The iVEC summer Research Internship program provides an excellent opportunity for undergraduate students to undertake cutting edge research in the application or development of advanced computing. Successful students will take part in a 10 week+ program including an induction week on advanced computing techniques, and culminating in a presentation at the iVEC eResearch Forum 2010. Students must be third year and above, studying at an iVEC member institution. Full details of the program and how to apply are available at http://ivec.org/interns.

Applications close 14th October 2009

We encourage students to consider projects at any of the universities and CSIRO.

The rapid progress of 3D visualisation technologies has resulted in their widespread use in a myriad of applications such as geospatial visualisation, surgical planning, design and entertainment. Although widely adopted in the area of visual analytics, little research has been carried out to quantitatively assess the impact of 2D and 3D visualisation modalities on the key task of object (target) interpretation (Carvaial, 2005). 'Target' here can refer to any object in an image, 2D or 3D dataset that is of particular interest to the viewer, which includes: geological features, seismic attributes, medical tissue diagnostics, financial trends, surveillance subjects and military targets. Currently it is difficult to ascertain “how much”, if any, a target spotter in a chosen field of expertise can gain in terms of “understanding” the scene/object when it is presented as a 3D and 2D visualisation.

Lately, advances in EEG signal processing (Hyvarinen, 1999) (Georgiev, 2004) have made it possible to effectively deal with the mapping of these complex signals to external evoked stimuli. The result of this development is the gradual introduction of effective brain computer interface applications in the form of brain activated text communication, control of wheel chairs etc. As EEG signals are analysed in terms of their response to visual stimulation, we will adapt the Rapid Serial Visual Presentation (RSVP) method used in Mathan (2006) to capture the Evoked Response Potential (ERP) of the subject in both single trial and multiple trial classification settings. The ERP can then be used to assess if a particular visual stimuli (e.g. 2D or 3D) has a positive or negative impact on the observer. Whenever possible, we will complement the EEG with data from the eye tracker to explore the visual behaviour associated with the cognitive event and in particular determine the focus of attention (where the subject is looking) corresponding to the EVP signals.

This proposal is an extension of a current Honours research project entitled “Improving image sorting via fusion of electroencephalography and eye tracking” that is investigating target detection and localisation. The eye tracker (MobileEye) used in this project was kindly loaned by iVEC.

In addition to the MobileEye eye tracker, this project will be using iVEC’s 3D visualisation platforms for the object presentations.

Required background: Computing skills, experience with Visual studio and Matlab. Preferably someone with EEG signal processing experience
For more information, view the Project Proposal.

Back to project list

Through the policy of mandatory detention, asylum seekers have been marginalised and isolated from the public sphere. This is nowhere more evident than on Christmas Island where the major detention facility is a high security immigration detention centre where few outside visitors or media have visited. Although some information is publicly available (see for example a "virtual tour" of the detention centre at http://www.newsroom.immi.gov.au/videos/17 ) there is scant information on the "human" element and what takes place inside the electric fences. Furthermore, the relationship of the detention centre to the context of Christmas Island is little known.

The project aims to challenge the "out of sight, out of mind" approach to immigration detention. Information will be provided to the student to guide him or her in developing visual imagery which both explains the island context and the detention domain. Although the project will evolve through supervision suggestions for what could be included are presented below.

The island

• Geographic location and isolation
• Mapping the island and the detention centre
• Ethnic composition of island
• Flora and Fauna
• Arrivals "boat" people and "plane" people (immigration staff and private contractors)

Back to project list

Open source physics engines are used extensively in animation, games, and simulators. Physics engines perform a number of tasks ranging from numerical integration, collision detection and response, many of these tasks are difficult to execute efficiently in parallel. Recently some physics engines have been extended to enable parallel execution on SMP computers, however are typically targeted at dual or quad-core. Further extensions to manycore architectures are necessary to achieve realtime performance for large environments or complex simulations.

This project will investigate if the parallelization approaches taken by these engines will scale to many-core (16+) architectures and analyse the speedup gained by the different parallelization techniques employed by each engine. These results can then be used to improve the physics engine and to build large-scale simulators.

The open-source Physics Abstraction Layer (PAL) software provides a uniform cross-platform interface to multiple physics engines and provides a set of benchmarks for evaluation physics engine performance. The software was developed by Adrian Boeing, and won the SONY/Eurographics game physics prize in 2007. It is currently used in a number of open source robot simulation packages and runs on Windows, Linux and Mac OSX.

