CN104618080B - Channel calibration method for extensive multiple-input and multiple-output tdd systems - Google Patents

Abstract

The invention relates to a channel calibration method for a large-scale multiple-input multiple-output time division duplex system, which belongs to the technical field of wireless communication relay channel calibration. The method includes: dividing the large-scale multiple-input multiple-output system into three subsystems. The subsystems include a single-input single-output system and two single-input multiple-output systems; the uplink and downlink channels of the three subsystems are respectively calibrated to obtain channel calibration coefficients. Then the calibration coefficients of the whole MIMO system are obtained according to the channel calibration coefficients of the three subsystems, and the relatively accurate downlink channel state information of the whole system is obtained according to the uplink channel state information. The method can reduce the feedback of the channel state information, improve the system capacity compared with the general traditional method, and reduce the power consumed by the system to send the pilot frequency.

Description

Channel Alignment Method for Massive MIMO Time Division Duplex System

technical field

The invention belongs to the technical field of wireless communication relay channel calibration, in particular to a TDD-LS-MIMO multi-antenna base station to multi-antenna relay downlink channel calibration method in the wireless communication field

Background technique

With the continuous increase of mobile terminals, the traffic volume of downlink data is also increasing. The multiple-input multiple-output (MIMO) system can increase the channel capacity, which has attracted extensive attention from the industry and academia. MIMO has two duplex modes, one is Frequency Division Duplex (FDD) and the other is Time Division Duplex (TDD). Frequency Division Duplex (FDD) is a duplex mode that is widely used now. The main feature is that the uplink and downlink channels work in different frequency bands. The pilot overhead is too large when estimated. In the time division duplex (TDD) mode, because both the uplink and downlink channels work in the same frequency band, there is reciprocity between the uplink and downlink channels, and the downlink channel state information can be estimated based on the uplink channel state information. However, due to the imperfection of uplink and downlink transmission and receiving devices, the channel reciprocity is destroyed, and the uplink channel status information cannot be accurately estimated by using the uplink channel status information. Therefore, it is necessary to calibrate the uplink and downlink channels to satisfy reciprocity. The traditional channel calibration technology is to obtain uplink and downlink channel state information separately, and then perform calibration to obtain the relationship between uplink and downlink channels. The efficiency is low, and a large amount of downlink channel state information needs to be fed back to the base station.

Contents of the invention

The purpose of the present invention is to promote the application of time-division duplex large-scale multiple-input multiple-output technology in the field of wireless communication, and propose a channel calibration method for large-scale multiple-input multiple-output time-division duplex systems, which can be more accurate To calibrate the uplink and downlink channels of a time-division duplex massive MIMO system, the accurate information of the downlink channel can be obtained relatively simply through the principle of reciprocity.

A channel calibration method for a large-scale MIMO time-division duplex system proposed by the present invention is characterized in that the method includes: dividing the large-scale MIMO system into three subsystems, and the three subsystems include a Single Input Single Output System (SISO), Two Single Input Multiple Output System (SIMO). Calibrate the uplink and downlink channels of the three subsystems respectively to obtain the channel calibration coefficients. Then, according to the channel calibration coefficients of the three subsystems, the calibration coefficients of the whole MIMO system are obtained, and then the relatively accurate downlink channel state information of the whole system is obtained according to the uplink channel state information.

The present invention specifically comprises the following steps:

1) Determine the reference channel, refer to the antenna, and divide the subsystems: divide the multiple-input multiple-output system a of the large-scale base station and the user into three subsystems, the user B sends the uplink pilot signal, and the base station A sends the pilot according to the B-side The signal estimates the uplink channel state information, selects a channel with the highest channel gain as the reference channel, and the antennas at both ends of the reference channel as the reference antenna, assuming that there are M antennas at the A end, which are A ₁ ~ A _M , and the reference antenna selected by base station A A ₁ , A ₂ ～A _M , A _i ∈ C (i=2,3,...M) represent other antennas of A; base station A sends information including reference antennas to base station B, and B has N antennas are B ₁ ～B _N , then base station B selects the reference antenna as B ₁ according to the information sent by base station A, and B ₂ ～B _N (B _j ∈ D(j=2,3,..,N)) means that B Antenna A ₁ and antenna B ₁ form a single-input single-output sub-SISO system b, reference antenna A ₁ and antennas A ₂ to A _M form a single-input multiple-output sub-SIMO system c, antenna B ₁ and antenna B ₂ ～B _N form another single-input multiple-output sub-SIMO system d;

2) Estimating the channel calibration coefficient β _b of system b: estimating the uplink and downlink channel state information of SISO system b, according to formula (3): G _b = H _b · β _b (4),

in, . Estimate G _b , H _b multiple times to generate a total least squares problem, and solve the least squares problem to find the calibration coefficient β _b ;

