SCRS Conference Proceedings on Intelligent Systems
Effective Intrusion Detection in IoT Environment: Deep Learning Approach
Authors: Sualiha Jan, Faheem Masoodi and Alwi M Bamhdi
Publishing Date: 26-04-2022
ISBN: 978-93-91842-08-6
Abstract
IoT is among the most important technologies in recent times, which has evolved for the good of human beings. It reduces the human efforts in every domain possibly by connecting things, making people work and live smarter. The use of smart devices in a variety of applications has helped pave the way for pervasive computing, which offers huge human, economic, and other advantages. However, these advantages come with a number of drawbacks that must be addressed, one of which is security.The issue addressed in this research will be an effective Intrusion Detection System deployed in the Internet of Things environment. Though, various Intrusion detection systems are available, some are using learning methods. However, they face challenges such as a lack of relevant data, with some relying on the KDD dataset, which isn't designed specifically for IoT. This paper presents Deep Learning-based techniques, such as DNN and LSTM-RNN classifier, for detecting classes of assaults and anomalies in a simulated smart environment and classifying them as abnormal or normal using the BoT-IoT dataset. The proposal framework was created with the Google Colab platform. The experimental results for all the attack classes using DNN and LSTM-RNN classifiers have achieved 99.7% and 99.8% accuracies respectively.
Keywords
IoT, Google Colab platform, KDD dataset, DNN and LSTM-RNN classifiers.
Cite as
Sualiha Jan, Faheem Masoodi and Alwi M Bamhdi, "Effective Intrusion Detection in IoT Environment: Deep Learning Approach", In: Raju Pal and Praveen Kumar Shukla (eds), SCRS Conference Proceedings on Intelligent Systems, SCRS, India, 2022, pp. 495-502. https://doi.org/10.52458/978-93-91842-08-6-47