A Discrete-time Sliding Mode Congestion Controller for Wireless Sensor Networks

Abstract Congestion is a challenging problem for Wireless Sensor Networks (WSNs) because it results in energy waste, packet loss, low throughput and short lifecycle. By regulating incoming and outgoing packets at a particular node, this paper first proposes a new traffic-based discrete congestion control model in WSNs. Then, based on this model, an exponential-reaching-law-based discrete-time sliding mode congestion controller (DSMC) is designed, which effectively adjusts bottleneck nodes’ queue length to desired value. Extensive NS-2.35 simulation results demonstrate that the proposed DSMC outperforms conventional control approaches (such as Fuzzy control, PID control and Fuzzy-PID), it significantly controls queue length and avoids congestion, and has better performance such as short delay, low packet loss rate, high throughput, long network lifetime, fast convergence and strong stability.