Title:
Resilient Networking Strategies for Manned-Unmanned Teaming Under Adversarial Conditions
Abstract:
Manned–unmanned teaming networks integrate human-operated platforms with unmanned systems, such as UAVs, to enhance situational awareness, mobility, and mission effectiveness in complex environments. The success of these teams depends critically on reliable wireless communication for coordinating task assignments and data exchange. However, the control and data links that sustain this coordination are vulnerable to disruption from congestion, mobility-induced intermittency, and adversarial interference such as smart jamming. These challenges motivate the development of resilient scheduling, routing, and decision-making frameworks that adapt to uncertain and dynamic network conditions.
This dissertation investigates resilient networking strategies for manned–unmanned teaming networks. First, we propose centralized scheduling policies that incorporate redundancy in scheduling information to improve control-message delivery under adversarial conditions. We also develop a hybrid scheduling policy that opportunistically switches between centralized and decentralized scheduling, improving delay performance at low arrival rates while maintaining stability under heavy traffic by leveraging the benefits of both approaches. Second, we formulate a decision framework for relaying traffic from the unmanned swarm to a constrained manned controller as a Constrained Markov Decision Process (CMDP) under intermittent link failures. We present simple threshold-based relay policies, and we use Lagrangian relaxation and dynamic programming to identify policies that minimize delay subject to a relay constraint. Finally, we investigate swarm mobility in a search and rescue mission scenario using path loss-based wireless models, where the unmanned swarm must identify a loiter location that balances reliable communication with the manned platform and reliable detection of an emitter of interest. We study the use of multi-armed bandit algorithms for loiter-location decision making when the swarm does not have complete knowledge of the communication environment. These algorithms enable the swarm to learn effective positioning strategies under uncertainty and communication-sensing tradeoffs. Collectively, these contributions aim to provide a unified analytical and computational foundation for adaptive and resilient networking in manned–unmanned teams, enabling effective coordination and robust performance in contested, resource-limited environments.
Research Advisor:
Prof. Alex Wyglinski
ECE Department, WPI
Committee Members:
Prof. Bo Tang
ECE Department, WPI
Prof. Bashima Islam
ECE Department, WPI
Dr. Thomas Stahlbuhk
MIT Lincoln Laboratory