Congestion Tracking Control for Wireless TCP/AQM Network Based on Adaptive Integral Backstepping
- PDF / 544,602 Bytes
- 8 Pages / 594.77 x 793.026 pts Page_size
- 82 Downloads / 200 Views
ISSN:1598-6446 eISSN:2005-4092 http://www.springer.com/12555
Congestion Tracking Control for Wireless TCP /AQM Network Based on Adaptive Integral Backstepping Lujuan Ma, Xiaoping Liu*, Huanqing Wang, and Yucheng Zhou Abstract: This paper extends the backstepping control strategy to wireless Transmission Control Protocol (TCP ) network to solve the Active Queue Management (AQM) problem. Different from the wired network, the uplink and downlink of the wireless network are asymmetric and the packet loss may be caused by congestion and may occurs during the transmission over wireless links. Inspired by the existing backstepping design idea for a wired network model, an adaptive tracking controller is proposed to deal with the congestion problem in wireless network. The uplink and downlink packet losses are estimated by adaptive update laws. The presented method provides satisfactory tracking performance in the network. Meanwhile, the packet loss is small. The simulation results show the feasibility of the proposed method. Keywords: Active queue management, backstepping, congestion control, wireless network.
1.
INTRODUCTION
With the rapid development of streaming media, network congestion control has become a hot issue. It is well known that Sack [1], New Reno [2], and Vegas [3] are typical congestion control mechanisms. This class of methods is conducted by terminal devices and has inherent problems of global synchronization and deadlock [4]. As a result, AQM algorithms are recommended by Internet Engineering Task Force (IEFT) [5], which are performed by routers. Random Early Detection (RED) [6] is the representative of AQM. Afterwards, a series of followers are presented [7–11]. RED algorithms have good performance in stability and fairness, but they are too sensitive to parameters [12]. After the fluid model [13] is proposed, control methods can be applied to congestion control [14–16]. In order to deal with the changing of the network, many advanced control design methods are adopted. A robust control algorithm is proposed in [17], targeting to solve problems caused by link capacity disturbance and model uncertainty. Dan et al. [18] studies the influence on congestion control from network topology uncertainty. Zhang et al. [19] aims to deal with the problem of adaptive neural network control by adopting one-to-one nonlinear mapping. Mohammadi et al. [20] combines fuzzy control [21] and PID to address the problem of network congestion. A
model predictive procedure is proposed to cope with the nonlinear variation of the network state and delay [22]. Optimization algorithms are also applied to solve the congestion problem [23–25]. Backstepping is a typical control design method, which has been applied for AQM in recent years [26–31]. Backstepping sliding mode control is used in [32] and [33]. Time-varying round trip delay is considered by applying the backstepping technique in [34–37]. Many researches combine backstepping with other control methods, such as minimax [38], H∞ [39], Prescribed Performance Control (PP
Data Loading...
