CN102136891A - Improved channel state information (CSI) limited feedback method for multi-user multiple input multiple output (MU-MIMO) system - Google Patents

Abstract

The invention discloses an improved channel state information (CSI) limited feedback method. In the method, a signal detection algorithm of a receiver and a channel direction information (CDI) selection are combined for designing to maximize E(SINRk). If a multi-user multiple input multiple output (MU-MIMO) system is required to be converted into a multi-user multiple input single output (MU-MISO) system by antenna combining on the premise of limited feedback, higher system performance can be realized by an improved antenna combining algorithm provided by the invention compared with the other currently researched combining algorithms.

Description

A kind of improved CSI limited feedback method in the MU-MIMO system

Technical field

The invention belongs to wireless communication technology field, relate to channel condition information (CSI) Limited Feedback technology.

Background technology

In present Communication System Design, be accompanied by always and how improve wireless data transmission rate.Past, the improved work of modulation and scheduling aspect was applied in the present wireless communication system gradually in chnnel coding.Next generation mobile communication system needs to make full use of the multifarious adaptive technique of channel.These adaptive techniques allow transmitting terminal signal and instant channel condition to be complementary.This has just shown that transmitting terminal need obtain some information of wireless channel, and this information is commonly referred to channel condition information (CSI) at transmitting terminal, such as: channel status, received power, interfere information.In the frequency division duplex system in the past, most of channel self-adapting technology is inapplicable, because the channel of up-downgoing is independent incoherent in the frequency division duplex system, and is taking different frequency bands.For digital signal, the feedback of these information all is to quantize by limited bit, so the feedback here is called the feedback of limited channel condition information again.

Compare with the SU-MIMO system, the MU-MIMO system separates a plurality of users on the spatial domain by appropriate signal processing thereby has bigger power system capacity, so obtained bigger concern.And restrictive condition is exactly, and the base station end of descending MIMO must accurately be known channel condition information.Even now, in the frequency division duplex system of reality, because the restriction of Limited Feedback, so the base station end is difficult to accurately know channel condition information.Therefore, for the feedback of MU-MIMO, research how to feed back with Limited Feedback be very important to Effect on Performance.

Pre-coding matrix in the CSI Limited Feedback of MIMO does not pre-set, and each user needs to base station feedback CSI, and the base station utilizes the information of each user feedback to carry out user's selection and pre-coding matrix calculates.Compare with the feedback based on code book, the CSI of feedback user makes base station scheduling have better flexibility, and this advantage is more outstanding in MU-MIMO.At present, the research of antenna merge algorithm in the MU-MIMO feedback scheme (Antenna Combining Algorithm) is also more and more, antenna merge algorithm in some documents requires vector of each user feedback as partial channel knowledge, and therefore also just having limited each user can only transmit a data flow.Many antenna merge algorithms of usefulness have based on the feedback of eigenvalue of maximum vector sum based on minimum quantization error at present, when they merge vector in the design weighting, all respectively merge the MISO channel that the back equivalent channel is designed to have maximum channel power and minimum quantization error through weighting.But these two kinds of antenna merge algorithms are not all from maximization E[SINR _k] angle set out.The present invention proposes a kind of improved CSI limited feedback method, this method is considered to select gang to design the signal detection algorithm of receiving terminal and CDI, in the hope of can be with E[SINR _k] maximization.

Summary of the invention

The present invention proposes a kind of improved CSI limited feedback method, this method is considered to select gang to design the signal detection algorithm of receiving terminal and CDI, in the hope of can be with E[SINR _k] maximization.

