Interview with Yashwant Gupta, Professor and Centre Director, NCRA-TIFR, Pune, India

Yashwant Gupta  is a Professor and Centre Director at the National Centre for Radio Astrophysics (NCRA), Pune, India.  NCRA is one of the national centres of the Tata Institute of Fundamental Research (TIFR).  Prof. Gupta’s research interests are in the broad area of radio astronomy, which special emphasis on the studies of pulsars – exotic neutron stars that are extremely compact (~ 10 km in radius),  rapidly rotating (~ 1 to 100s of times in a second) objects emitting powerful radio beams that reach us once every rotation (like a lighthouse); understanding the interstellar medium of our galaxy – the ‘empty’ space in between stars; as well as signal processing and instrumentation for radio telescopes.  The NCRA is host to one of the largest radio astronomy observatories in the world – the Giant Metrewave Radio Telescope,  located about 80 km from Pune, India.

AG: The Giant Metrewave Radio Telescope (GMRT) has been selected as a ‘Milestone’ facility by the IEEE. Tell us something about this project, the impact of this work,  and the contributions made by your team.

YG : The GMRT is one of the largest and one of the most sensitive low frequency radio astronomy facilities in the world today.   It consists of an array of 30 antennas of 45 m diameter each, spread out over a 30 km region about 80 km from Pune, with sophisticated electronics and computing for processing the data from all the antennas.  It operates over a wide range of frequencies covering 110 MHz to 1450 MHz, with the capability of processing an instantaneous bandwidth of 400 MHz from each of the dual polarisation receivers from all the 30 antennas. The GMRT was conceived of and proposed in the late 1980s, built and made operational during the 1990s, and opened for use by the global astronomy community in 2002.  At that time, it was one of the most powerful facilities at low frequencies, with several innovative features in its design and implementation. Over the years, it has been used by astronomers from all over the world (more than 40 countries) for a range of cutting edge science and discoveries that have significantly enhanced our knowledge and understanding about different aspects of our Universe.   During 2012 to 2019, it went through a major upgrade of its capabilities, that has preserved, and even enhanced, its status as a leading facility in the world for low frequency radio astronomy studies of the Universe. 

Both the original GMRT and the upgraded facility have been completely indigenous efforts – designed and executed entirely by the team at NCRA.   Given this, and the significant impact that this facility has had on the global scale, it was quite appropriate that IEEE India supported the move to have the GMRT considered for the ‘Milestone’ status.  I am extremely proud and pleased that we were able to convince the IEEE jury (after a fairly rigorous review process) about the suitability of the GMRT for this status – it is a great achievement for science and technology in India in the modern era, given that the only other such IEEE Milestones in India are ones that recognise the pioneering work of J.C. Bose (way back in 1895) to demonstrate for the first time, the ability to generate and receive radio waves and that of C.V. Raman (in 1928) for his Nobel prize winning discovery!

AG: Please share your other impactful work(s) with us and your current research work in the area of signal processing.

YG : I have been fortunate to have been associated with the GMRT and the NCRA since the very beginning.  This has allowed me to make significant contributions to the building of the GMRT, as well as to use it for some exciting science.  My major contributions to the making of the observatory have been in the area of signal processing : in the 1990s, I led the work on digital signal processing receiver (colloquially called ‘back-end receiver’) for the original GMRT; this was a hybrid design using ASICs and DSP chips to process data from the 30 antennas (coming at an aggregate rate of 2 Gsamples/sec) to implement a spectral correlator and a beamformer – this was the heart of the GMRT receiver system that processed the raw data and made it accessible by the astronomers in a volume and format that they could work with for extracting science results.  Later, around 2005-2010, my team improved on the original back-end by converting it into a ‘software receiver’ that carried out all the real-time signal processing, after digitising the data at Nyquist rate, on a network of CPUs.   Finally, when we upgraded the GMRT, we carried over the signal processing to GPU-based machines, to cater to a factor of 10 increase in the aggregate data rate (due to increase in the bandwith of the signals being processed).  

Currently, my team at the GMRT is engaged in developing next generation algorithms for real-time filtering of the data to mitigate against the effects of unwanted radio noise signals from man made activities in the vicinity of the observatory, that are picked up very strongly by the GMRT antennas an receiver systems.   We are also working on novel modes of signal processing of the astronomical data to enhance the quality of the images and results that we can get with the GMRT.  These include use of some of the modern techniques related to big data, high performance computing and machine learning – all very exciting and intense stuff!

AG: During these COVID times, what have been the challenges faced by your team to keep a large facility like the GMRT functional?   What about the impact on the research work in your institution? Do you have any suggestions for signal processing researchers? 

YG: An observatory like the GMRT runs 24x7, catering to the experiments of users from all over the world.   Furthermore, a large engineering team needs to work in a coordinated fashion for the upkeep of the antennas and the electronics, which are spread out over a region of 30 km in extent ! The restrictions due to the Covid pandemic posed quite a few challenges to our ability to operate the GMRT.   During the first one month of the lockdown, we had no choice but to shut down the observatory, as staff could not reach the site easily, nor could they work in the labs and control room  and travel to the various locations in the array was also difficult.   As the lockdown eased, we worked out schemes whereby a skeletal staff, taking all social distancing precautions, could be deployed to restart the observations from the control room, and also run the maintenance activities.   The data gathered is transferred to the users electronically, and all meetings and discussions moved to video conferencing.  Luckily, our plans worked, and we have been able to run the observatory in this fashion almost uninterrupted since then.  However, some of the development works and activities that our team members had planned did suffer delays; but fortunately none were such that they significantly impacted our plans for the GMRT. 

Similarly, the work of the research scholars and astronomers using the data from the GMRT also faced some delays and set-backs, but most of them are now back on track with their plans. Going forward, remote interactions is going to be the way of functioning, and all of us need to get accustomed to it!

AG: In your opinion, what are some of the most exciting areas of research in signal processing in your areas of work, for students and upcoming researchers?   

YG: Clearly we are entering the era of big data in my area of work also – the rate and volume of data generated by a modern radio observatory is mind boggling, and we will need smart ways to process the data, store the outcomes, and sift through these to extract our science results.   Plenty of fancy signal processing is still needed to achieve much of this; but most of it will move / is moving from hardware platforms to software based techniques, from the traditional approaches to more data driven approaches, to using more and more of machine learning and artificial intelligence techniques.   Some of the most exciting new innovations in radio astronomy will come from applications of these frontline technologies and ideas to the faint signals we receive from the distant reaches of the Universe.   Another thrust area of signal processing is going to be in the realm of trying to detect extra-terrestrial intelligence: can our techniques match up to the requirements of this vastly unexplored domain?  – that will be a billion-dollar question!    There is still plenty of excitement to attract the students and young researchers in this ongoing quest to unravel the mysteries of our Universe.