Research On Equalization Technologies Of Short Burst Signals

Due to its shortness and burstiness, the short burst communication has superior performance in anti-interference and resistance to intercept, and has found its applications in more and more fields, especially in the military, such as the submarine communication system and the widely equipped Joint Tactical Information Distribution System(JTIDS, Link16) in NATO countries. The short burst signals, which usually contain only tens or hundreds of codes, are difficult to detect or interfere for enemies, and it’s also much more difficult to receive than the long-time signals for the cooperators. In order to reduce the inter symbol interference(ISI) and the bit error rate(BER), the equalization about the channel for transmission at the receiving end deserves conducted to guarantee normal communication. The aim of this thesis is to study the channel equalization of the short burst signals.The main work and research results are listed as follows.(1) Introduce the short burst signal and its main features(i.e. the shortness and burstiness) briefly, describe the desired convergence performance aqccording to its signal character, and introduce the "data-reusing"(DR) method that suitable to deal with the short burst signals.(2) Describe the DR method briefly, and introduce the typical algorithms that based on DR, i.e. the affine projection algorithm(APA) and affine projection sign algorithm(APSA). Descriptions about the backgrounds, applications, the constrained optimization models and the effect of the main parameters(i.e. the step-size, projection order) of the APA and APSA are also given.(3) Reconsider the APA from 3 aspects, namely the variable step-size, the variable projection order and the combined algorithm. As for the variable step-size, the conventional update equation of the APA is modified based on the singular value decomposition and pseudo inverse of matrix, and then the variable step-size of the modified APA is derived via the mean-square deviation(MSD) analysis. As for the variable projection order, the approximate expression of the MSD is derived and the projection order is obtained by minimum the delta of the MSD at each iteration. As for the combined algorithm, the APA and normalized least mean square(NLMS) with variable step-sizes are adaptively combined according to the existing combination scheme. The variable mixing parameter used in the combination is derived by minimum the MSD at each iteration.(4) Reconsider the APSA, and a modification about the APSA is made since the output error contains more information about the unknown system than its sign information. The variable step-size of the modified APSA is derived based on a posteriori estimation error(APEE) analysis.(5) All the new algorithms in the thesis are fully analysed via theoretical derivations and computer simulations. Moreover, the convergence performance under different situations, the effect of the involved parameters and the selection of those parameters when dealing with short burst signals are presented.(6) Summarize and reflect on the work of the thesis, and consider the possible ideas that deserve further research.