Computer Laboratory

CST Part IB group projects

Selected design briefs from previous years

Design briefs from 2010

Winner in 2010: Party Line Detection

New online ‘social media’ products become political tools almost as soon as they are released. But not all politicians realise that online content is more dangerous than traditional party newsletters, because it can be analysed automatically. Your task is to apply the Twitter API to analyse whether UK politicians are showing independent thought when they use Twitter, or are simply following the party line. Are they addressing the same issues as others in their party, perhaps simply reusing material from press releases? Or are they adopting ideas from the opposition agenda? Your system should provide voters with a clear summary and trend analysis, perhaps using visualisations of the kind that are shown on TV on election nights.

Critical Care for the World

In wealthy countries, hospital intensive care units often use sophisticated data monitoring and capture systems like IMDsoft MetaVision to enforce procedures and collect information about various drugs and patient vital signs. The World Health Organisation needs to establish more rigorous care and data collection for the kinds of emergency field hospital that deal with outbreaks of new viral infections like Ebola. But such a system would need to be customised for each site, by local staff, to deal with the huge variation in local training and resources. Your task is to design a field- customisable critical care system that could report research data to the WHO, while continuing to operate with the range of network connections, power supply and hardware availability typical in remote regions.

Robot Chessboard (hardware project)

In the first Harry Potter film, chess pieces move themselves around in response to moves on the board. Last year a group project team built a robot draughts player where the pieces were moved by magnets. This year we have sophisticated new robot hardware, and a computer vision system should make it possible to build a similarly magical robot chess set.

Pedestrian Auto-Pilot (hardware project)

Design a ‘wearable’ computer (for pocket or beltpack) that keeps track of your location while walking or cycling, at a much finer accuracy than GPS does, by using inertial navigation techniques. When walking indoors, it should be able to remind you which direction to turn to go to your next lecture or supervision, remind you to buy your lunch when near the canteen and so on. The design should take account of known locations (for example when confirming task completion) to counteract the inevitable drift in inertial guidance systems. Your design will be based on the XMOS compact development system, and will include interfacing that system to components for inertial guidance.

Real Guitar Hero (hardware project)

One of the most irritating things about the game Guitar Hero (though there are several candidates) is that it doesn’t even use a real guitar. There is no reason why you shouldn’t connect the pickups of a real guitar to a digital processor board. And you needn’t play guitar hero either - the board can implement digital filters for standard guitar effects, and it can probably correct for the occasional wrong note. Your task is to build a digital guitar interface that uses signal processing techniques to produce a real musical instrument, not just a game.

Hyper-Resolution Camera

By taking many many pictures of the same scene, it is possible to build gigapixel images, where every new image is aligned with the previous ones and composited together. Develop an application that can be used to create such images, and then to browse them via a mobile or web interface in the style of Google Earth (‘flying down’ to zoom into the image, and then using false perspective to fly across the surface).


Most knowledge-workers (and professional scholars are no exception) have a serious struggle to differentiate routine correspondence from management of the new ideas that make the job interesting. Design an automated personal assistant that filters email, infers rules on the likely response to routine items, and looks for thematic patterns in the rest. This can be used to track research ideas, or even set up new clusters of collaboration. All rules should be customisable by the user, and confirm proposed actions before sending a message that might be harmful. After a little ‘bedding-in’, the customised result might be indistinguishable from a real professor.

Teach your Cat to Twitter (hardware project)

Now that everyone keeps in contact with their friends via Twitter, it’s a shame that your pets can’t send you tweets while you are away from home. This project means they can! Using covert camera products from ZedCam, detect what your cat is doing, and send updates via Twitter or some other Web 2.0 platform. The system should be remotely configurable from the same mobile device you would use to access Twitter.

Spot the Medicine – instant feedback from chemical structures

Computational predictive models that can generate results in fractions of seconds are used in the design of new medicines (e.g. In addition, many databases exist containing further data about previously synthesised molecules. However, this information can only be predicted or recalled for molecules for which the chemical structures are stored on a computer in some form. Chemists regularly have ‘visual access’ to molecule structures (papers, meeting presentations, notebooks) with no option but to make a copy and then ‘draw’ the molecule into their computer at a later time if they wish to find out more information. Your challenge is to design and prototype a method by which a chemist with visual access to a molecule structure can find out information about a molecule within seconds of seeing it, giving real-time feedback to spot potential new medicines.

