ELEC6255 IoT Networks

Coursework 2025/26

Assignment Set:                 Thursday 09th October 2025, during lecture

Assignment Type:               Individual coursework

Submission Deadline(s):   Fri 12th December 2025, by 4pm (written report)

Feedback:                           We aim to return your mark and written feedback by Friday 16th

January  2026,  though  will try to  get  this  to  you  before  the examination.

Mark Contribution:            This assignment is worth 40% of your mark for ELEC6255

Required Effort:                  You should expect to spend up to 60 hours on this coursework

Examiners:                          Professor Geoff Merrett and Dr Alex Weddell

Learning Outcomes

Having successfully completed this coursework, you will be able to:

•   demonstrate understanding of the principles of layered networking models, architectures and protocols which enable IoT networking

•   use simulation to test and evaluate networking algorithms, protocols and architectures

•   find, read and evaluate technical literature, and interpret standardisation documents

•   communicate your technical work

Task Summary

A billionaire property owner has contacted you to provide an IoT smart lighting system for one of their mansions. They want you to investigate and provide some assurances about the reliability, latency, and additional energy consumption of such a system. Your task, in this coursework, is to use OMNeT++ to simulate, evaluate and analyse the performance of such a smart lighting network, based loosely around the ‘Hue’ smart lighting system.

This coursework is solely assessed through your individual report, which must be submitted along with your source code. The report must document your investigation into this system, using network simulation as a tool. This is not a ‘lab’ exercise: we are not posing an exhaustive list of specific questions that we want you to answer, we want you to perform your own research and investigation of a system within the scope of the coursework brief, and this is reflected in the marking scheme.

Academic Conduct and Responsibility

Please ensure that you are aware of the University’s regulations on academic conduct and responsibility, particularly with regard to plagiarism (of written text or code) and collusion. This is an individual coursework and the report that you submit must be your own work. Any use of AI-based text- or code-generation tools, or collaboration with any other individual, must be clearly declared and is likely to result in a reduced mark. To avoid any potential issues with recycling, any student who is repeating the coursework should talk to the module leader (Professor Merrett) before starting it.

Full Task Description

Part 1: Model your Network

Your first task is to create an OMNeT++ network using the IEEE 802.15.4 PHY and DLL layers, and the AODV NET layer.

You can find the floor plan for the property that you are to model in Appendix 1 (listed by name), at the end of this document. You do NOT need to model this building perfectly, and can estimate the sizes and locations of rooms from the floor plan. You should arrange the following nodes throughout the property: 1 ‘hub’, 28 ‘smart light’, and 7 ‘light switch’ nodes. You can place them where you like, ensuring that you spread them out all around the property.

Each light switch controls all of the smart lights that are located in the same room as it. All packets sent from light switches get transmitted to the hub (possibly being routed via other light switches and/or smart lights). The hub uses a lookup table to identify the addresses of the smart lights that are located in the same room as the light switch, and then transmits packets to each of them (possibly being routed via other light switches and/or smart lights).

Your simulation should approximately (but realistically) represent the system that you are modelling. You should make appropriate and justifiable modelling decisions, including:

•    Consider your choice of parameters in the physical layer  (e.g. radio transmit power, sensitivity, data rates, frequency, bandwidth, etc). You could assume that the nodes will be based around the NXP JN516X radio module/system-on-chip. Consider your choice of an appropriate indoor wireless channel model, to model path loss and noise. Your chosen radio transmit power, channel model, hub location etc MUST ensure that multi-hop packet routing is required; it should not be possible for all nodes to be able to communicate directly (in a single-hop) with the hub. You may want to put your hub at one side of the network (not the centre) to help with this. Make sure you set sensible dimensions for your simulated floorplan, and locate devices in appropriate locations.

•    Consider how to model packet generation, and model/approximate the flow of data through the various types of node in the network. You can model light switch presses however you wish, in order to explore system operation. For example, do you think it is realistic that every light switch might be pressed once per second? Are they all pressed at the same time?

•    Consider how to model the energy properties  (e.g. the capacity of the energy stores, their initial values, and the radio power consumption in various receive/transmit/idle states) of the different devices. The light switches are battery powered, while the hub and smart lights are powered from mains. Your nodes should NOT harvest energy (i.e. you should not model any energy generation).

•    Consider how long your simulations should run for. You should simulate the network for  a  period  of  time  sufficient  to  explore  expected/typical  behaviour.  How  many simulation runs do you need in order to be able to draw general conclusions?

You are strongly encouraged to look at what you are required to write about in Section 1 of your  individual   report,  to  ensure  that  you  are  making  the  necessary  decisions  and appropriately recording your justification for them.

Part 2: Check your Model and Simulation Setup

To check that you have made appropriate modelling decisions, and setup the simulation correctly, you are encouraged to first check that the network performs as you would expect it to. Use OMNeT++ to simulate a small subset of the network; use a small number of devices and place them sufficiently close to the hub such that they can communicate directly over a single hop. Run some simulations, and confirm that it operates as you would expect (e.g. for latency, packets sent/received, energy).

Part 3. Analyse the Network Performance

Add all of the nodes back into the network (1 hub, 28 lights, 7 switches), located in appropriate places on your simulated floorplan. Your task is to analyse the performance of the full system you have modelled, in particular investigating the customer’s concerns, i.e. reliability, latency and energy consumption. This document does not provide a prescriptive or exhaustive set of questions for you to answer. However, you may wish to consider:

• Reliability: How often might a switch be pressed, but the light not respond? Is this acceptable? If a node fails, does the rest ofthe network recover and operate correctly?

