Advanced Algorithms for Secure and Spectrally Efficient 5G-IoT Networks

Communication in today's world is no longer limited to people, but also extends to machines, inanimate objects and everyday devices. The Internet of Things (IoT), supported by the 5th Generation of communications (5G) provides ubiquitous connectivity across a wide range of applications. This work provides advanced algorithms to enable two key targets of 5G-IoT wireless networks namely, security and gigabit data rates. In the security realm, privacy is a major concern in 5G-IoT applications. An emerging paradigm which protects wireless communications against eavesdroppers is known as Physical Layer Security (PLS). PLS algorithms provide privacy at low latency by merging security algorithms with Physical Layer signal processing. This research develops the analysis of Key-Based Physical Layer Security (KB-PLS) which enables transmission of completely private information by mapping secret information bits onto MIMO precoders from a codebook. The research proposes three ways in which KB-PLS schemes could be integrated with 5G networks. First, in addition to the 5G control and data planes, networks could be augmented with a separate secrecy plane to incorporate security and privacy functionalities. Second, the research presents a new method to leverage time-frequency resources in order to facilitate ultra-low latency two-way private communications. Third, privacy based modifications to the 5G Cell Search procedure are presented as an example of a practical application of KB-PLS. From a performance standpoint, closed form expressions are derived for the Key-Bit Error Rate (KER) of a KB-PLS private data channel over 5G-like stochastic geometries. New results include the derivation of expressions for the secret key transmission rate and KER as a function of Signal to Noise Ratio (SNR). Both of these metrics aid in 5G network planning from a privacy perspective, an approach not previously considered in other works. Lastly, in keeping with the research directions for beyond-5G and 6G networks, possible machine learning extensions for precoder detection in KB-PLS are presented. In the high data rate realm, interference cancellation is a significant enabler. This can be achieved through a technique known as Interference Alignment (IA). IA is often studied in the context of a multi-user MIMO framework called K-User MIMO X, which facilitates all-to-all communication between K access points and K mobile devices. For such a network, the research illustrates the demodulation of K2 independent data streams through a new interference cancellation beamforming algorithm that improves spectral efficiency compared to massive MIMO. The research derives a multi-user Shannon Capacity formula for K-User MIMO with K ≥ 3. An Orthogonal Frequency Division Multiplexing (OFDM) frame structure is defined to demonstrate the allocation of time-frequency resources to pilot signals for channel estimation. The capacity formula is then refined to include realistic pilot overheads. A key result is a practical upper-bound for MIMO array sizes that balances estimation overhead and throughput. Lastly, simulation results show the practical capacity in small cell geometries under Rayleigh Fading conditions, with both perfect and realistic channel estimation.