Trace Generation support for attack detection
- Project
- 19045 STACK
- Type
- New system
- Description
In order to train machine learning algorithms for attack detection, quite some training data is needed. For low-power wireless mesh networks, such data traces are not available. We have extended the widely used Cooja simulator (that stems from RISE and is used in both academia and industry) has been extended with functionality to create trace files. We have both implemented this tool and made it available as open source.
- Contact
- Niclas Finne, Joakim Eriksson, Thiemo Voigt, George Suciu, Mari-Anais Sachian, JeongGil Ko, Hossein Keipour
- thiemo.voigt@ri.se
- Research area(s)
- Security, IoT Mesh Networks
- Technical features
Able to generate data traces for attack detection
- Integration constraints
The tool is an extension of the Cooja simulator that is part of the Contiki-NG software
- Targeted customer(s)
Developers (security and machine learning experts) that want to design intrusion detection / anomaly detection etc mechanisms for IoT mesh networks
- Conditions for reuse
Part of open software with generous license, can be used by anyone
- Confidentiality
- Public
- Publication date
- 01-01-2021
- Involved partners
- RISE - Research institutes of Sweden (SWE)
- BEIA Consult International (ROU)
- Yonsei University (KOR)