Date of Award


Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy in Electrical Engineering (PhD)

Administrative Home Department

Department of Electrical and Computer Engineering

Advisor 1

Zhaohui Wang

Committee Member 1

Daniel R. Fuhrmann

Committee Member 2

Min Song

Committee Member 3

Ossama Abdelkhalik


Underwater acoustic (UWA) communication networks are promising techniques for medium- to long-range wireless information transfer in aquatic applications. The harsh and dynamic water environment poses grand challenges to the design of UWA networks. This dissertation leverages the advances in machine learning and signal processing to develop intelligent and secure UWA communication networks. Three research topics are studied: 1) reinforcement learning (RL)-based adaptive transmission in UWA channels; 2) reinforcement learning-based adaptive trajectory planning for autonomous underwater vehicles (AUVs) in under-ice environments; 3) signal alignment to secure underwater coordinated multipoint (CoMP) transmissions.

First, a RL-based algorithm is developed for adaptive transmission in long-term operating UWA point-to-point communication systems. The UWA channel dynamics are learned and exploited to trade off energy consumption with information delivery latency. The adaptive transmission problem is formulated as a partially observable Markov decision process (POMDP) which is solved by a Monte Carlo sampling-based approach, and an expectation-maximization-type of algorithm is developed to recursively estimate the channel model parameters. The experimental data processing reveals that the proposed algorithm achieves a good balance between energy efficiency and information delivery latency.

Secondly, an online learning-based algorithm is developed for adaptive trajectory planning of multiple AUVs in under-ice environments to reconstruct a water parameter field of interest. The field knowledge is learned online to guide the trajectories of AUVs for collection of informative water parameter samples in the near future. The trajectory planning problem is formulated as a Markov decision process (MDP) which is solved by an actor-critic algorithm, where the field knowledge is estimated online using the Gaussian process regression. The simulation results show that the proposed algorithm achieves the performance close to a benchmark method that assumes perfect field knowledge.

Thirdly, the dissertation presents a signal alignment method to secure underwater CoMP transmissions of geographically distributed antenna elements (DAEs) against eavesdropping. Exploiting the low sound speed in water and the spatial diversity of DAEs, the signal alignment method is developed such that useful signals will collide at the eavesdropper while stay collision-free at the legitimate user. The signal alignment mechanism is formulated as a mixed integer and nonlinear optimization problem which is solved through a combination of the simulated annealing method and the linear programming. Taking the orthogonal frequency-division multiplexing (OFDM) as the modulation technique, simulation and emulated experimental results demonstrate that the proposed method significantly degrades the eavesdropper's interception capability.