3) Estimate the channel calibration coefficient β _c of system b: SIMO system c composed of A ₁ and A ₂ ～A _M , obtain the uplink and downlink channel state information G _c , H _c of SIMO system c, according to formula (3):

G _c =β _c H _c (5)

in, Estimate G _c , H _c multiple times, generate M-1 total least squares problems, and solve the least squares problems, then find the calibration coefficient β _c : diag(β _c ) is the main diagonal element of β _c ;

4) Estimate the channel calibration coefficient β _d of system b: obtain the SIMO system d composed of B ₁ , B ₂ ～B _N , obtain the uplink and downlink channel state information G _d , H _d of the SIMO system d, according to the formula (3) Can get G _d ＝β _d ·H _d (6)

in Estimate G _d , H _d multiple times, generate N-1 total least squares problems, and solve the least squares problem, then find the calibration coefficient β _d : diag(β _d ) is the main diagonal element of β _d ;

5) Calculate the calibration coefficient β _a of the large-scale MIMO system a as:

Through steps 2), 3), and 4), it can be obtained that β _b , β _c , and β _c contain all the information of β _a , so β _a is represented by β _b , β _c , and β _c as follows:

6) Estimate the downlink channel state information of the massive MIMO system a according to formula (8):

[G] _mn = [H] _mn [β _a ] _mn (8)

[G] _mn is the downlink channel state information of the mnth channel, [H] _mn is the uplink channel state information of the mnth channel,

[β _a ] _mn is the channel calibration coefficient of the mnth channel, and the downlink channel state information is estimated according to the uplink channel information.

The method can reduce the feedback of the channel state information, improve the system capacity compared with the general traditional method, and reduce the power consumed by the system to send the pilot frequency.

Description of drawings

Fig. 1 is a schematic diagram of the system model of the present invention.

Fig. 2 is a schematic diagram of a system equivalent circuit of the present invention.

Fig. 3 is a flow chart of the method of the present invention.

detailed description

The channel calibration method for the large-scale MIMO time division duplex system proposed by the present invention, the structural drawings and the embodiments are further described as follows:

The present invention divides the MIMO system of large-scale base stations and users into three subsystems, as shown in Figure 1: the three subsystems divided into the large-scale multiple-input multiple-output system a (MIMO) include a single-input single-output system b ( SISO), two single-input multiple-output systems c and d (SIMO). The sending end is represented by base station A, and the receiving end is represented by user B. There are M antennas at terminal A, which are A ₁ ～A _M , A ₁ is the selected reference antenna, and A ₂ ～A _M (A _i ∈ C(i=2,3,...M)) represent other antennas of A antenna. There are N antennas at the B terminal, respectively B ₁ ~B _N , B ₁ is the selected reference antenna, B ₂ ~B _N (B _j ∈ D(j=2,3,..,N)) represents the other antennas of B antenna. Antennas A ₁ ～A _M and B ₁ ～B _N form the entire MIMO system. Reference antenna A ₁ and reference antenna B ₁ form a single-input single-output sub-SISO system, denoted by system b. Reference antenna A ₁ and antennas A ₂ ～A One of the single-input multiple-output sub-SIMO systems composed of _M is denoted by system c, and the other single-input multiple-output sub-SIMO system is _denoted by antenna B1 and antennas B2 _{- BN} .

Principle of the present invention:

The equivalent circuit of the above system is shown in Figure 2: T _A is the equivalent filter of the transmitting circuit at the A terminal, T _B is the equivalent filter of the transmitting circuit at the B terminal, and R _A is the equivalent filter of the receiving circuit at the A terminal , RB is the equivalent filter of the receiving circuit at the _B end, and C is an ideal electromagnetic channel.

Suppose: the downlink channel state information is: G=R _B C T _A (1),

The uplink channel state information is: H= _RA · _C ^T ·TB (2).

Due to the imperfection of _TA , TB, _RA , and _RB , the uplink channel state information H and the downlink channel state information _G are not equal, that is, G≠H. In order to use H to estimate G, the channel needs to be calibrated. which is

Elimination of C in formula (1), (2) can get: G=P _B H P _A (3),

in And without considering the crosstalk between the antennas, T _A , T _B , R _A , and R _B are diagonal matrices.

