2022-12-1920222022-122022-12-19December 2https://hdl.handle.net/10388/14379In wireless communication, channel fading refers to a destructive superposition of delayed copies of a transmitted signal at the receiver. The effect of channel fading appears as random amplitude attenuation and random phase shift of the transmitted signal. Unless properly dealt with, channel fading can severely disrupt reliable communication. In general, there are three different strategies in designing transmission schemes for wireless communication systems: coherent, non-coherent differential, and non-coherent energy-based modulation (EBM) schemes. Each technique has its own merits and suits certain application scenarios. In particular, non-coherent EBM enjoys a simple structure, a vital requirement in emerging technologies such as massive antenna systems. Furthermore, EBM is the only choice for very rapidly-varying channels, where both coherent and non-coherent differential schemes fail. The overall objective of this thesis is to advance the theory and practice of EBM. The focus of the research is on the integration of an emerging technique, known as index modulation (IM), into orthogonal frequency-division multiplexing (OFDM) to provide new design guidelines and obtain novel EBM schemes that are superior to existing schemes in terms of spectral efficiency and/or error performance. To date, OFDM is still the most popular transmission technique for wide-band channels in which the whole channel bandwidth is divided into multiple narrow-band sub-channels that are modulated independently by the information bits. On the other hand, using patterns of active and inactive sub-channels to transmit a portion of information bits in the index domain is the main principle of IM. The set of activation patterns in the IM technique can be conveniently represented as a subset of a binary codebook, i.e., comprising of codewords whose elements are 1 (corresponding to active) or 0 (corresponding to inactive). The thesis contains the following main contributions. First, a generalized non-coherent OFDM-IM system is proposed. Compared to the original system, the generalization relaxes the restriction that the set of activation patterns of OFDM sub-channels comes from a constant-weight and maximum-size codebook. By employing a more flexible codebook for activation patterns, an interesting rate-diversity tradeoff (i.e., a tradeoff between the data rate and error performance) is obtained. In particular, it is shown that the proposed system can achieve a diversity order equal to the minimum partial Hamming distance of the employed code. The tradeoff for achieving a higher diversity order (hence, better error performance) is the lower system's spectral efficiency. Our analysis takes into account a realistic assumption of correlated sub-channels, and we also develop a near-optimal low-complexity receiver for the proposed system. Furthermore, the error performance of the proposed system is analyzed in multiple-antenna settings, and in very fast-fading environments that cause inter-carrier interference. Next, the problem of code design for the generalized non-coherent OFDM-IM system is addressed. By establishing a channel equivalence relationship, it is shown that employing activation patterns that are obtained from a t-error correcting code designed for a Z-channel can achieve a diversity order of t+1 in the proposed system. To address the low spectral efficiency limitation of the generalized non-coherent OFDM-IM system, our next contribution develops a non-coherent multi-level IM system. In such a system, active indices are allowed to take multiple levels, not just binary levels. This is accomplished by using a non-binary code to represent the set of activation patterns. Performance of the proposed system is analyzed under the (optimal) maximum likelihood receiver and when the set of activation patterns comes from a multi-level code generated from asymptotically optimal alphabets. An asymptotic analysis of the pair-wise error probability shows that the proposed system can exploit a diversity order that is determined by the distance of the worst codeword pair in the l_1 metric, known as the Manhattan norm. The rate-diversity tradeoff for the proposed multi-level IM system as a function of the code length is obtained by packing in the Manhattan metric. It is shown that the proposed system can significantly outperform its two-level counterpart in terms of spectral efficiency and achievable diversity order, i.e., in both data rates and error performance. Although being optimal, the maximum likelihood receiver for the proposed multi-level non-coherent IM system has a very high complexity. The last contribution of this thesis develops a near-optimal low-complexity receiver for the proposed non-coherent multi-level IM system. The developed receiver is an iterative receiver that performs turbo processing between a soft-output demodulator and a soft-output channel decoder. The error performance of the resulting system is analytically studied. The analysis reveals how the mapping from coded bits to energy levels influences the diversity order and coding gain of the system. Furthermore, a design criterion for good mappings is formulated and an algorithm is proposed to find a set of best mappings for the proposed system. The main analytical findings as well as the excellent performance of the proposed system is thoroughly corroborated by simulation results.application/pdfenIndex modulationOFDM-IMdiversity ordernoncoherent detectionZ-channelachievable rateEnergy-based modulationbit-interleaved coded modulationiterative decodingThesis2022-12-19