Nonmonotonic-Based Congestion Control Schemes for a Delayed Nonlinear Network

摘要

In this paper, buffer dynamic modeling for wireless sensor networks as a highly nonlinear system is accomplished in discrete time, different subsystems are achieved based on delay, and the overall model is gained by blending them. According to nonlinear dynamics point of view, considering delay in the analysis of congestion control schemes is of paramount importance. In this paper, an adaptive back-off interval selection works with the proposed robust controller. Based on queue utilization and channel estimation algorithm, congestion is detected and a suitable rate is selected by adaptive back-off interval selection. An augmented form of our proposed system is utilized for controller synthesis. A new approach is proposed for controller synthesis based on non-quadratic and common quadratic Lyapunov candidates where the former is generalized to be more relaxed. Also, the monotonicity requirement of Lyapunov's theorem is relaxed. The closed-loop systems are globally asymptotically stable in case of delay changes resulted from queue size changes. Extended simulation results confirm the effectiveness of our proposed schemes.