The demand for timely information updates is growing with the proliferation of real-time applications such as remote monitoring, autonomous... Show moreThe demand for timely information updates is growing with the proliferation of real-time applications such as remote monitoring, autonomous vehicles, and Internet of Things (IoT) networks. The age of Information (AoI), a critical performance metric, quantifies the freshness of received information. This dissertation addresses the challenge of minimizing AoI in practical wireless networks employing carrier sense multiple access (CSMA). We propose a comprehensive framework combining analytical modeling and deep learning techniques to optimize AoI under diverse network conditions.First, we introduce a deep learning-facilitated framework that enables a tagged node to adaptively adjust its update rate based on local observations and background traffic, minimizing its AoI. Extensive simulations in IEEE 802.11 networks validate this frameworkâs effectiveness. We then extend the framework to IEEE 802.15.4 networks, incorporating key protocol characteristics such as clear channel assessments and retransmissions. Simulation results demonstrate its generalization capability and significant AoI improvement.Next, we develop a novel stochastic hybrid systems (SHS)-based analytical model to evaluate AoI in finite-buffer CSMA networks. By incorporating collision probabilities and leveraging deep learning to handle heterogeneous background traffic, this model achieves accurate AoI analysis. Numerical results from ns-3 simulations confirm its robustness and scalability across various scenarios. Furthermore, we explore the interplay between AoI and other metrics, introducing ``throughput weighted age of information (TwAoI)" to evaluate the joint performance of information freshness and channel utilization, and investigating the trade-off between information freshness and sampling cost.This research bridges theoretical insights and practical implementation, providing a suite of tools for AoI optimization in CSMA networks. The findings serve as a guideline for system designers and network administrators aiming to support real-time applications in dynamic and distributed wireless environments. Show less
