An energy efficient selection of cluster head and disease prediction in IoT based smart agriculture using a hybrid artificial neural network model

Abstract

The advent of new technologies paved the way for the proliferating productivity of agriculture and farming activities in a cost effective way. Internet of Things is the one among them and used for automatic smart farming applications. Hence to overcome the issues faced in the wide agriculture sector most of the farmers are now changed to IoT based applications and in the meantime many scientists also used artificial intelligence techniques to predict the diseases of the crops from the data collected using the IoT sensors in the agriculture farming. The sensors gather data from crops and fused with the local unit and forwarded to the cluster heads of network. To achieve the disease free crops and therein improve the productivity we have proposed a novel energy efficient IoT based smart farming approach, in which we have utilized K-means algorithm for the cluster formation and Adaptive Mud Ring optimization algorithm (AMR) for CH and energy efficient optimal path. The collected data via the CH are then stored in the cloud storage. Subsequently, the data are accessed via the proposed Hybrid Artificial Neural Network (HANN) to predict the diseases. The Artificial Neural Network (ANN) is hybridized using the Google Net in order to extract the exact features to predict the diseases. Performance validation is effectuated with Network Simulator-2 (NS-2) software and the results are compared with the state-of-art works. Several parameters such as energy efficiency, delay, network lifetime, accuracy and etc, are used for analyzing and our approach surpasses all the other approaches.

Keywords

• Internet-of-Things
• Smart agriculture
• Energy efficient routing
• Artificial neural network and adaptive mud ring optimization algorithm