This project will port the PAL software and the PAL benchmark to a shared-memory supercomputer and evaluate the computational performance. The open source physics engines that will be investigated include ODE, Tokamak, and Bullet.

Required background: Bachelor (CS/Engineering) or final year.
Computing Skills Essential:
- Knowledge of *nix and strong understanding of compiling software for *nix. (eg:Makefiles or premake/lua, shared objects, libraries, etc.)
- C/C++ development skills
Highly desirable:
- Experience in porting software / cross platform development
- Experience with developing parallel software
For more information, view the Project Proposal.

Back to project list

Brain-Computer Interface (BCI) is a communication platform that translates brain signals into computer commands to affect communication and control of external interfaces and devices. This process bypasses the normal motor neural pathways which makes it a useful aid for paralysed but cognitive active patients.

Over the past decade, advances in statistical signal processing (Hyvarinen, 1999) (Georgiev, 2004) and clustering and classification algorithms (Barros, 1998) (Mansour, 2000) (Xu, 2005) have been instrumental in providing effective means to analyse EEG recordings and to determine EEG patterns associated with the human motor movement (Guan, 2008). The objective of this proposal is to develop techniques to classify various aspects of limp movements from EEG signals alone. The main feature of the proposal is the development of a motor-visualisation presentation approach that will help translate the mental imagination of movements into good quality brain signals that can be subsequently recorded by the EEG sensors. iVEC’s 5DT Data Glove and the Inertia Cube motion sensor will be used to provide the ground truth data for training the EEG signal classification system. We will study the performance of various feature classification techniques like support vector machines, neural network, decision trees, naïve bayes and etc available from WEKA, a data mining software. Because the offline training of the classification system can be computationally involving, the proposal will harness the super computing facility of the ARRC facility to speed up the training time (using WEKA Parallel). Finally, the project will investigate the online adaptation of the system to actual EEG input without the presence of motor stimuli. The outcome of this research will provide future pathways to control the actuations of an exoskeleton suit from thoughts.

This proposal is an extension of the Hons research currently investigated by Daniel Paoliello. The 5DT Data Glove and the wireless Inertia Cube motion sensors used in this project are kindly on loan from iVEC.

Required background: Computing/Electronics Engineering. Experience in Visual studio and Matlab. Preferably someone with EEG signal processing experience.
For more information, view the Project Proposal.

Back to project list

Simultaneous Localization and Mapping is extensively used in robotics to simultaneously build a map of the environment and accurately determine the robots location. Mapping involves interpreting uncertain data from the robots sensors to determine what the world looks like from the robots point of view. Localization involves determining the position of the robot from a map. Thus, these problems are an interconnected “chicken-and-egg” problem, but can be solved using state-estimation techniques, such as an Extended Kalman Filter (EKF) to correlate the mapping and position data.

In small environments the amount of data required to perform SLAM is minimal, however for robots roaming large areas (eg: 500x500m) a significant amount of data is generated. This requires the computation of very large jacobian matrices to solve the SLAM EKF problem.

This project will profile a EKF SLAM implementation and develop a GPU-based EKF. The code will be benchmarked to determine the speedup gained from a GPU implementation. These results can then be used to improve standard SLAM approaches, allowing robots to map much larger areas.

The open source SLAM code available at http://www.openslam.org/ will form the basis of this project. The UWA Computer Vision Group also have a complete CPU Visual-SLAM implementation and have indicated they will be able to provide assistance.

Required background: Bachelor (Engineering / Computer Maths) or final year.
Essential:
- Engineering / Computing Maths background
- C/C++ development skills
- Strong math skills
Highly desirable:
- Experience with developing GPU software (OpenCL/CUDA)
- Knowledge of DSP/filters/control engineering (eg: Kalman filter)
For more information, view the Project Proposal.

Back to project list

Reconstruction of tilt-axis tomographic datasets obtained with Energy Filtered Transmission Electron Microscopy (EFTEM) can be a time consuming process requiring significant compute resources. However, the resulting volumetric datasets are enabling scientists to explore materials with complex topology and composition in ways which are difficult or impossible with other imaging modalities.

