POST-DOCTORAL POSITION « GOAL-ORIENTED DATA SAMPLING, SCHEDULING AND MULTIPLE ACCESS FOR REAL-TIME MONITORING «

Subject: Goal-oriented data sampling, scheduling and multiple access for real-time monitoring Supervisor : Quang Trung LUU Over the past decade, communication systems have undergone explosive growth in both data traffic and global infrastructure. Internet of Things (IoT) systems are rapidly gaining traction in domains ranging from environmental monitoring and autonomous driving to smart manufacturing and energy management. As interconnected devices equipped with sophisticated sensing, computing, and learning capabilities continue to proliferate, they are expected to generate unprecedented volumes of data. This trend will place significant pressure on existing communication networks. In particular, scenarios such as smart factories and autonomous driving require continuous monitoring of physical processes to support timely decision-making and control. In classical data communication systems, functions such as data sampling, transmission, reception, and protocol design have been treated as separate modules, without considering the underlying purpose of the communication. While this modular approach has been effective for conventional applications, it becomes inefficient in large-scale real-time systems. This approach may lead to excessive data generation, resulting in network congestion and limiting the scalability of future connected systems. Recently, goal-oriented performance metrics, such as the Age of Information (AoI) and its variants, have been proposed to better capture the timeliness and usefulness of transmitted information [1], [2]. Incorporating such metrics into communication system design, particularly in data sampling policies, can significantly reduce unnecessary data transmissions [3]. Although recent studies have examined the role of AoI in real-time systems, most have concentrated on specific mechanisms, such as sampling policies and resource allocation schemes, often assuming either a single information source or a limited number of independent sources [3], [4]. In this project, we aim to extend existing work in the following directions:
• Propose new goal-oriented metrics that go beyond AoI to capture both the purpose and the characteristics of data in emerging applications (e.g., by considering the value of information alongside its freshness). These metrics will guide the design of more efficient communication and networking strategies. We will identify concrete use cases in smart factory and autonomous driving scenarios and evaluate the relevance of the proposed metrics in those contexts;
• Develop joint goal-oriented sampling, scheduling and multiple access schemes to reduce data overhead and increase the number of supported devices by integrating data generation and sampling processes directly into the communication framework. Unlike prior work, which has largely focused on efficient resource utilization through AoI-based metrics, we will address joint sampling, scheduling and multiple access using richer goal-oriented metrics that account for correlated sources—since multiple sources may serve the same application. Incorporating application-specific notions such as the value of information introduces new analytical challenges. The resulting problem involves the joint optimization of three interdependent stochastic processes: (i) the evolution of the information source, (ii) the time-varying value of its data, and (iii) the randomness of the communication channel. One promising direction to address these complexities is to formulate the problem as a Markov decision process (MDP) and develop low-complexity solutions, for instance, using Whittle index policy [5]. References [1] S. Kaul, R. Yates and M. Gruteser, « Real-time status: How often should one update?, » in Proc. IEEE INFOCOM, Orlando, FL, USA, 2012, pp. 2731-2735. [2] R. D. Yates, Y. Sun, D. R. Brown, S. K. Kaul, E. Modiano and S. Ulukus, « Age of Information: An Introduction and Survey, » in IEEE Journal on Selected Areas in Communications, vol. 39, no. 5, pp. 1183-1210, May 2021.. [3] Y. Sun, E. Uysal-Biyikoglu, R. D. Yates, C. E. Koksal and N. B. Shroff, « Update or Wait: How to Keep Your Data Fresh, » in IEEE Transactions on Information Theory, vol. 63, no. 11, pp. 7492-7508, Nov. 2017. [4] I. Kadota, A. Sinha, E. Uysal-Biyikoglu, R. Singh and E. Modiano, « Scheduling Policies for Minimizing Age of Information in Broadcast Wireless Networks, » in IEEE/ACM Transactions on Networking, vol. 26, no. 6, pp. 2637-2650, Dec. 2018. [5] P. Whittle, “Restless bandits: Activity allocation in a changing world,” Journal of Applied Probability, vol. 25, no. A, pp. 287–298, 1988.