SUGGESTIONS FOR STUDENT RESEARCH PROJECTS

By RACHEL CARDELL-OLIVER
Last updated July 2008

Some specific project areas including practical implementations of sensor network applications, formal methods, and software engineering, are listed below.  Also see my CV for publications relating to other areas I am working on, and happy to supervise projects in.  A PROJECT REPOSITORY of proposals and theses for past and present students in the Mobile, Ad Hoc and Sensor Networks group is also available online, and contains some good ideas for new projects.

 

Sensor Network Projects

 

A Matlab Workbench for Analysing Data from Environmental Sensor Networks

A “data graveyard” is a collection of scientific data that is hard to use: it is not fit for purpose.  Sensor networks gather spatial-temporal observations of natural environments at a scale not previously possible, promising new breakthroughs in scientific understanding of these environments.  But unfortunately this data often ends up in a data graveyard because it is highly variable, comes from heterogeneous sources, and may have many errors.  This project is to develop a workbench of tools, based on Matlab, for interpreting and analysing sensor network data.  Interpretation tools will be needed for data entry, normalization and data scrubbing.  Analysis tools include 2D and 3D visualization tools, and tools for investigating relationships between gathered variables.

Develop a workbench for using sensor network data

This project is linked to projects with UWA EcoHydrology Centre of Excellence and CSIRO Land and Water and may attract a scholarship for a professional computing group, or an honours or practicum student with suitable skills.   For some background on the functionality required, see Cathy Rye’s 2005 Honours thesis

 

TinyPortal: support for interacting with sensor networks

TinyPortal is a free software application for managing a wireless sensor network, under development by Mark Kranz of the University of Western Australia. It is an application that provides a network gateway, as it provides the means to communicate with a sensor network through traditional computer networks.  See

http://wsnwiki.csse.uwa.edu.au/index.php/Software:TinyPortal  for more information. 
And also http://www.csse.uwa.edu.au/~mkranz/wiki/index.php/Summer_Project

There are various projects available in this area for people who enjoy systems programming.
Paid field work is available as part of this project  with Prof Keith Smettem in the School of Environmental Engineering.

 

Data Management for Heterogeneous Sensor Networks

Modern environmental monitoring systems are hugely complex systems, containing many different types of sensing hardware and software, measuring different environmental parameters, and changing over time.  Furthermore, sensing is a highly error prone process, and so significant data cleaning may be required to make the raw data suitable for scientific analysis.  PhD candidate Joseph Ziegler is working with CSIRO Land and Water, Perth, and the SWAN Futures project, to develop an active database for managing the collection and analysis of data from complex sensing systems.  There are many more projects in this area, ranging from the implementation of web services to the design of new database query languages for this domain.

 

Dependable Sensor Networks

Obtaining dependable spatial-temporal data is crucial for informed decision making in many fields from environmental science to health care.  Sensor networks offer new technology for gathering such data.  However, the almost universal experience of sensor network deployments over the past decade is of unacceptably low data yields, and “data graveyards” of observation data that proves hard to use once collected.  This project will address the open problem of building dependable sensornet systems, by extending classical dependability methods for the sensor networks domain.  The project involves formal methods, software engineering, and systems engineering for sensor networks.

 

Space and Time Logic for Programming Sensor Networks

Situations are a way of expressing user requirements for sensor networks involving spatial and temporal constraints.  For example, “if chemical levels or temperatures exceed safe thresholds for at least 30 percent of available measurements in room 1.20 within a period of 15 minutes, then activate the extraction fan and, unless a human check is due in the next 5 minutes, register an alarm for that room”.  With Mark Reynolds, I am working on a novel spatial-temporal logic for expressing situations, and on a compiler for generating sensor network programs from situation requirements.  There are many projects in this area from implementation and experimental applications (eg with SunSPOTS) to theoretical projects on new situation logics and their properties.

References:

Space and Time Logic for Programming Sensor Networks, Rachel Cardell-Oliver, Mark Reynolds and Mark Kranz, in proceedings of ISOLA 2006, IEEE
Tim Burrough, 2007 BE(SE) Honours project SENSID on SunSPOTS

Mark Kranz, 2005 BCompSci Honours Project SENSID: a situation detector for sensor networks

 

Verification of Complex Distributed Systems
Investigate the use of tools such as PRISM (probabilistic model checker) and Uppaal (timed automata model checker) for verifying properties of complex systems such as sensor networks.

