Optimized Support Vector Machine Based Congestion Control in Wireless Sensor Network Based Internet of Things

As the Wireless sensor network (WSN) has significant part in Internet of Things (IoT), it is utilized in various applications such as sensing environment and transmitting data via the internet. Nevertheless, due to the problem of heavy congestion, WSN based IoT obtains longer delay, higher ratio of packet loss and lower throughput. Although machine learning algorithms have been presented by researchers for detecting the congested data in IoT, detection accuracy is further to be improved. So, to control the congestion in WSN based IoT, artificial flora algorithm (AF) based support vector machine (SVM) is presented in this paper. To improve the performance of SVM, penalty parameter and kernel parameter of SVM is optimized using AF algorithm. In this proposed SVM-AF, the performance factors are given as input such as queue size (que), packet loss (pkt loss), cwnd (congestion window size), and throughput (throu). Based on these input factors, the prediction model SVM-AF predicts the congested data and decides whether to offload each device task to the server. Simulation outcomes show that the proposed SVM-AF outperforms the model such as Genetic Algorithm based SVM (SVM-GA) and SVM based on throughput, energy consumption, delivery ratio, and overhead.