This project will establish a semi-automated distributed workflow for performing a tomographic EFTEM reconstruction using distributed parallel processing. This will require developing the system infrastructure to allow clients to submit tilt-axis EFTEM image sequences to be processed via a Grid-based batch queuing system and implementing a novel tomographic work-flow pipeline to produce fully reconstructed volumes. The project will cover tomographic re-construction algorithms, EFTEM data processing, as well as grid-based data and workflow management (including complex job dependencies).

Required background: BSc in Computer Science, Mechanical Engineering, NanoTech or related.
Experience with microscopy (eg. SEM, TEM)

Computing Skills:
- Unix/Linux experience required, including scripting and system configuration.
- Interest or experience in distributed computing, scheduling and batch queuing systems.
- Previous experience with web-based technologies (php, javascript, ajax, etc).
For more information, view the Project Proposal.

Back to project list

In this project, we aim to study the quantum dynamics of electrons confined in nano-structured devices under the influence of external magnetic field. In particular, we will be searching for the signature of quantum chaos, as the systems under study would be chaotic in the classical re-gime.

This project will require vast amounts of computational time as the spatial grid required to fully resolve the momentum details in the time propagation of electron wavefunctions is expected to be large and the spectral analysis requires sampling at a very large number of time steps. Effi-cient management of memory usage will also be crucial.

Required background: B.Sc major in physics or computational science
Fortran programming, basic Unix, general plotting tools, knowledge of computational Libraries
For more information, view the Project Proposal.

Back to project list

Atherosclerosis is responsible for development of coronary artery disease, the leading cause of death in Australia. Atherosclerosis is a chronic, inflammatory, fibroproliferative disease primarily involved with large and medium-sized conduit arteries. Although the entire coronary circulation is exposed to systemic risk factors, coronary atherosclerotic lesions primarily form where there is disturbed coronary flow, specifically bifurcations such as the branch point of the left anterior descending and left circumflex arteries (the two main coronary arteries providing blood supply to the heart). These focal lesions result primarily from haemodynamic forces, with endothelial shear stress (ESS) playing the fundamental role. In arterial lesions with disturbed flow, low ESS induces endothelial dysfunction and exposes the endothelium to atherogenic effect, resulting in formation of the atherosclerotic plaque.

The proposed study aims to investigate the noninvasive evaluation of coronary artery endothelial shear stress with the goal of achieving noninvasive coronary vascular profiling as an adjunct to the standard morphology assessment from cardiac CT imaging. Data concerning plaque morphology (calcified, non-calcified or mixed type of plaques, local or extensive plaques along the coronary artery wall) and the haemodynamic environment in which the plaque is located can be used to predict the development of high-risk plaque which is likely to progress to plaque rupture and precipitation of an acute clinical cardiac event.

Computational fluid dynamics (CFD) evaluation of endothelial shear stress will be performed using OpenFOAM® simulations carried out on arterial geometries extracted from CT scan data and converted to appropriate input meshes.

It is expected that the following research outcomes will be achieved from the study:

• Risk stratification of individual coronary plaques that extends beyond the morphologic assessment now available with standard cardiac CT data; identification of high-risk of rupturing plaques;
• In vivo understanding of the local haemodynamic environment responsible for future behaviour of individual plaques;
• Potential clinical applications of identifying and treating individual high-risk, non-flow limiting coronary lesions.

Back to project list

Cryogenic preservation of biological tissues at very low temperatures seeks to avoid the damaging formation of ice by using so-called cryosolvents that promote the vitrification (glassification) of water. Glycerol, ethylene glycol, propylene glycol and ethanol are widely used cryosolvents whose mechanism of action is still poorly understood. This project will use molecular dynamics (MD) simulation methods to predict the transitions of supercooled water to the glassy state in aqueous mixtures of these cryosolvents, and to investigate the structure and dynamics of these solutions. The elucidation of the mechanism of solvent vitrification will allow the future rational design of optimal multi-component aqueous mixtures of solvents with improved vitrification and, thus, cryopreservation properties. This will have applications in areas as diverse as the freezing of eggs and embryos and the preservation of germplasm from endangered plant species.

MD simulation methods are based on highly parallelised algorithms that are ideally suited for high performance computing facilities with multiple CPUs. This project will make use of the MD program GROMACS, which can run at both the ARRC and Informatics facilities in order to benefit from its nearly linear scalable performance. A number of simulations will be carried out under different concentration and temperature conditions, as well as using ‘annealing’ algorithms to represent fast cooling/heating. The resulting ‘trajectories’ will be analysed to determine the thermodynamic and structural properties of aqueous solutions of cryosolvents. Animations of these simulations will illustrate the ability of these methods to capture molecular behaviour in atomistic detail and to exemplify the power of high performance computing to achieve this. This project will thus encompass the areas of scientific computing, visualisation and data storage in a high performance computing environment.