References:

Formal Specification and Analysis of Performance Variation in Sensor Network Diffusion Protocols, Sule Nair and Rachel Cardell-Oliver, In 7th ACM Symposium on Modeling, Analysis and Simulation of Wireless and Mobile Systems (MobiSWiM), Venice, October 2004, ACM Press, pp. 170-173, http://doi.acm.org/10.1145/1023663.1023694

Evaluating the Impact of Limited Resource on the Performance of Flooding in Wireless Sensor Networks, Patrick Downey and Rachel Cardell-Oliver, In International Conference on Dependable Systems and Networks (ICDSN), Florence June 2004


Environmental Sensor Network Applications

Wireless sensor technology offers a major advance in monitoring basic environmental variables required for assessing the performance of plants and systems in controlling dryland salinity. Not only are the systems extremely cost-effective for field monitoring programs but they also provide unique opportunities to collect high resolution event-based data which can be critical for evaluating plant system designs to maximize water use (e.g. tree belts on slopes that are deployed to intercept perched water generated during rainfall events).

The aim of this project, funded by Motorola and the Salinity CRC, is to develop state of the art, robust and long-lived sensor networks to support environmental monitoring for research into the management of dryland-salinity.  Sensor networks will be designed, deployed and tested for monitoring and studying:

a) water use and micro-climate of native vegetation

b) the dynamics of soil water movement and water uptake by perennial vegetation, especially in relation to rainfall events.

References:
Recent project status: http://wsnwiki.csse.uwa.edu.au/index.php/Projects

An Experimental Evaluation of Temporal Characteristics of Communication Links in Outdoor Sensor Networks, Jingbo Sun and Rachel Cardell-Oliver, in ACM Workshop on Real-World Wireless Sensor Networks, REALWSN ’06, Uppsala June 2006

A Reactive Soil Moisture Sensor Network: Design and Field Evaluation, Rachel Cardell-Oliver, Keith Smettem, Mark Kranz and Kevin Mayer, in International Journal of Distributed Sensor Networks, pp. 149 – 162, Volume 1, Number 2 / April-June 2005

 

Generic Roles for Sensor Network Nodes

Monitoring the performance of sensor networks and allowing user interaction with the network differs from traditional network management because of the energy constraints of sensor networks.  For example, network nodes are usually asleep, only waking periodically to perform actions.  Although there have been a number of special purpose tools proposed, there are no general framework for expressing sensor network policies, or algorithms for their implementation.  An interesting generic scheme for one aspect of the problem, role assignment, is presented by Frank and Romer in SenSys 05.   Open problems include extending the role assignment algorithm for heterogeneous networks, and developing a more expressive language for role assignment and other network management issues.

References:

Algorithms for Generic Role Assignment in Wireless Sensor Networks, Frank and Romer, Sensys 05, www.vs.inf.ethz.ch/res/papers/sensys05.roleassignment.pdf

 

For further sensor network project suggestions please also see http://wsnwiki.csse.uwa.edu.au/index.php/Development

 

Software Engineering Projects

 

See Dependable Sensor Networks above (this is an SE project too)

 

Using automatic testing to support student learning in Java courses

Tools such as JUnit4 provide automated support for writing and executing test cases for Java programs.  In this project you will develop and evaluate on-line tutorials using automatic test tools to helping students who are learning to program.  Automatic tests can be used to provide immediate feedback, to help in understanding a program, and can also be used as an aid for marking assignments.  These tutorials will use the JUnit4 test tool and an IDE such as Eclipse.

References:

David Hng, 2007 BE(SE) Honours project

Philip Armour, The Laws of Software Process: A New Model for the Production and Management of Software. Auerbach Publications, 2004 (Armour argues that software engineering is an activity which consists of `effectively acquiring knowledge' and 'transcribing it into the active form'.)
Sui Jimm Boh, 2007 BCompSci Honours project

 

Capturing Rationale in Software Projects

Following Armour’s idea of software development as encoding knowledge, this project investigates ways to document decisions and their rationale in software projects.  The argument mapping tool, Rationale, will be used as a starting point.  An industry partner with some non-trivial software projects for investigation would be a great asset for this project.

References:

Bruegge and Dutoit, Object Oriented Software Engineering, has a good chapter on rationale.