Q-Learning Based Emergency Event Detection in Secure Internet of Things

Energy efficiency and network lifetime maximization-based event detection are still major challenges in wireless sensor networks-Internet of Things (IoT). However, current efforts have significant issues, including a hierarchical architecture that increases complexity, time, and energy use. This research presents a secure IoT paradigm with the help of reinforcement learning to address these issues. First, the network is clustered and data aggregation is performed in the monitoring area. The data may have or not the events that needed to be detected. In each cluster, CH collects the data. Then, the data is hashed for security purpose by using PHOTON which is the lightweight hashing algorithm. With the secure data, BS is responsible to detect events by executing Q-Learning algorithm. The overall work is evaluated in NS-3 based on multiple performance metrics such as delay and throughput.