Quality of service evaluation and assessment methods in wireless networks

Wireless networks are capable of facilitating a reliable multimedia communication. The ease they can be deployed is ideal for disaster management. The Quality of Service (QoS) for these networks is critical to their effectiveness. Evaluation of QoS in wireless networks provides information that supports their management. QoS evaluation can be performed in multiple ways and indicates how well applications are delivered. In this work, fuzzy c-means clustering (FCM) and Kohonen unsupervised neural networks were compared for their abilities to differentiate between Good, Average and Poor QoS for voice over IP (VoIP) traffic. Fuzzy inference system (FIS), linear regression and multilayer perceptron (MLP) were evaluated to quantify QoS for VoIP. FCM and Kohonen successfully classified VoIP traffic into three types representing Low, Medium, and High QoS. FIS, regression model and MLP combined the QoS parameters (i.e. delay, jitter, and percentage packet loss ratio) with information from the generated clusters and indicated the overall QoS.