Integrating Phones with the Web

Networks ( and previously known as Data Connection Ltd) is a UK-based company which is a market leader in providing next generation, voice over IP telephony equipment. Their equipment provides ways to start phone calls, listen to voicemails, read a web addressbook and get lists of phone calls made and received using JavaScript. The purpose of this project is to work out, build and demonstrate a Web mash-up to include those controls with other Web applications such as Facebook, mySpace, Google Maps or anything else you think would make your life and the way you use the telephone better.

Database in the Clouds

‘Cloud’ applications that allow people to create and maintain online shared spreadsheets or word processor documents are becoming pretty common. But there aren’t many good database systems for cloud users. Systems like Amazon SimpleDB and LongJump are aimed at technical users, but it would be good to have a simple, free data management service for use in the voluntary / non-profit / community sector. Your task is to provide extremely simple web-based definition of schemas, end-user query and reporting language, with import/export to formats that are recognisable to non-technical users, such as plain text email, html tables, Word lists and Excel. Imagine your grandparents as the target users, and that they needs this to run membership, event management, collections & libraries for all their different hobby societies and community groups.

Phone Programming for Children

Design a programming language suitable for use by children, with a sufficiently compact user interface that the development environment can be run on a mobile phone. Children should be motivated to use this language to develop applications that are useful and of interest to them - it shouldn’t be like school! The Android platform will be used.

Remote DIY software plumbing

Systems like (from Cambridge) were developed to help remote monitoring of your house from the web. The next step is remote control of your house. Create a web-based programming environment, that will allow homeowners to do DIY software plumbing (scripting and configuration) of home media, energy and control systems.

Projects offered in 2010, but not taken

Accelerated Interaction

iPhone and iPod Touch include an accelerometer, but few Apps make good use of it. Driving and steering games work pretty well, as do some novelty items like beerglasses and spirit levels. But gestures like ‘shake to clear the screen’ are appalling. Your task is to design a useful and usable user interface (not a game) that can be controlled through tilt and gestures alone. Note that iPhone programming uses the slightly odd Objective C language, and the OSX operating system. Interest in learning these, or previous experience, will be beneficial.

Unified Firmware Development in Eclipse (hardware project)

The aim of this project is to set up a new firmware development tool chain within the Eclipse programming environment. It will apply open source technology from Embecosm for communicating with embedded devices, whether as silicon or model. The target will use a System-on- Chip, implemented with the Altera board FPGA and as high level, cycle accurate and event-driven simulation models. The challenge will be in getting the Eclipse debugger to display non-processor state while debugging (e.g. registers of peripheral logic) with any of the targets and packaging the product in a form suitable for delivery to customers.

A Virtual Phillips Machine

Over in the Economics Department there is a Phillips Machine (aka MONIAC), a 1950’s analogue computer that solves the equations of Keynesian macroeconomics using water flowing through a system of pipes, pumps and pulleys to simulate the flow of money through the national economy. Although more advanced mathematical models now exist, they lack the vivacity and didactic appeal of their hydraulic precursor. The aim of this project is to construct a vivid graphic realisation, possibly using JavaScript or Flash, of a Virtual Phillips Machine - an online demonstrator where valves and other settings can be adjusted in real-time via the mouse to replicate the workings of the original.

Hello Theremin World (hardware project)

The Theremin was the first successful electronic musical instrument, like a ‘Hello World’ application for the whole field of electronic music. 100 years later, many artists and musicians use small digital controllers like the Arduino and mBed to prototype new interactive concepts. But it’s surprisingly difficult for them to implement a simple digital Theremin (it’s easy to do it with electronic parts). Your task is to create an open source toolkit, and a sample hardware demonstrator, that would make it easy for new digital artists to create a simple theremin as their ‘hello world’ experiment.