A cognitive QoS management framework for WLANs

Abstract

Due to the precipitous growth of wireless networks and the paucity of spectrum, more interference is imposed to the wireless terminals which constraints their performance. In order to preserve such performance degradation, this paper proposes a framework which uses cognitive radio techniques for quality of service (QoS) management of wireless local area networks (LANs). The framework incorporates radio environment maps as input to a cognitive decision engine that steers the network to optimize its QoS parameters such as throughput. A novel experimentally verified heuristic physical model is developed to predict and optimize the throughput of wireless terminals. The framework was applied to realistic stationary and time-variant interference scenarios where an average throughput gain of 344% was achieved in the stationary interference scenario and 70% to 183% was gained in the time-variant interference scenario.

Throughput optimization of wireless LANs by surrogate model based cognitive decision making

“…In this condition, the dominant external parameters influencing the throughput of the SUT is the traffic characterization of the homogeneous collocated interference links. As authors in [7] suggest, the most crucial parameters of the interference, namely channel occupancy degree (COD) and transmission rate of the interference could be practically measured by means of packet sniffers and looking up the radio tap header information of sniffed packets. Two dedicated monitoring agents are incorporated that are basically Wi-Fi interfaces operating on IEEE 802.11 monitor mode [19] on every of the channels in the vicinity of the SUT (see Fig.…”

“…The monitors are assumed to capture all present traffic without dropping any packets. If for any reason the interference is not perceived by the sniffers, an event detector as proposed in [7] could assist by detecting the unpredicted throughput degradation in the network. Secondly, deploying both the CDE and the REM imposes data and processing overhead on the network.…”

“…rule-based [2], case-based [3]- [5], search-based [6], knowledge-based [4] reasoning systems and many others. Such methods often require a certain set of representative cases or domain expert knowledge [7] to derive important analytical formulas or rules that steer the decision making process. The reader is referred to [1] for a detailed survey.…”

Surrogate modeling based cognitive decision engine for optimization of WLAN performance

“…rule-based [2], case-based [3, 4, 5], search-based [6], knowledge-based [4] reasoning systems and many others. Such methods often require a certain set of representative cases or domain expert knowledge [7] to derive important analytical formulas or rules that steer the decision making process. The reader is referred to [1] for a detailed survey.…”

Throughput optimization strategies for large-scale wireless LANs

“…In [4] the authors propose a framework, which utilizes radio environment maps to make informed decisions on the operating channel of the wireless network in response to performance degradation of the network due to the interference. However, the framework was verified in a controlled pseudo-shielded environment with a limited number of devices.…”