kumoh national institute of technology
Networked Systems Lab.


NSL > Education> Correction
Deep Reinforcement Learning with Edge Computing in Faulty-Node Detection for Industrial Internet of Things
By : Alifia Putri Anantha
Date : 2019-10-04
Views : 208

In industrial internet of things (IIoT), wireless sensor network (WSN) is one of essential elements that need to be considered. However, WSN deployed for IIoT is prone to system failures due to the harsh conditions of IIoT. Concurrently, most of the IIoT applications have real-time requirements. Thus, based on the aforementioned issues, this paper proposes a novel faulty node detection scheme for IIoT by leveraging edge computing and deep reinforcement learning (DRL). In the proposed scheme, edge computing is used to perform a DRL task, and as a result, it can reduce the processing time and network latency. Moreover, DRL brings the detection accuracy to be more precise. Therefore, the proposed scheme is able to detect the node failure in real-time, and the catastrophic condition of the system can be avoided. Results of the simulation indicate that the technique proposed can enhance the reliability of the identification of faulty nodes.
Attachment 1:   ÷ 108._赿.docx(51.3KB)