Detecting Anomaly in IoT Devices using Multi-Threaded Autonomous Anomaly Detection
Muhammad Yunus Iqbal Basheer, Azliza Mohd Ali, Nurzeatul Hamimah Abdul Hamid, Muhammad Azizi Mohd Ariffin, Rozianawaty Osman, Sharifalillah Nordin · 2021
In this paper, we proposed a multi-threaded anomaly detection algorithm that works autonomously and without any prior assumption in streaming data. The proposed algorithm was upgraded from the autonomous anomaly detection (AAD) algorithm. The difference is we change the AAD algorithm from working offline to online by using IoT devices. Furthermore, to make it fast, we change the algorithm to process asynchronously. Before this, the AAD algorithm works in offline and worked in synchronous processing. Although AAD was altered to receive online data from IoT devices, the algorithm performance is still slow. Hence, this paper aims to prove that multi-threaded or asynchronous processing is much more efficient and faster in handling dynamic streaming data from IoT devices than single-threaded AAD.