Congestion Avoidance Aware using Modified Weighted Fairness Guaranteed DRED-FDNNPID Congestion Control for MWSN

Mostly, the critical challenge is Wireless Sensor Network (WSN) is congestion control and its avoidance when traffic becomes higher than channel capacity. This can cause high packet loss ratio and less efficiency by degrading the overall network performance. To control the congestion, certain considerations are needed to detect, avoid and resolve congestion through the network. As a result, Dynamic Random Early Detection with Fuzzy-Deep Neural Network Proportional Integral Derivative (DRED-FDNNPID) controller based congestion control technique was proposed. However, it requires network traffic estimation at congested nodes to divert the traffic or drop the packets via an alternate path in the network. Hence in this article, a network traffic estimation and diversion method is proposed to avoid the congestion. In this technique, a modified additive increase decrease algorithm is proposed to control the congestion rapidly than the conventional additive increase and multiplicative decrease algorithm. This reduces the frequent data packet loss and the traffic is adjusted based on the queue size and capacity of number of packets through the alternate paths. Finally, the simulation results show that the effective improvements on proposed technique compared to the existing technique.