STEM Research Methods (ENGE817) is designed to give students an understanding of the requirements for a logical, well-designed research proposal. Students must write a research proposal and in the case of computing students, undertake a small part of that research. For MCSDF students, this will likely require a small, practical part of the research being undertaken with results discussed and conclusions from the research drawn. It is important to always bear in mind that this is aresearch methods course and not a research course. Therefore, the purpose of the course, and the majority of the marks, will come from producing a research design with well-thought-out research questions, steps in the research and a good understanding of the issues and challenges of the research, partially drawn from the practical component.

A common approach is to design a research project with a number of stages that will lead to a robust research project with comprehensive results, but to only undertake the first few stages and leave the following stages to be discussed as future work. In this way, the pilot study undertakes a small part of the project that acts as a ‘proof of concept ’ for the wider research discussed in the project. It is, however, not mandatory to take this approach. It is important to demonstrate a reasonable amount of research has been done. Trivial or very small pilot studies will not be viewed favourably.

The following topics are suggestions only. You may select one of these topics or you may design your own research topic. The choice is yours.

Topic 1: Forensic reconstruction of mesh networks utilizing trilateration.

This topic requires programming in a graphical programming language of your choice. The point of this topic is for students to understand the complexities of network reconstruction from artefacts recorded during network operation and to highlight the challenges that emerge as the software is written and simulations are run. The example below is from software written in Matlab,but any language is acceptable.

Imagine a mesh network forming where every computer (node) in the network can calculate its distance to every other node in the network. This ‘perfect world ’ communication is useful for the pilot study, but in reality, calculating distances from signal strength is not overly accurate.

Several nodes appear on a grid 2500 metres x 2500 metres square. The centre of the grid is coordinate (0,0). The first node is always placed here. The second node is always placed on the y axis, either above or below the first node. Each node from there is randomly placed. To obtain the x andy coordinates, random numbers between - 1250 and +1250 are used. The nodes appear on the grid and the distances  from  every node to  every  other node  are  calculated using Pythagorean theory. The name of each node (you may utilise letters of the alphabet or numbers) is recorded and the distance in metres to the node is recorded. In this way, a table or array of nodes and distances is created.

The nodes are shown on the grid along with lines and distances to every other node. For the sake of clarity of the figure, begin with 5 nodes but it should be scalable. Your software should accept input from a text file so that various features on the grid can be switched on or off. This includes circles which show the distance to the nodes.

5 nodes with lines and distances

5 nodes with circles, lines and distances

This is the first part of the wireless network plotting, but now the forensic reconstruction must be  done.  This  involves  only  having  the  following  information recorded: node name  and distance from node to every other node. Coordinates of the nodes is not known. To reconstruct the network, only the node and the distances is used. You must demonstrate a thorough understanding of why this is simple on one hand and complex on the other. What will happen if a  communication  distance  of 300  metres  only  is  introduced.  How  will this  affect  the reconstruction and the plotting of the nodes?