Anomaly Detection In IOT Networks

This project is supervised by prof. Sikora, Axel as a part of academics, University of Freiburg.

This repository deals with the analysis and implementation of Intrusion Detection in Industrial Internet Of Things(IIOT) network based on ML models and TinyML Inferences.

References

Dataset

“Sebastian Garcia, Agustin Parmisano, & Maria Jose Erquiaga. (2020). IoT-23: A labeled dataset with malicious and benign IoT network traffic (Version 1.0.0) [Data set]. Zenodo. http://doi.org/10.5281/zenodo.4743746”

Updates:

1. Added XGBoost classifier and converted it to ONNX model to make it deployable at microcontrollers such as STM32. This provided an accuracy of 99% for a considerably medium-sized subset of the dataset.
2. Quantized the model using ONNX dynamic quantize.
3. Updated inferencing code to check model output using ONNX runtime.

To do: further optimizations of the XGBoost model to reduce memory footprints.