Agent based Load Optimization in Sensor Cloud

Abstract

Sensor Cloud is a prominent technology that integrates sensor network and cloud computing to provide integrated capabilities to a wide range of applications. Sensor network sense physical parameters in the sensor cloud, captures them and, eventually, the information is placed in cloud servers via gateway. At Every occasion untreated information from the sensor network is placed in the cloud server resulting in the overall system being unbalanced. Because of the handling of raw information, the energy of the node eventually drain as well as the cloud server stores most of the undesirable information which would degrade its efficiency. Load balancing in sensor cloud using Neuro-fuzzy and agents is proposed in this paper. By utilizing Neuro-fuzzy method, unwanted information is eliminated and only expected information with better accuracy is stored into cloud server. At the physical sensor network, agents have been continuously imposed to retrieve the sensed information from various clustered nodes and finally it is been submitted to cluster head minimizing usage of node energy. The combinational Neurofuzzy optimization technique balances the overall load of the system by rejecting similar information generated from multiple sensors and making decisions on the information gathered. Accuracy of the Neuro-fuzzy output is improved by incorporating weights to the decision outcome which further improves system reliability significantly. The result explores that there is an inconsistent increase in network life of overall system as well as improved information accuracy and small information is processed and saved in cloud servers with high accuracy, hence enhancing system stability by using load balancing method.