Hiba Dakdouk

• • Designed and implemented a reinforcement learning control system for remote actuators operating over unreliable wireless networks, using Python and Gymnasium/MuJoCo for simulation. Developed and benchmarked multiple approaches to improve robustness under quantized, noisy state observations.

• • Investigated multi-armed bandit (MAB) algorithms for resource allocation in industrial IoT systems over unreliable communication channels, including development and implementation of a stochastic queue sampling strategy to improve reward estimation accuracy.

• • Evaluated energy-performance tradeoffs via simulation (CodeCarbon).

• • Developed signal processing and blind-source separation techniques to detect and predict moving obstacles in dense mmWave wireless networks, contributing to 6G integrated sensing research.

• • Built a cooperative sensing framework combining multiple device inputs to track object movement in real time.

• • Developed collaborative multi-armed bandit algorithms to optimize resource allocation and fairness across IoT de vices.

• • Built a LoRa network simulator in C, and benchmarked algorithms performance against state-of-the-art methods in terms of reliability and energy consumption.