Preparatory studies towards SKA processing architectures

This proposal is for a studentship to develop key systems for the Square Kilometre Array (SKA), to be held jointly by the University of Oxford Dept of Physics, which is one of the leading contributors to the SKA design and specifications, and Selex Galileo, a leading European electronics company. The SKA is a project to build by far the largest radio telescope ever constructed, which will have a transformative effect on many areas of astrophysics and cosmology. The key science goals of the SKA include mapping the development of the structure of the universe by measuring the positions in time and space of a billion galaxies; testing fundamental theories of physics by precision measurements of the most extreme objects in the universe; and studying the formation of earth-like planets. In its final form, it will consist of an array of around 2000 15-m dish antennas, plus around 500 large (~200-m diameter) phased array antennas covering two different frequency bands (totalling over ten square kilomtres of physical collecting area), plus a large central data processing facility, costing a total of around E1.5 billion. The antenna arrays will be concentrated in a remote desert site, with some elements spread over continental distances. The aperture arrays will be the largest digital signal processing networks ever built, with a total input data rate of several petabits per second (greater than the entire current internet) and processing power of many peta-operations per second (comparable to all the personal computers in the UK combined). The period covered by this proposed studentship coincides with the pre-construction phase of the SKA, during which the current plans and outline designs will be converted into functional prototype sub-systems which are capable of being manufactured and installed on an industrial scale. This is therefore the ideal time to have a joint academic-industrial studentship working on a critical aspect of the SKA Phase 1 system design, the low-band aperture array. By using a phased array instead of a large dish antenna, it is possible to image a large number of independent fields of view simultaneously, vastly increasing the survey speed. Once digitized, the signals from each antenna element are broken in to narrow frequency channels, then combined heirarchically into phased beams consisting of weighted sums of all the ~10,000 antenna elements in an array station. Complex gain factors applied to each input element both calibrate out amplitude and phase imbalances in the elements, and generate well-defined beams pointing to different parts of the sky. The project student will study the detailed implementation of the RF analogue electronics, digitization, and initial signal processing of the data streams. The performance requirements for the SKA, in terms of bandwidth, data throughput and volume of output data are far greater than any currently implemented system, and will require innovative design solutions, and a close interaction between science requirements and engineering implementations. Of particular importance is that the designs must be optimized for low manufacturing cost, ease of initial testing, low power consumption, and long service life with little or no maintenance. These are areas where Selex Galileo has vast experience and will bring a major input to the system design. The precise areas of study for the student will be fixed during the initial phase of the project rather than now; the student will not start until October 2011, and the overall system design will have moved on by then, and the scope of work available is also much greater than any one student could cover in a PhD. The supervisors will agree a programme of work which makes best use of the interests and skills of the student and the capabilities of the industrial partner, with the aim of maximising the impact of the project on the SKA design and hence the UK industrial involvement in the SKA construction phase.