My Master's by Research degree looked at ways of localising RFID tags from a mobile robot as part of the ENRICHME project funded by the EU (www.enrichme.eu).

At the time of writing, the research is currently ongoing. A paper produced from this research has been peer­-reviewed and was accepted for publication at the International Workshop on Intelligent Environments Supporting Healthcare and Well­being (WISHWell'16) in London this year, with another paper currently under review.

The work makes use of particles filters to combat large amounts of noise present in the readings to create probabilitic estimates of locations. In addition, phase and frequency analysis is factored in to highlight areas of interest. This is fed into the Robot Operating System (ROS) so that these calculations can be performed in real-time and fed into other on-board systems which can make use of the information.

My final year MComp project looked at developing a software image which could be burned onto an SD card, which could then turn low-cost SOC computers (such as a Raspberry Pi) into feature rich home routers. The software was commended by academic staff and was featured in the University of Lincoln Computer Science showcase.

The software looked at the networking hardware present, and would guess a suitable network configuration. As an example, if the router picked up an internet connection available through an Ethernet port, and also a wireless access point-capable NIC present, the router would provide a wireless local area network, and route outgoing internet connections to the Ethernet port.

The modular NIC architecture of the project allowed advanced infrastructure configurations such as wireless WAN ports, something only otherwise available in very high end network routers. The router also provided other features not commonly available such as a full forward proxy server.

As part of my BSc third year project, I constructed a desktop application which allowed data streams to be routed and forwarded effortlessly around a PC. This is amongst the most complex pieces of software I've worked on and was awarded distinction status by the university.

The program could communicate with various different inputs and outputs, such as USB and serial ports, and also network ports with either TCP or UDP, running as a server or client. The program linked these various ports together, and allowed data to stream fluidly between them. As an example, the program could be set up so that various TCP or UDP network ports could be opened, with incoming data being mixed together and then outputted to a serial port.

Architecturally, the program could handle port errors or crashes because of its inherently safe design. Each port was assigned its own dedicated thread which could operate independently. Therefore the program was designed to have many different threads running and could synchronise communication between them dynamically. This allowed for the outward port communication to be blocking, pushing down the CPU usage of the program to next to nothing.

The software was designed with cross-platform use in mind, and could run under Windows, and Ubuntu and Fedora Linux. The project included a suite of testing and diagnostic tools, which simulated the various different types of ports, fed in dummy data, and would occasionally simulate problems such as corrupt data, connection loss, and the port hanging or crashing.