Energy-efficient QoS-based Congestion Control for Reliable Communications in Wireless Multimedia Sensor Networks

Congestion is a major cause of energy wastage and Quality of Service (QoS) degradation in wireless communication systems. Saving energy and maintaining high QoS levels are especially important in Wireless Multimedia Sensor Networks (WMSN), due to the limited sensor node energy resources and QoS-related application requirements. This paper proposes an energy efficient and QoS-aware congestion control scheme for reliable communications over WMSNs (eqCC). The proposed solution makes use of QoS feedback and current battery energy levels of sensor nodes in order to adapt sending data rate. We employ reinforcement learning by formulating the problem in terms of a Markov Decision Process (MDP) and solve it using the Q-Learning technique. The proposed eqCC is validated using simulations and is compared with classic TCP and Flush, another congestion control algorithm for Wireless Sensor Networks (WSN). The results show how eqCC outperforms the other solutions under high and low network load.