Congestion Control Mechanism in Wireless Sensor Network - QoS Evaluation and Performance Analysis

Wireless Sensor networks (WSN) has emerged as the effective tool for gathering the information from the remote environment. The sensor node communicates with the base station simultaneously when the data is available. This results in the packet loss at the receiving node. Since the nodes operated in the WSN are battery operated, retransmission of the data packets due to the congestion results in the failure of the system. Thus avoiding the congestion in the nodes has emerged as...
More Description

Wireless Sensor networks (WSN) has emerged as the effective tool for gathering the information from the remote environment. The sensor node communicates with the base station simultaneously when the data is available. This results in the packet loss at the receiving node. Since the nodes operated in the WSN are battery operated, retransmission of the data packets due to the congestion results in the failure of the system. Thus avoiding the congestion in the nodes has emerged as important research topic. In this book work utilizes a congestion threshold for determining the congested node in the WSN. Since the data packets have its own priority, the low priority data packets are dropped during the transmission to reduce the congestion. The presence of the congestion in the node can be lowered by making the share rate of the child nodes less than the service rate of the parent node. In this thesis, the congestion avoidance is done by choosing the optimal share rate with three proposed works. The proposed work included in this thesis is ACSRO, EACSRO, and NNRA algorithm. The proposed ACSRO algorithm modifies the existing CS algorithm and uses a proposed fitness function for finding the optimal share rate. The proposed ACSRO algorithm uses the adaptive step size for modifying the existing CS algorithm. The proposed EACSRO algorithm modifies the ACSRO algorithm with the use of the epsilon constraint. The proposed NNRA algorithm uses the NARX based neural network of the optimization process.