On Low-Complexity High-Rate Space-Time Codes

Space-time signal processing theory in multi-antenna wireless systems developed atthe end of the 20th century becomes a novel field in the wireless communication technol-ogy research. Studies on high-rate space-time codes, especially encoding (construction)and decoding of low-complexity codes are of significance to the future high-rate high-quality wireless communication systems.With the increasing requirements on the quantity and the quality of information trans-mission in the modern society, the wireless communication system capacity should begreatly improved. Hence the increasing shortage of usable frequency spectrum is furtheraggravated, which has become one of the maj or problems faced currently in theoretical re-search and system implementation. Recently, high-rate space-time codes in multi-antennawireless systems supply a new approach to solve this problem. Moreover, constructionson low-complexity high-rate space-time codes guarantee the possibility of project imple-mentation. Under this circumstance, we focus on high-rate space-time codes, especiallyconstructions of low-complexity codes and project implementation in this dissertation.We analyze the common character of these existing low-complexity codes and thenstudy two non-pure orthogonal space-time codes in this dissertation: investigation andconstruction of group-decodable space-time codes; definition, construction and decod-ing of block-orthogonal space-time codes. Summarily, the contents of this dissertationinclude:1 .The identifiability of space-time coding systems is deeply studied.(1) The space-time coding matrix-form system model is improved;(2) The condition for identified space-time coding systems is derived;(3) The proof of the maximum achievable code rate of identified space-time codingsystems is presented.The research results show that the identifiability of space-time coding systems isdetermined by both the linear independence among dispersion matrices and the receiveantenna number, based on which we give a proof to the conclusion "the code rate of iden-tified space-time coding systems can not exceed the minimum of transmit and receiveantenna numbers." 2. Based on the space-time decoder using QR decomposition, we analyze the com-mon property of the existing low-complexity codes, and find that the orthogonality be-tween any two information symbols in the code matrix may bring the decoding complexityreduction, which is out of the scope of the conventional (quasi-)orthogonal codes.3. From the theory of linear equations, the theoretical fundamental of group-decodablestructure is analyzed, and constructions of group-decodable codes are proposed systemat-ically.(1) Based on the analysis on quasi-orthogonal constraint (QOC), group-decodableequation is derived;(2) The characteristic of unbalanced 2-group-decodable codes is analyzed, then theupper bound of code rates of unbalanced 2-group-decodable codes is found and a system-atic construction algorithm is proposed;(3) Based on unbalanced 2-group-decodable codes, constructions of balanced andquasi-balanced 2-group-decodable codes are presented;(4) A construction algorithm for multigroup-decodable (diversity embedded) space-time codes is also proposed;(5) With the usage of group-decodable equation, a method to verify whether the max-imum code rate of existing codes has been achieved is described.The research works show that the proposed construction algorithms for group-decodablecodes is more scalable in code length, transmit antenna number and code rate. Existinghigh-rate 2-group-decodable code constructions are special cases of our balanced 2-group-decodable code construction. Moreover, high-rate multigroup-decodable code construc-tion is proposed and high-rate multigroup-decodable code example is presented for thefirst time.4. A new non-pure orthogonal space-time codes, block-orthogonal space-time code,is defined. Then a simplified decoding with reduced complexity for block-orthogonalstructure is studied for the first time. At last, constructions and code examples of block-orthogonal space-time codes are presented.(1) Block-orthogonal code is defined based on the character of the block-orthogonalstructure. Some existing high-rate codes have held block-orthogonal structures(2) A simplified decoding with reduced complexity is proposed for the first time, where the complexity reduction is brought from the block-orthogonal structure, but with-out any extra performance penalty;(3) To further explore the block-orthogonal structure, block-orthogonal code con-structions are proposed newly and several code examples are presented;(4) For project implementation, simulation results on bit error rate (BER) againstdecoding complexity are given, and the performance of space-time coding systems withlimited computational capacity is studied deeply.Research results show that block-orthogonal structure can bring decoding complexityreduction without performance penalty. The proposed block-orthogonal space-time codescan outperform the best previously known codes with certain decoding complexity.5. The expression of the complexity of a breadth-first search algorithm, QR de-coder based on M-algorithm (QRDM), is derived for both non-considering and consider-ing block-orthogonal structure cases. Then the decoding-complexity-BER curves of anSTC can be plotted for the first time, which can help us to evaluate the true value of thiscode for a multi-antenna wireless system.Finally, the proposed codes (i.e., group-decodable codes and block-orthogonal codes)are compared and the suitable environment for application is analyzed. The principles ofspace-time code selection are also given for proj ect implementations. And future prospectsare presented.