(A) study on machine learning aided physical layer security for wireless communications

Abstract

This dissertation considers machine learning aided physical layer security for wireless communications. At first, we propose an artificial noise (AN)-aided secure multiple-input single-output non-orthogonal multiple access (NOMA) scheme. In the design of the proposed scheme, we consider fairness that all the users have higher secrecy rates as compared to those given by a competing orthogonal multiple access (OMA) scheme. Despite its importance, the fairness aware design has remained rarely touched since it is mathematically intractable. This work shows that this problem can be efficiently solved by utilizing a deep neural network as the precoder for the information and AN signals even without resorting to some assumptions such as a large antenna array and/or a high signal-to-noise ratio. We will also propose an adaptive mode that switches the access protocol from the OMA scheme to the NOMA scheme only when fairness is met. The performance of the proposed secure NOMA scheme will be extensively evaluated and compared with existing NOMA and OMA schemes. The comparisons clearly show that the sum secrecy rate can be significantly improved while guaranteeing the fairness, which however cannot be achieved with the existing NOMA scheme. Secondly, we study a covert communication scheme for an uplink multi-user scenario in which some users are opportunistically selected to help a covert user. In particular, the selected users emit interference signals via an orthogonal resource dedicated to the covert user together with signals for their own communications using orthogonal resources allocated to the selected users, which helps the covert user hide the presence of the covert communication. For the covert communication scheme, we carry out extensive analysis and find system parameters in closed forms. The analytic derivations for the system parameters allow one to find the optimal combination of system parameters by performing a simple one-dimensional search. In addition, the analytic results elucidate relations among the system parameters. In particular, it will be proved that the optimal strategy for the non-covert users is an on-off scheme with equal transmit power. The theoretical results derived in this work are confirmed by comparing them with numerical results obtained with exhaustive searches. We demonstrate that the results of work can be utilized in versatile ways by demonstrating a design of covert communication with energy efficiency into account. Finally, we study a covert communication scheme in which some users are opportunistically selected to emit interference signals for the purpose of hiding the communication of a covert user. The study is conducted in a generic setup where the channels between pairs of entities in the scheme are allowed to be correlated. For the setup, we optimize system parameters of the scheme utilizing Q-learning, which however is plagued with long learning time and large storage space when the dimension of the state gets large and/or a fine resolution of reward function value is necessary. To resolve the technical challenge, this work proposes a scalable Q-learning which recursively narrows down the discretization level of the continuous state in an iterative fashion. To confirm the results in this work, the system parameters are evaluated with theoretical results for independent channels and compared with the ones from the proposed scalable Q-learning results. In addition, this work demonstrates that the channel correlation is beneficial to have a higher throughput of covert communication, which has never been discussed before.