Research On Performance Of Space Time Diversity And Precoding

The information theory of MIMO system,established by Telatar and Foschini,shows the capacity of MIMO system is as several times as the capacity of SISO system.Their new information theory breaks the theory frame of wireless communicationsystem founded by Shannon,which promotes traditional communication techniquestaking a conversion to the MIMO techniques rapidly.Among those pop fields of MIMOtechniques,the BER of MIMO system,another important parameter of acommunication system,is concerned greatly by researchers now.The development onBER analysis of MIMO system is very slow,because of the complex structure of therandom channel matrix,and deep mathematical foundation problem.We have a deepresearch on this problem.Meanwhile,the precoding technique of orthogonal space-timecoding is discussed in this dissertation.A class of MIMO system is researched in this dissertation.In this MIMO system,symbol will be transmitted several times to obtain better temporal diversity.Differenttransmitting and receiving gain is adopted when symbol is transmitted within a blockperiod.The power of the transmitted signal in MIMO system can be controlled by thegain of transmit antennas,and we can maximize the SNR of received signal byadjusting receiver antennas gain.Three classes of methods describe the powercontrolling of transmit antennas,which divide the communication system into threegroups:System with spatial transmitted diversity,Trace constrained system,andEigenvalue constrained system.The different weight vectors of the symbol form aweight matrix of the symbol for each antenna.If there are several different symbolsneed to be transmitted in a block period,we can select proper weights for each symbolto construct a special space-time block coding.In the research of BER of a MIMO system,based on the deduced equivalentlimited integral expression of function Q(x)and the special expression of the integralfunction in Q(x),we turn the problem from the difficult expectation of a randomfunction to the expectation problem of a class of random function,which is mild.Meanwhile,a Karhunen-Loeve transformation is introduced in this dissertation,which turns the coefficients vector of correlated subchannels to be uncorrelated one.Applyingthis obtained methods to the MIMO system with different power constrains,the BERproblem of some systems is researched under correlated Rayleigh fading channel andindependent Rayleigh fading channel.Five results are obtained.(1) BER of MIMOsystem with spatial transmitted diversity.(2) BER of MIMO trace constrained system,under a kind of channel,where the vectors with all channel coefficients to jth receiveantenna,and the vector with all channel coefficients to kth receive antenna,are linear.(3)BER of MIMO eigenvalue constrained system.(4) The BER formula and its low andupper boundness of MIMO system with eigenvalue constrain,when the fading channelis correlated in the transmitter.(5) The BER formula and its low and upper boundness ofMIMO system with eigenvalue constrain,when the fading channel is correlated in thereceiver.In the research of precoding of orthogonal space-time coding,based on theestimation inequation of the expectation of a complex random function of Q(x),theauthor obtained a new boundness of BER for the orthogonal space-time coding MIMOsystem.Based on the new boundness of BER,we propose a new criterion for precodingof orthogonal space-time coding,and format the optimal problem for the precodingmatrix.Precoding matrices are obtained when the number of transmit antennas is notlarger than the space dimension of an orthogonal space-time coding,under the fadingwith form of the Kronecker product or non-Kronecker product at the transmitter.Noproper precoding matrix is found to meet all constraints when the number of transmitantennas is larger than the space dimension of an orthogonal space-time coding MIMOsystem.The optimal problem proposed by A.Hj(?)rungnes and J.Akhtar,can be deducedeasily by the methods formed in this dissertation.The caculation difficulty to obtain theprecoding matrix is decreased obviously when the proposed criterion is applied to thesystem correlated at the transmitter.