For the upcoming academic year I would like to propose the following topics/areas for obtaining a master’s degree:
- Optimizing IEEE 802.11 Networks with Machine Learning – There are at least two tools (ns3-gym, ns3-ai) which allow applying machine learning algorithms to ns-3. You can also look at a previous work I supervised (paper, GitHub repo), in which reinforcement learning was used to dynamically set 802.11’s contention window value. The exact topic in this area is still to be agreed upon.
- A Python-based Simulator of Wi-Fi/NR-U Channel Access Coexistence – Write a simulator of channel access for Wi-Fi and New Radio Unlicensed (NR-U) coexistence using the Python SimPy library. There are two implementations for these technologies working separately: DCF-SimPy and NR-U channel access in SimPy. You can either integrate this code or write the simulator from scratch and then conduct a performance analysis (similar to Section 6 of Downlink channel access performance of NR-U).
- Analyzing Gap-based Channel Access for NR-U in ns-3 – Extend the 5G LENA simulator (a module for ns-3) by adding gap based channel access for NR-U. The channel access types (gap and reservation signal) are explained, e.g., in Section III of Revisiting LAA or in Section 3 and Figure 2 of Downlink channel access performance of NR-U. Next, conduct a performance analysis in ns-3 (similar to Section 6 of Downlink…).
- A Monte Carlo Simulator for ns-3 – Write an extension/overlay/framework/library for ns-3 allowing to run Monte Carlo (MC) simulations of a network. An MC simulator iterates over rounds (stages); an example application in the networking domain would be as follows. Consider a node receiving a video stream. At the end of each stage it checks if the stream’s QoS parameters are met and makes a decision (say, continue or disconnect for a random number of stages). Dividing the time into stages (the MC component) enables game theoretic research, while ns-3 provides credible performance results. I wrote an MC simulator for ns-3 for this research paper and I can share the code, but it’s more of a hack than a proper implementation. The end goal of this thesis should be an app submitted to ns-3’s app store.
I put forward these topics for the consideration of students who plan to complete their master’s degrees by defending their diploma theses in 2022.
If you are interested in writing a thesis under my supervision, please send an e-mail to szott@agh.edu.pl with the subject “Praca magisterska 2022”. Please write a few words about yourself and your motivation. Relevant information (such as past grades from related courses) is appreciated. Due to the holiday season, please be patient in awaiting my reply in July/August.