Information Freshness for Monitoring and Control over Wireless Networks

Abstract

In this thesis, we study the optimization of general information freshness metrics in wireless networks, with the goal of applying our theoretical results to problems in real-time monitoring and control. We make contributions in three directions.

First, we consider the optimization of general cost functions of Age of Information (AoI). Here, we develop computationally efficient scheduling algorithms for optimizing information freshness in both single-hop and multi-hop wireless networks. We further develop an online learning formulation when the cost functions of AoI are unknown and propose a new online learning algorithm for this setting called Follow-the-Perturbed-Whittle-Index.

Second, we consider weighted-sum AoI minimization. In this setting, we study how correlation impacts information freshness. We also propose a near-optimal distributed scheduling protocol called Fresh-CSMA for AoI minimization, that has provable performance guarantees.

Third, we apply our theoretical results to problems in multi-agent robotics and monitoring – both via simulations and practical system implementations. We use simulations to demonstrate significant performance improvements in the collection of time-varying occupancy grid maps using multiple robots via the Whittle Index framework. Further, to demonstrate the benefits of our theoretical contributions, we built a real system (WiSwarm) for mobility tracking using a swarm of UAVs, communicating with a central controller over WiFi. Our experimental results show that, when compared to the standard IEEE 802.11 MAC layer + TCP/UDP, our system can reduce AoI by a factor of 109x/48x and improve tracking accuracy by a factor of 4x/6x, respectively.