Technical scheme of the present invention may further comprise the steps:

Step 1: the CSI-RS that the user utilizes the base station to send carries out channel estimating; For k user, through u _kCan represent by following formula as its received signal after the input, wherein

y _kBe the signal that user k receives, n ' _kBe the white complex gaussian noise that is superposeed on the equivalent channel, G _iBe the precoding vector of i user M * 1:

$r_k=u_k{y}_k=u_k{H}_k{G}_k{d}_k+u_k{H}_k\underset{i\∈S,i\≠k}{\mathrm{\Σ}}{G}_i{d}_i+n_k^{\′}$

Step 2: the user is by quantizing directional information CDI and the corresponding channel quality information CQI that code book feeds back its equivalent channel:

(2a) merge algorithm:

The SINR that the selected user k that transmits data is obtained _kCan be expressed as:

${\mathrm{SINR}}_k=\frac{p_k{\left|u_k{H}_k{w}_k\right|}^2}{1+\underset{i\∈S\backslash \left\{k\right\}}{\mathrm{\Σ}}{P}_i{\left|u_k{H}_k{w}_i\right|}^2}$

W wherein _kAnd p _kBe respectively k user's precoding vector and transmitted power, supposing the system carries out the constant power distribution to each user here, so

The selection that improved antenna merge algorithm is intended to the input of associating receiving terminal and CDI makes each user's E[SINR _k] maximization.

(2b) calculating of the selection of channel directional information (CDI) and channel quality information (CQI):

K user's CQI can be expressed as:

${\mathrm{CQI}}_k=\frac{\frac{P}{M}{\left|u_k{H}_k{v}_k\right|}^2}{1+\frac{P}{M}{\left|\right|u_k{H}_k-\left(u_k{H}_k{v}_k\right)v_k^H\left|\right|}_2^2}$

If ${\mathrm{CQI}}_k=\frac{u_k{A}_k{u}_{\mathrm{<figure-callout\; id="1"\; label="k\; H"\; filenames="HSA00000496490700011.png,HSA00000496490700021.png"\; state="\{\{state\}\}">k</figure-callout>}}^H}{1+u_k{B}_k{u}_k^H},A_k=\frac{P}{M}\left(H_k{c}_i{c}_i^H{H}_k^H\right),B_k=\frac{P}{M}\left[H_k(I-c_i{c}_i^H)H_k^H\right]$

At c _iDuring as the CDI of equivalent channel, select the MMSE signal detection algorithm as maximization E[SINR _k] the receiving terminal merge algorithm; Wherein, the weighting of devise optimum merges vectorial u under fixing quantization code _k:

$u_k={\left({(I+B_k)}^{-1}\sqrt{\frac{P}{M}}{H}_k{c}_i\right)}^H$

(2c) its corresponding weighting merging vector of each quantification vector sum will produce a CQI like this, and the user therefrom selects the channel quality information of the CQI of a maximum as its feedback, feeds back the pairing c of this CQI simultaneously _iCDI as its local channel.The base station end utilizes the CDI of each user feedback and CQI carries out user's scheduling and pre-coding matrix calculates, and the receiving end signal detection algorithm here adopts MMSE.

The algorithm that the present invention proposes has superior performance through simulating, verifying:

Simulation result shows that improved antenna merge algorithm all provides more excellent systematic function with respect to the antenna merge algorithm based on main characteristic vector and minimum quantization error under all signal to noise ratio conditions.And along with the increase of signal to noise ratio, it is increasing that the gap of their performances becomes; If need the MU-MIMO system is merged the MU-MISO system that is converted into by antenna under the prerequisite of Limited Feedback, so improved antenna merge algorithm can obtain more excellent systematic function than present other merge algorithms of studying.

