Synthesis of a Non-Recurrent Regularazing Algorithm of M-QAM Signal Detection in MIMO

This paper proposes a regularizing non-recurrent algorithm based on a modified method of least squares (LS) for reception of the signal with quadrature amplitude modulation (M-QAM) in a system with multiple transmitting and receiving antennas (MIMO), in order to obtain soft decisions. The regularizing parameter (Lagrange multiplier) was determined from the condition of knowing the variance of the additive noise and the information sequence. Hard decisions were found according to the criterion of the minimum distance between soft decisions and possible values of the symbols of the information sequence. The signal was received under conditions of a communication channel with Rayleigh fading and Doppler spread of the spectrum, as well as with additive Gaussian white noise. The channel matrix takes into account the distortions caused by the signal propagation in the channel and in the path of the direct conversion receiver. The noise immunity of 4-QAM signal reception in a system without coding with eight transmitting and receiving antennas (MIMO 8×8) was compared using a regularizing algorithm and a procedure with a Lagrange multiplier equal to one, as well as an algorithm that works according to the criterion of the minimum mean square error (RMS), and Zero Forcing algorithm. The computational experiment was carried out for two cases: precisely known and estimated parameters of the communication channel.