Required background: Undergraduate studies in physics, chemistry, nanotechnology and/or materials science
PC or Mac, preferably some experience with molecular modelling software.
For more information, view the Project Proposal.

Back to project list

We have developed a generic code, convergent close-coupling (CCC), for electron/positron/photon collisions with atoms. It is the only one in the World that is valid irrespective of the projectile energy, and for both excitation and ionisation processes.

Presently, the CCC code is being modified to make it suitable for distributed memory architectures. One project would be to test and improve the efficiency of this implementation.

The original implementation of the CCC code is within the non-relativistic formalism. We now are in the process of developing a fully relativistic version. Another project would be to contribute to the development of this code. More specifically, we require an MPI implementation, as well as some small self-contained problems to be coded.

We are also developing a new CCC code for electron-molecule collisions. Hence another project would be to contribute to this code development. This is at the very early stages and an organisation around an MPI implementation would be very useful.

This is an ideal time for our future Honours students to join us and learn a bit about code development and application to real physics problems of interest to the international community. I have three Honours (in 2010) students, Mark Zammit, Jeremy Savage and Arwin Karlon who have expressed considerable enthusiasm in joining our group. Their background is ideal for undertaking the above-mentioned projects. If some extra funding was required to enable all of them to participate then I would be happy to contribute.

Required background: Complete 3d year University with a strong record in mathematics and physics
Some coding experience of scientific problems would be ideal
For more information, view the Project Proposal.

Back to project list

Boundary layers form on the surface of structures subjected to a flow. In the case of aerofoils, the development of the boundary layer, its separation from the surface and reattachment can significantly influence its aerodynamic performance, ie its lift and drag characteristics. Active flow control refers to the process of controlling the development of the boundary layer and the separation and reattachment of the boundary layer to gain improvements in aerodynamic performance (or hydrodynamic performance in the case of hydrofoils).

A recent student project within our School included the development of a code for the numerical simulation of active flow control of a boundary layer by a zero-net-mass-flux jet. The code was written in Matlab and, while the code was successful in simulating the flow to some extent, the size of the simulation in terms of the number of fluid elements that was practical in Matlab did restrict the usefulness of the code. Apart from restricting the number of fluid elements, the simulation was also restricted to two-dimensional flow.

The proposed student project for the iVEC Internship is to convert the code to a language that is compatible with the iVEC facility, such as C/C++, and to parallelise the code to take advantage of the high performance computing resources available.

The data from the simulation would be time resolved and transient. Initial data would be two dimensional although part of the proposal is to extend the simulation to three dimensions. A complete understanding of a fluid flow can be greatly enhanced by visualising the flow data, and we would like to also use the visualisation facilities available at WASP to further investigate the data and the representation of the data from the numerical simulation.

Required background: Engineering, maths or computer science
Student should be able to transfer the code from matlab to a language that is compatible with the supercomputing facility (eg basic C/C++ required)
For more information, view the Project Proposal.

Back to project list

Density functional theory (DFT) is the standard approach to computing the electronic properties of condensed phases and plays an important role in the understanding and development of new materials at the atomistic level. Although considerably faster than traditional wavefunction-based methods, density functional calculations are still computationally demanding, especially when applied to complex systems.

The aim of the present project is to explore whether GPU processors might represent a viable approach to significantly accelerating the performance of DFT calculations. This project will specifically explore whether some of the computationally most demanding operations performed within the SIESTA methodology can be adapted to GPUs. In particular, this will focus on sparse matrix operations and eigensolution methods that are used during the determination of electronic states.

Required background: Computer science / mathematics
Previous experience in GPU programming.
For more information, view the Project Proposal.

Back to project list

This Project has 3 aims:

• To further develop a custom discrete particle solver in OpenFoam, in order to simulate particle nucleation and particle and fluid dynamics
• To apply the custom solver to novel – realistic scale – filtration problems, for which an HPC is required.
• To conduct a detailed study into the scalability of OpenFOAM on the Murdoch Informatics Machine.

