This website hosts materials related to the collaborative NSF-sponsored project entitled "Fundamental Limits on Information Freshness" with award numbers 1836690 and 1836695.


PIs: Andrew G. Klein (WWU) and D. Richard Brown III (WPI)

WWU undergraduate student researchers: Arick Grootveld, Leah Lackey, Vlad Bugayev

WPI graduate student researchers and post-docs: Shahab Farazi


Information freshness is of critical importance in a variety of networked monitoring and control systems such as Internet of Things (IoT), intelligent vehicular systems, feedback in wireless communication systems, environmental monitoring, robotic networks, and real-time monitoring and control in cyber-physical systems. In addition it is important in several information-update and data analytics applications including financial trading, social networks, crowdsourcing, consensus systems, and online learning. In all these applications, stale information can lead to incorrect decisions, unstable control loops, and even compromises in safety and security. Since timely updates are critical in these and many other contemporary applications, an important problem is to understand the fundamental limits of information freshness in networked status update systems. This project will further our understanding of the fundamental limits of information freshness in general multi-source/multi-monitor settings, and will develop strategies to achieve these limits. The project also includes engaging research experiences for both undergraduate and graduate students.

The approach in this project differs from prior work in age of information (AoI) by considering general multi-source/multi-monitor networks with explicit contention and using techniques from graph theory to study information freshness under the AoI metric in general network topologies with interference constraints. This approach is expected to lead to new insights in realistic networks with general topologies. The research plan is organized into three related tasks to better understand the theoretical foundations of AoI in the explicit contention framework: (i) simultaneous transmission and network coding, (ii) generalized information mappings and age weightings, and (iii) large-network limits. The focus in all of these tasks is on characterizing fundamental limits of information freshness and developing strategies to approach or achieve these limits in general classes of wireless and wired networks. The proposed work will generalize the explicit contention framework for AoI to a rich, realistic class of problem settings including settings where nodes can simultaneously broadcast information updates, settings where nodes have access to common or correlated information, and settings where certain information is more critical than other information.


Intellectual merit activites: This project has explored various aspects of multi-source status update systems, with special attention devoted to the impact of unreliable communications due to packet errors. We developed an algorithm that generates schedules for any network with a connected topology. We also derived a closed-form expression for the average number of time slots to update all tables as a function of fundamental graph properties. Under the same theme of multi-source age of information with unreliable packets, we investigated a more classical queueing theoretic age of information setting with multiple sources communicating over a single hop to a single monitor. Specifically, we considered a setting first studied by Yates where any source can preempt the packets of any other source currently in service. Since this sort of preemption may not be desirable in some settings, we studied the case where newly generated packets can preempt only the packets in service from the same source. We derived a closed-form expression for the average age of information in this setting in terms of the system parameters, e.g., packet generation rates of each source, service rate of the server, and packet loss probability. Our work considers both stationary randomized policies as well as round-robin policies under three different packet retransmission policies when errors occur, including: (i) retransmission without resampling, (ii) retransmission with resampling, and (iii) no retransmission. Expressions for the average AoI have been derived for each of these cases, and we have shown that the round-robin schedule policy in conjunction with retransmission with resampling when errors occur achieves the lowest average age of information among the considered cases. For stationary randomized schedules with equiprobable source selection, we have shown that the average AoI gap to round-robin schedules with the same packet management policy scales as O(N). And, for stationary randomized policies, the optimal source selection probabilities that minimize a weighted sum average AoI metric was derived.

Broader impact activities The investigations conducted under this award have led to a better understanding of the fundamental limits of information freshness in broad classes of wireless networks and led to the development of strategies to achieve or approach these limits. Since information freshness is important in a wide range of applications ranging from environmental monitoring to intelligent vehicular systems to social networks, the foundational studies conducted under this award have the potential for significant technological and societal impact. In terms of educational impacts, this project has supported the training of a PhD student as well as three undergraduate researchers through the REU program.