**Advertisement**

By a

The patent's inventors are Wagner, Sebastian (Antibes, FR); Couillet, Romain (Antibes, FR).

This patent was filed on

From the background information supplied by the inventors, news correspondents obtained the following quote: "

"An advanced

"The performance of

"Precoding algorithms used in MU-MIMO are divided into non linear and linear precoding types.

"A non linear precoding technique is the so-called Dirty-Paper Coding (DPC) which pre-cancels the interference of the signal transmitted by the transmitter without any power penalties. The transmitter is assumed to know the interference signal regardless of channels state information knowledge. However, the use of this technique is inconvenient according to its cost and complexity and, for such reason, a linear precoding is often preferred.

"Indeed, linear precoding can achieve reasonable throughput performance with lower complexity and cost relative to nonlinear precoding. Linear precoding includes zero-forcing (ZF) and Regularized Zero-Forcing (R-ZF) precoding while non linear precoding includes Dirty-Paper Coding (DPC).

"Zero-forcing (ZF) technique addresses the drawback of high complexity and cost and performs close to the system capacity. However, it is required that the transmitter knows the downlink channel state information. Accordingly, in the case of limited or inaccurate channel state information, significant loss of the system throughput may occur.

"Advantageously to Zero Forcing (ZF), Regularized Zero-Forcing (R-ZF) precoding manages to compensate unfavorable channel conditions by computing a regularization parameter .alpha.. This computation is performed at the transmitter for each channel estimation (performed by the receiver) and has to satisfy the transmit sum power constraint. Unfortunately, the computation of a requires knowledge of the downlink channel at the transmitter and thus, in the case of ill-conditioned channel estimates the system throughput is degraded.

"In such a context, there is a strong desire to propose a simple solution that provides a satisfactory computation of the regularization parameter .alpha. and is independent of the channel estimation."

Supplementing the background information on this patent, VerticalNews reporters also obtained the inventors' summary information for this patent: "It is an object of the present invention to provide a simple process for computing the regularization parameter .alpha., said process being independent of the channel estimation.

"It is another object of the present invention to provide a process for computing the regularization parameter .alpha., said process depending on the spatial correlation of the transmit antennas at the transmitter and satisfying the transmit sum power constraint.

"These and other objects of the invention are achieved by means of a precoding process for a transmitter of a MU-MIMO communication system comprising M antennas in the transmitter and K User Equipments (UE), said precoding being based on a Regularized Zero Forcing (R-ZF) linear precoding of the type:

".times..function..times..times..alpha..times..times..times..times..times.- .times..alpha..times..times..times. ##EQU00001## With:

"H being a complex matrix representative of the K.times.M channel estimates, said H matrix comprising channel coefficients which are correlated with corresponding correlation coefficients being the entries of a correlation matrix .crclbar..sub.T,

"I.sub.K and I.sub.M being the identity matrix, respectively of size K.times.K and M.times.M;

"And .alpha. being the regularization parameter,

"The precoding process comprises the steps of:

"a) computing the correlation matrix .crclbar..sub.T;

"b) initializing two parameters .alpha..sub.max and .alpha..sub.min;

"c) computing .alpha.=1/2(.alpha..sub.max+.alpha..sub.min);

"d) determining for z=-.alpha. the value of S (S.sub.H'.sub..omega..sub..THETA..sub.T.sub.H'.sub..omega..sub.H(z)) solving the fixed-point equation:

".function..intg..lamda..mu..THETA..function..lamda..times.d.lamda..lamda.- .beta..times..function. ##EQU00002## With .lamda. being the eigenvalues of the correlation matrix .crclbar..sub.T, .mu. being the empirical eigenvalue distribution functions of .crclbar..sub.T and .beta. being M/K; e) determining for z=-.alpha. the value of S.sub.d

"dd.times..omega.'.times..THETA..times..omega.'.times..times..function. ##EQU00003## solving the fixed-point equation:

".function..intg..lamda..beta..times..times..lamda..beta..times..function. .mu..THETA..function..lamda..times.d.lamda..function. ##EQU00004## f) computing the equation

".beta..beta..beta..alpha..alpha..beta..beta..beta..alpha. ##EQU00005## g) adapting the interval [.alpha..sub.mm, .alpha..sub.max] as follows:

"if F-P

"In one particular embodiment, the computation of regularization parameter .alpha. is based on a WHILE loop being executed as long as the absolute value of F-P (ABS(F-P)) is superior to a predetermined value .epsilon..

"Preferably, the precoding is based on two successive FOR loops.

"A first FOR loop executed for a number of N occurrences which includes the steps of:

"initializing S to a predetermined value, for instance 0,

"performing, for n=1 to N, the following computation:

".alpha..times..times..lamda..lamda..beta..times. ##EQU00006##

"A second FOR loop which comprises the steps of:

"initializing S.sub.d to a predetermined value;

"performing, for n=1 to N, the following computation:

".times..times..lamda..beta..times..lamda..beta..times. ##EQU00007##

"Preferably, the transmitter comprises a number of antennas (M) which is equal to the number of UE(s) (K).

"The invention also achieves a transmitter for a MU-MIMO communication system comprising M antennas in the transmitter and K User Equipments (UE), said precoding being based on a Regularized Zero Forcing (R-ZF) linear precoding of the type mentioned above,

"The transmitter particularly includes:

"means for computing the correlation matrix .crclbar..sub.T;

"means for initializing the two parameters .alpha..sub.max and .alpha..sub.min

"means for computing .alpha.=1/2(.alpha..sub.max.alpha..sub.min);

"means for determining for z=-.alpha. the value of S solving the fixed-point equation:

".function..intg..lamda..mu..THETA..function..lamda..times.d.lamda..lamda.- .beta..times..function. ##EQU00008## means for determining for z=-.alpha. the value of S.sub.d

"dd.times..omega.'.times..THETA..times..omega.'.times..times..function. ##EQU00009## solving the fixed-point equation:

".function..intg..lamda..beta..times..times..lamda..beta..times..function.- .times..mu..THETA..function..lamda..times.d.lamda..function. ##EQU00010## means for computing the equation

".beta..beta..beta..alpha..alpha..function..beta..beta..beta..alpha. ##EQU00011## means for adapting the interval [.alpha..sub.min, .alpha..sub.max] by using the following process:

"if F-P

"In one embodiment, the transmitter includes means for performing one WHILE loop and two FOR loops for computing the regularization parameter .alpha."

For the URL and additional information on this patent, see: Wagner, Sebastian; Couillet, Romain. Precoding Process for a Transmitter of a MU-MIMO Communication System. U.S. Patent Number 8649457, filed

Keywords for this news article include: Technology,

Our reports deliver fact-based news of research and discoveries from around the world. Copyright 2014, NewsRx LLC