The Space Science Behind Road Security
Dr Roberto Trotta applies statistical and data analysis methods developed for astrophysical research to the problem of road security. In collaboration with an industrial partner, Dr Trotta is translating astrophysics research tools for the analysis of in-vehicle monitoring systems data.
Such data set constitutes a unique testbed for a novel data analytical technology. Aims include improving road safety, refining car manufacturers' understanding of the actual behaviour and performance of their vehicles often in challenging, off¬road conditions or poorly mapped areas, and evaluating drivers’ behaviour to improve their safety.
Imperial College participated in the specification, design and operation of the SPIRE instrument. The Astrophysics Group led one of the Data Processing and Science Analysis Software Centres. Research areas included developing data reduction and analysis software, contributions to lab testing of the instrument, instrument health and trend analysis, and in flight calibration.
The primary science interest is in mapping the light emitted by stars and black holes, that has been reprocessed by dust. This enables a census of the energy outputs of galaxies at these wavelengths, over the history of the Universe.
Herschel will achieve these aims by making maps of particular fields within two surveys, ATLAS and HERMES. Notable results include the discovery of the most distant rapidly star forming galaxy known, and uncovering previously unsuspected high redshift groups or clusters of star forming galaxies.
High Efficiency Space Solar Cells
Dr Ned Ekins-Daukes leads research into highly efficient solar cells in the Quantum Photovoltaics group. This research includes the development of multi-junction solar cells that are used routinely in space arrays.
The group is presently investigating candidate semiconductor alloys that would enable quad-junction solar cells to be made with an AM0 efficiency of 35%, as well as developing solar cell device designs that are highly radiation resistant.
Real-time Optimization on FPGAs
The Control & Optimisation team of Dr Eric Kerrigan, and Digital Computing team of Professor George Constantinides, have been working together to design computing architectures and optimisation algorithms for optimal control and estimation problems arising in aerospace.
They have also developed analytical methods for ensuring reliable operation under reduced precision arithmetic. Implementation of the proposed architectures in FPGAs show that satisfactory control performance at a sample rate beyond 1 MHz is achievable, even on low-end and low-power processors.
Robotics for Space Exploration
The Aerial Robotics Lab, led by Dr. Mirko Kovac, is developing next-generation robots for exploration of unknown and challenging environments. Current research projects focus on (i) multi-modal mobility using combinations of aerial, aquatic and ground based locomotion and (ii) autonomous construction of structures with swarms of flying robots. The long term vision of the lab is to create fully self-sustained robot colonies, may it be on earth or in space, which will build their own habitats and act as exploration tools for humanity.
LISA Pathfinder, an ESA mission launching in 2015, will demonstrate technology required for ‘Cosmic Vision 2015-2025’, ESA’s third large-class missions programme. LISA is a gravitational wave observatory in space. On-board LISA Pathfinder, 2-kg masses are kept in perfect free-fall below 10-13 g. It will open a new astronomical window on the universe, and test Einstein’s theory of General Relativity with unprecedented accuracy.
An Imperial team, led by Professor Sumner, has developed and manufactured hardware to maintain the electrical charge of the masses at zero using photoelectron emission under UV illumination. Imperial scientists will also be part of the operations team carrying out six months of experiments to assess the performance of the spacecraft in-orbit.
Earth Observation and Remote Sensing
Case study to come
Array Signal Processing and Array Telecommunications
Professor Manikas’ research explores how space sensors - working together as large flexible sensor/antenna array systems - could enhance space technology, earth observation, remote sensing, and telecommunications.
For instance, a group of satellites or Unmanned Aerial Vehicles (UAVs) can be treated as a sensor array, receiving spatial and temporal information in the form of multiple paths or multiple telecommunication channels. This enables spatiotemporal SIMO/MIMO array telecommunication techniques to augment the dimensions of the system. Therefore, resolution, telecommunication quality-of-service and link capacity are all improved.
Positioning, Navigation, Timing
Protecting Global Navigation Satellite Systems
Dr Wolfgang Schuster is leading the development of a sensor network to protect Global Navigation Satellite Systems (GNSS). GNSS is used practically in every aspect of our lives for positioning, navigation, and timing. It is currently being taken for granted, despite being highly vulnerable to interference.
Significant GNSS failures are a matter of national security: 6-7% of GDP is dependent upon GNSS, including power grids, telecommunication networks, transport services and financial trading systems.
The proposed network will combine new resilient hardware, novel context and failure-mode adaptive integrity monitoring techniques, and new failure correction models to maximise GNSS availability.
It uses an integrated approach to GNSS user protection based on novel approaches that combine internal (within receiver) and external monitoring techniques to improve failure detection, identification and mitigation.
Computing and Data Science
Professor Julie McCann researches how frugal computing techniques can contribute to very low-cost science, educational and commercial activities. Redundant arrays of inexpensive and disposable thin-film, printed circuit board, and CubeSat spacecraft are used.
Research areas include efficient energy harvesting; using self-optimising algorithms to overcome resource constraints and manage low power peer-to-peer communications. Research also focuses on maximising the amount of data captured and reliably returned by an extensive and customisable range of sensors, hosted by spacecraft swarms.
Data Science Analytics
Professor Yike Guo, director of the Data Science Institute, is leading the way in big data and data science research. In the DSI, analytics is applied to a broad range of areas including medicine, urban informatics, human behavioural analysis, and social network analysis, with a particular focus on applications where learning processes take place in real-time and need to be adaptive to changing conditions. Examples include deep learning for traffic pattern prediction and compressive-sensing for fMRI analysis.
Searching the Sky
Dr Daniel Mortlock is pioneering the use of modern statistical methods and machine learning in the search for the most distant astronomical objects.
Modern astronomy is increasingly using large surveys, which provide data on billions of objects in some randomly chosen area of the sky. Identification of the most interesting tiny fraction of this data increasingly requires the use of methods from statistics and data science.
This approach led to the discovery of the most distant known quasar, and will be extended in the future to exciting projects like the Large Synoptic Survey Telescope (LSST).