skip to content

Department of Computer Science and Technology

  • Senior Research Associate

Project Suggestions:

I am happy to supervise Part II, Part II, or MPhil ACS projects in my research area, which includes Internet of Things (IoT), privacy preservation, edge computing, or Wireless Sensor Networks. Possible projects might align with, but are not limited with the following topics:

IoT Device Identification

Propose a new way how an IoT device can be identified using its network traffic, or re-evaluate existing approaches using publicly available datasets as well as a dataset collected by our lab.

IoT Model Retraining at the Edge

It has been shown that the accuracy of NN models used to identify IoT devices degrades over time. Therefore, there is a need to update the models by retraining them. The open question remains, what is the best way to retrain the models: a) in a cloud using a single global model, b) using federated learning, or c) refine the models at the edge.

Rust at the Edge

Using the Rust programming language to compile a simple program with well-defined input and output for different edge platforms, depending on their capabilities and CPU architecture.

Research

His research focuses encompasses the wide are of the Internet of Things. This includes small embedded systems with low-powered wireless radio ensembled in Wireless Sensor Networks as well as home IoT devices including smart speakers, smart TVs, or doorbells. The focus of his research is the identification of these IoT devices using modern machine learning and artificial intelligence techniques such as random forests, Bayesian networks, or neural networks. Roman is also interested in the area of users' privacy preservation in the context of IoT. Finally, his research is also oriented on the area of security and protection of IoT devices deployed in our homes.

Currently, he participates in two research projects:

Publications

Reports

  • Kolcun, R. and Mortier, R., 2023. Scalable Device Identification for IoT Networks using Binary Classification Models at the Edge
  • Journal articles

  • Kolcun, R., Popescu, DA., Safronov, V., Yadav, P., Mandalari, AM., Xie, Y., Mortier, R. and Haddadi, H., 2020. The Case for Retraining of ML Models for IoT Device Identification at the Edge
  • Conference proceedings

  • Popescu, DA., Safronov, V., Yadav, P., Kolcun, R., Mandalari, AM., Haddadi, H., McAuley, D. and Mortier, R., 2019. "Sensing" the IoT network: Ethical capture of domestic IoT network traffic: poster abstract. SenSys '19,
    Doi: 10.1145/3356250.3361953
  • Popescu, DA., Safronov, V., Yadav, P., Kolcun, R., Mandalari, AM., Haddadi, H., McAuley, D. and Mortier, R., 2019. "Sensing" the IoT network: Ethical capture of domestic IoT network traffic: poster abstract. SenSys '19,
    Doi: 10.1145/3356250.3361953
  • Kolcun, R., Popescu, DA., Safronov, V., Yadav, P., Mandalari, AM., Mortier, R. and Haddadi, H., Revisiting IoT Device Identification
  • Pan, W., Kolcun, R. and Mortier, R., Anonymising Video Data Collection at the Edge Using DeepDish 2023 IEEE 24th International Conference on High Performance Switching and Routing (HPSR),
    Doi: 10.1109/HPSR57248.2023.10147953
  • Software

  • Kolcun, R., IoT Device Identification at the Edge
  • Contact Details

    Room: 
    FN07
    Office phone: 
    (01223) 3-34647
    Email: 

    roman.kolcun@cl.cam.ac.uk