OPTIMIZATION OF SPACE TIME CODED MULTI-USER SYSTEMS THROUGH A VARIETY OF ALGORITHMS

Abstract

Multi-user systems such as Multi-Carrier Code Division Multiple Access (MC-CDMA) and Orthogonal Frequency Division Multiple Access (OFDMA) are very frequently use in modern communication networks like 4G, MOTO4A, etc. These systems utilize the available bandwidth effectively. In order to maximize the capacity and encounter the high data rate demand, much emphasis given to coding techniques for effective utilization of resources at the transmitter end. Similarly, a variety of algorithms is evolve to separate and detect the data of users at the receiver end of multi-user system. Optimization of both transmitter and receiver is still a very attractive research area for future communication systems. One of the famous coding techniques is Space Time Block Codes (STBC). STBC are implement in single layer and does not give so attractive results in multi-user systems. Nowadays, two layer coding mechanism is being used in Multiple Input Multiple Output (MIMO) systems. In this approach, coding process is apply in two layers for better data recovery. In multicarrier systems, Fourier transform is common for frequency domain spreading. It consumes more bandwidth due to addition of cyclic prefixes at transmitter end. Another transformation namely, Slantlet transform is receiving good attention due to its benefit of bandwidth cutback as it does not need addition of cyclic prefixes. The received data can be recover with the help of different conventional adaptive algorithms like Least Mean Square (LMS) and Recursive Least Square (RLS). The conventional LMS and RLS have slow convergence rate and high Bit Error Rate (BER) because of static step size iv and forgetting factor. The performance of these algorithms can be improve with adaptive step size in case of LMS and adaptive forgetting factor in case of RLS for multi-user systems like MC-CDMA and OFDMA. Unconventional algorithms like Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) also used to achieve better performance of multi-user systems. In this thesis, the initial work is divide in two parts on MC-CDMA systems. Initially, the two layered-CDMA codes are focus on MC-CDMA systems. The Multi User Detection (MUD) of this system is implement with conventional LMS and proposed Fuzzy logic based LMS. Later, the single layered space-time block codes is adopt on MC-CDMA systems for exploring the MUD different conventional LMS types. The GA is implement on the same system with proposed GA depth study for MUD. The cooperative coevolutionary algorithms are also test for MUD. In order to overcome high bandwidth consumption issue, the Slantlet transform is propose instead of Fourier Transform for Frequency/Time domain transformations. Lastly, the in depth of MUD on OFDMA is also the part of this thesis. The OFDMA systems are study with proposed Fuzzy LMS and Fuzzy RLS algorithms for MUD. The Soft-PSO is also implement for MUD. Lastly, the proposed algorithm namely Piranha Fish Optimization (PFO) is implement for MUD on OFDMA systems. The opposition based learning concept implemented on PFO. All the proposed schemes evaluated in well comparative manner, which shows its key significance in domain of multi-user systems.