Joint adaptation framework in mobile ad hoc networks: A control theory perspective

Abstract

To enhance the performance of wireless communications in mobile ad hoc networks, existing methods focus on tuning one certain wireless variable such as rate adaptation, or two variables together such as joint power-rate adaptation. However, field tests reveal that not only the single controllable variable but also their correlation affect the performance. Tuning them one-by-one and ignoring their correlation cannot achieve the optima. In this paper, we study the adaptation problem from a distributed control perspective and present a general joint adaptation framework (JAF). Leveraging the multiple-input-multiple-out control model, JAF is scalable, which embraces all controllable variables as its inputs and target performance metrics as its outputs. Moreover, based on the closed-loop control theory, JAF adapts the optimal combination of variables through the feedback of the real-time measurements. Extensive simulations are conducted to evaluate the distributed JAF. As an example, every node using JAF jointly adapts its data rate and transmission power in our simulations. The simulation results demonstrate that JAF outperforms the existing methods by improving the throughput up to 13% and the packet delivery ratio up to 15% simultaneously.