Document Type
Dissertation
Date of Award
Fall 1-31-2016
Degree Name
Doctor of Philosophy in Electrical Engineering - (Ph.D.)
Department
Electrical and Computer Engineering
First Advisor
Ali N. Akansu
Second Advisor
Ali Abdi
Third Advisor
Richard A. Haddad
Fourth Advisor
Alexander Haimovich
Fifth Advisor
Eliza Zoi-Heleni Michalopoulou
Abstract
Discrete Fourier Transform (DFT) is a restricted version of Generalized DFT (GDFT) which offers a very limited number of sets to be used in a multicarrier communication system. In contrast, as an extension on Discrete Fourier Transform (DFT) from the linear phase to non-linear phase, the proposed GDFT provides many possible carrier sets of various lengths with comparable or better performance than DFT. The availability of the rich library of orthogonal constant amplitude transforms with good performance allows people to design adaptive systems where user code allocations are made dynamically to exploit the current channel conditions in order to deliver better performance.
For MIMO Radar systems, the ideal case to detect a moving target is when all waveforms are orthogonal, which can provide an accurate estimation. But this is not practical in distributed MIMO radars, where sensors are at varying distances from a target. Orthogonal waveforms with low auto- and cross-correlations are of great interest for MIMO radar applications with distributed antennas. Finite length orthogonal codes are required in real-world applications where frequency selectivity and signal correlation features of the optimal subspace are compromised. In the first part of the dissertation, a method is addressed to design optimal waveforms which meets above requirements for various radar systems by designing the phase shaping function (PSF) of GDFT framework with non-linear phase.
Multicarrier transmission such as orthogonal frequency-division multiplexing (OFDM) has seen a rise in popularity in wireless communication, as it offers a promising choice for high speed data rate transmission. Meanwhile, high peak-to-average power ratio (PAPR) is one of the well-known drawbacks of the OFDM system due to reduced power efficiency in non-linear modules. Such a situation leads to inefficient amplification and increases the cost of the system, or increases in interference and signal distortion. Therefore, PAPR reduction techniques play an essential role to improve power efficiency in the OFDM systems. There has been a variety of PAPR reduction methods emphasizing different aspects proposed in the literature. The trade-off for PAPR reduction in the existing methods is either increased average power and/or added computational complexity. A new PAPR reduction scheme is proposed that implements a pre-designed symbol alphabet modifier matrix (SAM) to jointly modify the amplitude and phase values of the original data symbol alphabets prior to the IFFT operation of an OFDM system at the transmitter. The method formulated with the GDFT offers a low-complexity framework in four proposed cases devised to be independent of original data symbols. Without degrading the bit error rate (BER) performance, it formulates PAPR reduction problem elegantly and outperforms partial transmit sequences (PTS), selected mapping technique (SLM) and Walsh Hadamard transform (WHT-OFDM) significantly for the communication scenarios considered in the dissertation.
Recommended Citation
Wang, Yuewen, "Generalized DFT: extensions in communications" (2016). Dissertations. 63.
https://digitalcommons.njit.edu/dissertations/63
