Packet Loss Characterization Using Cross Layer Information and HMM for Wi-Fi Networks.

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Packet loss significantly degrades wireless network performance and impacts internet service quality.
• Existing packet loss models, like the Gilbert-Elliot (GE) model, often fail to accurately represent real-world Wi-Fi network behavior.
• Current models typically focus on a single network layer, neglecting crucial network state variables.

Purpose of the Study:

• To develop a novel packet loss model for Wi-Fi networks that enhances prediction accuracy.
• To incorporate both the temporal dynamics of losses and key network state variables into a unified model.
• To address the limitations of existing models by considering signal-to-noise ratio and network occupation simultaneously.

Main Methods:

• Utilized a Hidden Markov Model (HMM) framework for its training and forecasting capabilities.
• Characterized packet loss burst-lengths at each HMM state using probability distributions.
• Incorporated signal-to-noise ratio and network occupation as primary input variables.
• Validated the model using computer simulations and real-world network data, analyzing burst-length distributions and Mean Square Error (MSE).

Main Results:

• The proposed HMM-based packet loss model demonstrated superior performance compared to existing models for Wi-Fi networks.
• The model accurately captures the temporal behavior of packet losses and network state variables.
• Validation using real network data confirmed the model's effectiveness and outperformance over competitors.

Conclusions:

• The novel packet loss model offers a more accurate representation of Wi-Fi network behavior than traditional models.
• Simultaneous consideration of temporal dynamics, signal-to-noise ratio, and network occupation is crucial for effective packet loss modeling.
• This research provides a valuable tool for improving the reliability and performance of wireless communication systems.