NNRA-CAC: NARX Neural Network-based Rate Adjustment for Congestion Avoidance and Control in Wireless Sensor Networks

ABSTRACT

A wireless sensor network (WSN) is an application area that is valuable in various fields, such as healthcare monitoring, environmental monitoring, and so on. Application areas require WSNs with high throughput and low degree of packet loss. Due to congestion in the network, the throughput of the network is affected, which imposes the need for congestion control in the network. This article proposes a method, titled NARX Neural network-based Rate Adjustment (NNRA) for avoiding and controlling congestion in the network. Initially, congestion in the network is avoided by dropping packets and the NNRA is used to control congestion in the network when congestion is present. Performance analysis is carried out in terms of throughput, delay, size of the queue, packet loss, and the level of the congestion using two setups. The results of the proposed method are compared with the existing methods to prove the effectiveness of the proposed method. The proposed method attained a maximum throughput at a rate of 0.9585 and minimum values for delay, queue size, packet loss, and the congestion level.

KEYWORDS:

• NARX neural network
• LM algorithm
• rate adjustment
• congestion
• packet dropping