Congestion-aware Data Acquisition with Q-learning for Wireless Sensor Networks

In wireless sensor networks (WSNs), congestion is one of the challenging issues. It degrades the network's performance in various operating variables such as throughput, latency, energy consumption, packet loss, lifetime of the network and etc. Congestion occurs in WSN when the rate of sensor data flow outreach the channel or buffer capacity. The congestion is controlled in two ways in WSN, such as controlling network traffic or efficiently managing the resources. This paper performs the resource control mechanism by providing the alternative path towards to base station using a Q-learning for congestion alleviation. This congestion-aware data acquisition (CADA) mechanism initially identifies the congestion node (CN) where the nodes buffer occupancy ratio is higher. Further, we recognize the proper next node to construct the dynamic alternative route to the base station. The CADA is evaluated in various network conditions by comparing it with recent congestion-aware algorithms. The simulation tests show that the CADA efficiently ameliorates the congestion and enhances the performance across multiple performance metrics.