• Latency: How long does it take for a light to turn on after pressing a switch? Is this an acceptable delay? What is the delay caused by; how might you improve it?

• Energy Consumption: How much energy do the nodes consume? How large would a  switch’s battery  need to be to last  a year;  is  this  realistic?  Consider  the  energy consumption of switches and lights, but ignore the energy consumption of the hub.

Interpret your results; if the reliability/latency/power differs at different nodes in the network, what are the causes of this? How can you evidence this? Explore the routing table formed by the network. What does it look like, and does it change during simulation?  Do the results depend on the type of node, or its distance/number of hops from the hub? Consider the different  results  you  could  obtain/plot  to  explore  behaviour   –  for  example,  for  energy consumption you could obtain data for Remaining Energy vs Time, or you could obtain results for the Remaining Energy at the end of the simulation. What is most appropriate?

Think about suitable statistical methods to use to analyse and evaluate your results. It is unlikely that the average (mean) alone will ‘tell the whole story’, and you may wish to consider additional appropriate metrics   (e.g.  range,  variance,  standard  deviation,  interquartile range). Note that analysis is more than just presenting data/results and describing them. You should discuss why the behaviour is seen, and what is happening to cause this. You may need to do some reading of the technical literature and/or further experiments to answer this.

If you are struggling to get results, or to make sense of the results you have obtained, consider reducing the size of the network (substantially fewer switch and light nodes, potentially distributed over a smaller area). If and when you get interpretable results from this, scale your network back up to the full size (1 hub, 28 lights, 7 switches).

Deliverables and Marking Scheme

There is a single deliverable, marked out of 40, contributing 40% of the credit for this module.

Individual Report [total of 40 marks]

(due by 4pm on Friday 12th December 2025, online hand-in)

You should produce a written report documenting your investigation. The report must be single-spaced, single-column, use 12pt Times New Roman, and contain the following sections:

Section 1: Network Model [maximum of 1000 words]

You should include a screenshot showing the network topology, i.e. where various device types (light switches, smart lights and the hub) were spatially located. State and justify the physical dimensions of your floorplan (as defined in your .ned file).

For each modelling decision that you made (i.e. the models and parameters that you chose) in Part A of the coursework, you should then 1) state what you chose, 2) justify why this was an appropriate choice, and 3) refer to the line(s) in your simulation’s .ini file (which you will include in Appendix 1) where this was implemented. As a minimum, you should ensure that you state and justify all of the following:

• Radio and wireless channel

o Radio transmit power, sensitivity, data rate, frequency, bandwidth

o Channel background noise, path loss model, and associated parameters

•    Power/energy

o Initial and nominal capacities of energy storage devices

o Power consumption models and parameters in various states

•    Packet flows

o How did you configure the devices to model packet flows through the network?

o Packet generation start times, send intervals, packet lengths etc

•    Simulation setup

o Simulation time/duration

o Number of simulation runs

Where appropriate, you should substantiate your modelling decisions using references to standards, datasheets, and the technical literature (e.g. conference and journal articles).

Section 2: Results and Analysis [maximum of 1500 words]

Present the results you obtained from simulating your full network in Part 3, and discuss the conclusions that you draw from them. Do not underestimate the importance of results (i.e. figures to show data/graphs/networks etc) 一 this is where you explore various aspects of your simulation, e.g. energy, latency, lost/received packets. Make sure that you provide results to evidence all of the conclusions that you make, and explain them in detail (i.e. don’t leave the reader(s) to draw their own conclusions from your results). Label all axes on graphs, and state  units  where  appropriate.  We  encourage  you  to  replot  data  obtained  from  your simulations (e.g. in Excel), rather than just pasting OMNeT++ screenshots in your report (as these can be difficult to read and omit important information).

Explore why the network performs as it does. For all of the results that you present, you should consider ‘Is this what I expect?’ and ‘Why is this the case?’. For example:

•    If only 50% of sent packets are received, why are some of them getting lost? If 100% of sent packets are received, what’s ensuring that none of them get lost? Do you have results  that  evidence  that  this  is  indeed  what  is  causing  this  behaviour?  Is  this supported by the technical literature?

•    If some nodes deplete their energy store more/faster than others, why is this? What influences how much energy they consume? Do you have further results that confirm that this is what is causing the behaviour? Is it supported by the technical literature?

Support your analysis and conclusions using the technical literature, where appropriate.

Appendix 1 - Word Count

Complete the following table, and include it in Appendix 1:

* The Word Limit/Count excludes any figures or tables

Appendix 2 - Your .ini file

You  should  include  the  contents  of  your .ini (probably  named omnetpp.ini)  file  in  this appendix. Use a non-proportional/monospacedfont (e.g Courier) for readability, and add line numbers (many text editors can do this for you), so that you can refer to them in Section 1 of your report.

References

Include references to any sources you have used in your report; you must use the IEEE referencing  style.  You  are  expected  to  engage  with  technical  literature  to  justify  your modelling choices (for example technical papers on path loss models in indoor environments, datasheets on devices modelled, etc), and to understand and interpret your results.

Design Archive

Please also submit your code in a ‘design archive’: a zip file containing your project folder but NOT the results (as this will likely be very large). While the design archive is not marked, it is a compulsory submission and maybe looked at by the examiners if they have any concerns about academic responsibility and conduct. Late penalties will apply from the  time  that BOTH the report and design archive are submitted.

The marking criteria, and examples of corresponding descriptors (these are not exhaustive; academic judgement will extrapolate these to assess submissions), are shown below.