Back to project list

The prediction of fluid flow through fractured media has applications in petroleum production, water resources, geothermal energy, storing of high-level nuclear waste and in situ leaching. Despite these applications, fluid flow through rock fracture networks is not well understood, largely due to the inherent difficulties in identifying and characterising fracture systems.

Whilst studies have been conducted investigating the relations between some fracture network properties and hydraulic properties, little work has studied the influence of the connectivity of fracture systems on large-scale hydraulic properties

Using explicit discrete fracture models to represent fracture systems, it is proposed to study the relationships between the geometric properties of fracture networks and the connectivity of the fractures using percolation models.

This workflow requires running large percolation models on Cognac at iVEC’s ARRC facility. Further to this the project will investigate the effect of the interconnected nature of the fractures on the large-scale permeability of the fracture network.

The internship would support the Final Year Project of Ben Wisbey, enrolled in UWA’s School of Mechanical Engineering.

Required background: 4th year Student in Mechanical engineering.
For more information, view the Project Proposal.

Back to project list

AuScope Grid is enabling the geoscience communities to use live data sources via interoperable OGC web services. There are currently no platform independent technologies which allow for the data providers used as input for simulation tools. The project would be to build a data provider processor which could request, translate (via shared templates), and output the data in a suitable form to be used in computational simulations. The proposed geologic use cases will be run by Guillaume, and involve requesting spatially referenced rock properties as inputs into a real-world model for Gale simulations.

The same pattern can be used across many science fields, as more and more fields (astro, marine, health, government) are beginning to embrace OGC web service delivery for spatial data, the need for this connectivity technology will increase, and the concepts are completely reusable. Designing and building the generic solution to this information processing tool will enable the data to be used in the specific cases via configuration/templating, and will provide links into other communities for its reuse.

The project is best suited to a Software Engineer or someone who wishes to design and engineer both the tool and its application into the science world. The project will be supported by a large group of software engineers in AuScope, and engage the larger geosciences community via AuScope.

The produced code will be exported into the wider communities, and licensed under an ‘Apache Style’ or permissive open source license. The Student will have the option of remaining as an active member in the development team, as the software is intended to live beyond the life of the internship.

Required background: Software engineer or CS or IT
Experience with Python and XML.
For more information, view the Project Proposal.

Back to project list

GeoPIXE (http://nmp.csiro.au/GeoPIXE.html) incorporates the Dynamic Analysis method for real-time PIXE/SXRF spectral deconvolution. Currently this software package is PC bound and processing can be time consuming and problems analysed are limited to the resources of a PC. There is benefit in “speeding” this process up and making it possible to conduct parallel analysis – something that HPC resources can provide to applications.

This project will consist of two parts:

• migrating GeoPIXE to work on a HPC resource and making it accessible via the Australian grid.
• Augmenting a client program to submit GeoPIXE jobs to the grid for processing and retrieval of results

Back to project list

The distribution of neutral hydrogen in the Universe is of fundamental importance as both a tracer of the underlying Dark Matter across cosmological distances and the basic fuel for star formation and galactic growth. A ground-breaking new radio telescope in Western Australia, the Australian Square Kilometer Array Pathfinder, will revolutionise our view of this key quantity. To both predict and understand the results from this facility we need high resolution simulations of significant volumes of the Universe. Just such a series of simulations, with the accompanying datacubes, have been created in an international collaboration between researchers in the Netherlands, UK and ICRAR. This project involves the visualisation of these simulation in a manner that is able to cope with the large spatial dynamic ranges together with the orders of magnitude variations in the signal of the neutral hydrogen. The outcome will be a series of two and possible three-dimensional movies which will allow researchers to better comprehend the simulation and better present the simulation output to an audience.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

We are developing a new approach to data correlation, whereby the majority of the effort is performed at the antenna site inside the existing data sampling hardware on GPUs. This program will develop a truly distributed FX-correlator, in collaboration with the Australian Government research organisation, CSIRO and iVEC. Modern Radio Astronomy correlators work in frequency space where the data from each antenna is Fourier transformed (F) and then the frequency spectrum is cross-correlated (X). It is now easier for the conversion to frequency domain (known as channelisation) to be implemented in GPU hardware at the antenna site. This distributed FX correlator is a completely new approach, and potentially offers significant power and processing savings for all Radio Astronomic applications, such as the SKA correlator. The student project will be to develop an autocorrelator to be installed at all Australian VLBI antennae. This autocorrelator will the be the platform for developing the channelisation for the next generation cross correlator.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