Description of drawings

Fig. 1 MU-MIMO system block diagram

The average system capacity of Fig. 2 different antennae merge algorithm

The power system capacity CDF of Fig. 3 different antennae merge algorithm

Embodiment

Technical scheme of the present invention may further comprise the steps:

Step 1: the CSI-RS that the user utilizes the base station to send carries out channel estimating; For k user, through u _kCan represent by following formula as its received signal after the input, wherein

y _kBe the signal that user k receives, n ' _kBe the white complex gaussian noise that is superposeed on the equivalent channel, G _iBe the precoding vector of i user M * 1:

$r_k=u_k{y}_k=u_k{H}_k{G}_k{d}_k+u_k{H}_k\underset{i\&Element;S,i\&NotEqual;k}{\mathrm{\&Sigma;}}{G}_i{d}_i+n_k^{\&prime;}$

Step 2: the user is by quantizing directional information CDI and the corresponding channel quality information CQI:(2a that code book feeds back its equivalent channel) merge algorithm:

The SINR that the selected user k that transmits data is obtained _kCan be expressed as:

${\mathrm{SINR}}_k=\frac{p_k{\left|u_k{H}_k{w}_k\right|}^2}{1+\underset{i\&Element;S\backslash \left\{k\right\}}{\mathrm{\&Sigma;}}{P}_i{\left|u_k{H}_k{w}_i\right|}^2}$

W wherein _kAnd p _kBe respectively k user's precoding vector and transmitted power, supposing the system carries out the constant power distribution to each user here, so

The selection that improved antenna merge algorithm is intended to the input of associating receiving terminal and CDI makes each user's E[SINR _k] maximization.

(2b) calculating of the selection of channel directional information (CDI) and channel quality information (CQI):

K user's CQI can be expressed as:

${\mathrm{CQI}}_k=\frac{\frac{P}{M}{\left|u_k{H}_k{v}_k\right|}^2}{1+\frac{P}{M}{\left|\right|u_k{H}_k-\left(u_k{H}_k{v}_k\right)v_k^H\left|\right|}_2^2}$

If ${\mathrm{CQI}}_k=\frac{u_k{A}_k{u}_{\mathrm{<figure-callout\; id="1"\; label="k\; H"\; filenames="HSA00000496490700011.png,HSA00000496490700021.png"\; state="\{\{state\}\}">k</figure-callout>}}^H}{1+u_k{B}_k{u}_k^H},A_k=\frac{P}{M}\left(H_k{c}_i{c}_i^H{H}_k^H\right),B_k=\frac{P}{M}\left[H_k(I-c_i{c}_i^H)H_k^H\right]$

At c _iDuring as the CDI of equivalent channel, select the MMSE signal detection algorithm as maximization E[SINR _k] the receiving terminal merge algorithm; Wherein, the weighting of devise optimum merges vectorial u under fixing quantization code _k:

$u_k={\left({(I+B_k)}^{-1}\sqrt{\frac{P}{M}}{H}_k{c}_i\right)}^H$

(2c) its corresponding weighting merging vector of each quantification vector sum will produce a CQI like this, and the user therefrom selects the channel quality information of the CQI of a maximum as its feedback, feeds back the pairing c of this CQI simultaneously _iCDI as its local channel.The base station end utilizes the CDI of each user feedback and CQI carries out user's scheduling and pre-coding matrix calculates, and the receiving end signal detection algorithm here adopts MMSE.

Effect of the present invention can further be verified by following emulation:

1) emulation content:

The power system capacity of utilizing antenna to merge among the MU-MIMO is carried out emulation, wherein adopt different antenna merge algorithm (based on the antenna merge algorithm of main characteristic vector, minimum quantization error, improved antenna merge algorithm)

2) simulated conditions:

The mode of the single carrier emulation that emulation is adopted.Here the quantification code book of Cai Yonging is for quantizing code book (RVQ) at random, the CDI of each user feedback 4 bit and an analog quantity CQI.Each user has 2 reception antennas, always has 20 users and waits for scheduling.The base station end has 4 transmitting antennas, simultaneously send data to 4 users, each user receives only a data flow, and user's selection algorithm of maximized system capacity is adopted in the base station, the scheme that each user's transmitted power adopts constant power to distribute, channel are independent identically distributed multiple Gaussian channel.

