy, and society by developing new imaging technologies essential to support forthcoming challenges in astronomy, and by delivering a new class of compressive imaging algorithms that can in turn be transferred to many applications, starting with biomedical imaging.

Firstly, CIRI will bring new advances for ultra-high resolution wide-band radio-interferometric imaging and calibration, together with theoretical and algorithmic evolutions in generic compressive imaging. By bringing new knowledge from a multi-disciplinary perspective, our results will induce paradigm shifts that would be unreachable by the astronomical and signal processing communities separately.

This foreseen multi-disciplinary academic impact will subsequently enable each community to deliver its own academic, economic and societal impact. On the academic side, high resolution imaging in radio interferometry (RI) will enable the new science envisaged with future radio telescopes such as the SKA, where the PI and Co-Is have exisiting collaborations. The theoretical and algorithmic evolutions in compressive imaging will reinforce the compressive sampling (CS) revolution for information theory. The generic CS algorithms developed to handle large data sets on multi-core architectures will be transferred for adaptation to other imaging applications, in particular for MR imaging and ultrasound imaging. On the economic and societal side, CS typically targets the design of new simpler, inexpensive, energy-efficient devices for technological applications, thereby linking to industry. These applications may relate to healthcare via devices for biomedical imaging applications, to digital economy and manufacture via inexpensive devices such as cameras, and to the global issue of "energy" via energy-efficient devices such as body sensor networks. Finally, astronomy targets philosophical societal questions relative to the structure and origin of our Universe.

In summary, the direct multi-disciplinary academic impact of CIRI will constitute a critical pathway towards (indirect) economic and societal impact. While the former will scale on the duration of the project, the latter will require progressive integration and implementation of our techniques and algorithms at various levels of application, which will scale in decades. CIRI will also have a direct societal impact through dissemination of scientific culture via public events (project timescale).

Secondly, CIRI will bridge the gap between signal processing and application, and deliver first direct economic and societal impact. Upon completion of the project, new concepts will have integrated validated imaging technologies. CIRI will target direct economic impact through patenting procedures for our imaging techniques and algorithms and by establishing R&D relations with industry. Potential collaborations are with Siemens, General Electrics, and Philips in biomedical imaging, and the with the UK Astronomy Technology Centre, a major industrial actor in astronomy and numerous other applications. This will represent a first step before transfer to non-astronomical applications and commercialisation. Commercial products may appear that would represent the production of wealth within the UK.

Finally, CIRI will have a direct impact on people (project timescale). CIRI researchers will sharpen their skills for the mathematical formulation of complex problems and for analytic reasoning. Beyond these hard (i.e. technical) skills, they will also acquire essential soft (i.e. interpersonal) skills for management, leadership, team work and communication. This portfolio of skills will participate to their professional development and help them to become leaders in any employment sector.