The discovery of fast transients opens a new field of astronomy which has never been previously explored. The search for these fast transients will require an order of magnitude improvement in existing methods and approaches. The student will test a number of transient detection methods to investigate their success rates and sensitivities, and to determine the sampling parameters needed for applications in the field. These tests will be run on high level development platforms, such as Mathematica or Matlab, which will allow us to investigate the computational costs and complexities for the test algorithms. The conclusions from the project will discover the most effective algorithm, which will be implemented in an optimised form in future work.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

Molecular hydrogen is the basic ingredient in star formation and is seen to follow a well defined mass function, a result that will be investigated to extraordinary precision and accuracy with the successfully launched mid-infrared telescope Herschel. In order to understand the results from this billion dollar instrument an international collaboration involving the Netherlands, UK and ICRAR has created some of the highest resolution hydrodynamical simulations ever made. The project will demand knowledge of both IDL and astronomy, the outcome of which should be a unique understanding of the distribution of molecular hydrogen under the influence of a range of astrophysical phenomenon including Supernova and Supermassive Blackholes. Additionally, the project will result in a testable prediction for the evolution of this key quantity over cosmic time, of quality and interest sufficient to be published in a peer-reviewed journal.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

Supernova 1987A, which exploded in a satellite galaxy of the Milky Way, was the brightest supernova since the invention of the telescope. It has been the subject of much observational and theoretical research in the intervening 22 years. Moreover, the remnant of the supernova continues to get brighter at radio and other wavebands as it collides with neighbouring gas clouds. A recent series of hydrodynamic simulations has been conducted by researchers at ICRAR in order to better understand the source of the radio emission, and to predict its future evolution. The aim of this project is to help us to better visualize the supernova simulations which are very large in both time and spatial dimensions. The student will conduct a review of possible visualization methods and work with researchers to produce realtime and pre-computed visualizations in two and three dimensions in order to assist with the characterization of the simulations and comparison with the observations.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

The current paradigm of the Universe is one in which a gravitationally repulsive force termed Dark Energy is driving apart the large scale structure, of which gas and stars (baryons) are a subdominant matter component, the majority of which is in the form of Dark Matter (DM). The DM forms gravitationally bound structures, termed haloes, which merge to build up progressively larger sized structures. This is the hierarchical Lambda CDM paradigm. The merged haloes are prone to disruption; however due to the very nature of the DM such processes can only be probed by simulations run on supercomputers. We are interested in the role in which the hitherto neglected baryonic component influences this disruption. To that end a number of high resolution cosmological simulations have been completed within an international network of the UK, The Netherlands and Australia. These simulations include feedback from supernovae as well as supermassive black holes and are currently amongst the best resolved in the world. Within the allotted time the student should have a definitive answer on the scale at which these processes determine the survival of DM haloes within the current cosmology. Ideally the prospective student should be proficient in IDL, with knowledge of simulations; however neither are a prerequisite.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

Transient sources are very useful in the world of astronomy. The regular signals from radio pulsars have been used to map the electron content of the Milky Way, test Einstein's theory of gravity and investigate the properties of ultra-dense matter. Unfortunately, performing blind surveys for pulsed or transient sources is very computationally intensive. This project will investigate the use of a distributed computing system (similar to BOINC) called Nereus-V. We will develop a pipeline to process data from a prototype transient telescope in WA using a local server and a network of donated CPUs.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list

Meteor showers happen at various times throughout the year as the Earth passes through the debris trails left behind by comets and asteroids with Earth-crossing orbits. But not all meteor showers peak during the night. Some showers are known to have such short peaks that, if they occur during daylight, we miss them altogether. A radio monitoring system can observe 24 hours a day and pick up meteor streams that might otherwise not be detected.

While meteors can be very bright optically, they are not radio emitters. One way they can be detected is using a method known as forward-scatter, a form of bistatic radar where the transmitter and receiver are physically separated. Rather than picking up a signal from a meteor itself, the receiver detects the reflection of a strong signal (such as a TV station) from the ionised trail created as a meteor passes through the atmosphere.

The aim of this project is to construct a simple meteor monitoring station, find a suitable frequency to monitor and implement a method of collecting and analysing the data.

Required background: At least two full years of their undergraduate studies in physics, astronomy, computer science or a relevant engineering discipline.
See also http://www.icrar.org/outreach-education/scholarships

Back to project list