3) simulation result:

Accompanying drawing 2 has provided the MU-MIMO power system capacity of different antennae merge algorithm.Simulation result shows that improved antenna merge algorithm all provides more excellent systematic function with respect to the antenna merge algorithm based on main characteristic vector and minimum quantization error under all signal to noise ratio conditions.And along with the increase of signal to noise ratio, it is increasing that the gap of their performances becomes.Accompanying drawing 3 has provided the MU-MIMO power system capacity CDF of different antennae merge algorithm when signal to noise ratio equals 20dB.

Comprehensive above-mentioned simulation result and analysis, if need the MU-MIMO system is merged the MU-MISO system that is converted into by antenna under the prerequisite of Limited Feedback, so improved antenna merge algorithm can obtain more excellent systematic function than present other merge algorithms of studying.

Claims (2)

1. the improved CSI limited feedback method that proposes of the present invention, the signal detection algorithm and the CDI that are based on receiving terminal select gang to design, in the hope of can be with E[SINR _k] maximization.

2. the present invention includes following steps:

Step 1: the CSI-RS that the user utilizes the base station to send carries out channel estimating; For k user, through u _kCan represent by following formula as its received signal after the input, wherein

y _kBe the signal that user k receives, n ' _kBe the white complex gaussian noise that is superposeed on the equivalent channel, G _iBe the precoding vector of i user M * 1:

$r_k=u_k{y}_k=u_k{H}_k{G}_k{d}_k+u_k{H}_k\underset{i\&Element;S,i\&NotEqual;k}{\mathrm{\&Sigma;}}{G}_i{d}_i+n_k^{\&prime;}$

Step 2: the user is by quantizing directional information CDI and the corresponding channel quality information CQI that code book feeds back its equivalent channel:

(2a) merge algorithm:

The SINR that the selected user k that transmits data is obtained _kCan be expressed as:

${\mathrm{SINR}}_k=\frac{p_k{\left|u_k{H}_k{w}_k\right|}^2}{1+\underset{i\&Element;S\backslash \left\{k\right\}}{\mathrm{\&Sigma;}}{P}_i{\left|u_k{H}_k{w}_i\right|}^2}$

W wherein _kAnd p _kBe respectively k user's precoding vector and transmitted power, supposing the system carries out the constant power distribution to each user here, so

The selection that improved antenna merge algorithm is intended to the input of associating receiving terminal and CDI makes each user's E[SINR _k] maximization.

(2b) calculating of the selection of channel directional information (CDI) and channel quality information (CQI):

K user's CQI can be expressed as:

${\mathrm{CQI}}_k=\frac{\frac{P}{M}{\left|u_k{H}_k{v}_k\right|}^2}{1+\frac{P}{M}{\left|\right|u_k{H}_k-\left(u_k{H}_k{v}_k\right)v_k^H\left|\right|}_2^2}$

If ${\mathrm{CQI}}_k=\frac{u_k{A}_k{u}_k^H}{1+u_k{B}_k{u}_k^H},A_k=\frac{P}{M}\left(H_k{c}_i{c}_i^H{H}_k^H\right),B_k=\frac{P}{M}\left[H_k(I-c_i{c}_i^H)H_k^H\right]$

At c _iDuring as the CDI of equivalent channel, select the MMSE signal detection algorithm as maximization E[SINR _k] the receiving terminal merge algorithm; Wherein, the weighting of devise optimum merges vectorial u under fixing quantization code _k:

$u_k={\left({(I+B_k)}^{-1}\sqrt{\frac{P}{M}}{H}_k{c}_i\right)}^H$

(2c) its corresponding weighting merging vector of each quantification vector sum will produce a CQI like this, and the user therefrom selects the channel quality information of the CQI of a maximum as its feedback, feeds back the pairing c of this CQI simultaneously _iCDI as its local channel.The base station end utilizes the CDI of each user feedback and CQI carries out user's scheduling and pre-coding matrix calculates, and the receiving end signal detection algorithm here adopts MMSE.

Images