The present invention is used for the channel calibration method of large-scale multiple-input multiple-output time division duplex system, as shown in Figure 3, it is characterized in that, comprises the following specific steps:

1) Determine the reference channel, reference antenna, and divide the subsystem: User B sends an uplink pilot signal, and base station A estimates the uplink channel state information according to the pilot signal sent by B, and selects a channel with the highest channel gain as the reference channel, and the reference channel The antennas at both ends are used as reference antennas. Suppose the reference antenna selected by base station A is A ₁ . Base station A sends information containing the reference antenna to base station B. Then base station B selects the reference antenna as B ₁ according to the information sent by base station A. Antenna A ₁ and antenna B ₁ form a single-input single-output sub-SISO system b;

2) Estimating the channel calibration coefficient β _b of system b: estimating the uplink and downlink channel state information of SISO system b, according to formula (3): G _b = H _b · β _b (4),

in, . Estimate G _b , H _b multiple times, then generate a total least squares problem, solve the least squares problem, then you can find the calibration coefficient β _b ;

3) Estimate the channel calibration coefficient β _c of subsystem c: SIMO system c composed of A ₁ and A ₂ ~ A _M , obtain the uplink and downlink channel state information G _c , H _c of SIMO system c, according to formula (3) can be obtained : G _c =β _c ·H _c (5)

in, (Because there is no crosstalk problem between antennas, PC is also a diagonal matrix, and system _c is decomposed into M-1 SISO systems) Estimates of G _c and H _c multiple times can generate M-1 total least squares problems , to solve the least squares problem, the calibration coefficient β _c can be obtained, and diag(β _c ) is the main diagonal element of β _c , then

4) Estimate the channel calibration coefficient β _d of subsystem d: obtain the SIMO system d composed of B ₁ , B ₂ ～B _N , and obtain the uplink and downlink channel state information G _d , H _d of the SIMO system d, according to the formula (3 ) can get G _d ＝β _d ·H _d (6)

in (Because there is no crosstalk problem between antennas, P _D is also a diagonal matrix, so system d can be decomposed into N-1 SISO systems) Estimates G _d , H _d multiple times can generate N-1 overall least squares The multiplication problem and the least squares problem can be solved to obtain the calibration coefficient β _d . diag(β _d ) is the main diagonal element of β _d ,

5) Calculate the calibration coefficient β _a of the whole system a as:

Through steps 2), 3), and 4), it can be obtained that β _b , β _c , and β _c contain all the information of β _a , so β _a can be expressed by β _b , β _c , and β _c as follows:

6) Estimate the downlink channel state information of system a according to formula (8):

[G] _mn = [H] _mn [β _a ] _mn (8)

[G] _mn is the downlink channel state information of the mnth channel, [H] _mn is the uplink channel state information of the mnth channel, [β _a ] _mn is the channel calibration coefficient of the mnth channel, which can be based on the uplink channel information Estimate the downlink channel state information, (because the coherence time of the antenna transmitting circuit and the receiving circuit is longer than that of the electromagnetic channel, it can be calibrated multiple times in the coherence time of the antenna transmitting circuit and the receiving circuit, which can be estimated according to formula (8) Multiple downlink channel state information.

Example

In the specific embodiment of the method of the present invention, make A-end antenna M=128, B-end antenna N=16;

The specific steps of the present embodiment method are as follows:

1) User B sends the uplink pilot signal, base station A estimates the uplink channel state information according to the pilot signal sent by B, selects the channel with the highest channel gain as the reference channel, and the antennas at both ends of the reference channel as the reference antenna, then base station A can Select the reference antenna A ₁ , base station A sends the information containing the reference antenna to the base station B, then the base station B can select the reference antenna B ₁ according to the information sent by the base station A, the system b composed of the antenna A ₁ and the antenna B ₁ .

2. Estimate SISO system b uplink and downlink channel state information G _b , H _b , according to formula (3): G _b = H _b · β _b (4),

in, . Estimate G _b , H _b multiple times, then a total least squares problem can be generated, and the channel calibration coefficient β _b of system b can be obtained by solving the least squares problem;

3) The SIMO system c composed of A ₁ and A ₂ ～ A ₁₂₈ can obtain the uplink and downlink channel state information G _c , H _c of SIMO system c. According to the formula (3), it can be obtained: G _c =β _c H _c (5 )

in, (Because there is no crosstalk problem between antennas, PC is also a diagonal matrix, and system _c is decomposed into 127 SISO systems.) Multiple estimations of G _c and H _c can generate 127 overall least squares problems, and solve least squares multiplication problem, the channel calibration coefficient β _c of system b can be obtained, diag(β _c ) is the main diagonal element of β _c , then

4) Obtain the SIMO system d composed of B ₁ , B ₂ ～B _N , and obtain the uplink and downlink channel state information G _d , H _d of the SIMO system d. According to formula (3), G _d =β _d ·H _d can be obtained (6)

in (Because there is no crosstalk problem between antennas, P _D is also a diagonal matrix, so system d can be decomposed into 15 SISO systems) Multiple estimations of G _d , H _d can generate 15 overall least squares problems, and solve By using the least square problem, the channel calibration coefficient β _d of system b can be obtained. diag(β _d ) is the main diagonal element of β _d ,

5) Calculate the calibration coefficient of system a as:

Through steps 2, 3, and 4, it can be obtained that β _b , β _c , and β _c contain all the information of β _a , so β _a can be expressed by β _b , β _c , and β _c as follows:

6) Based on [G] _mn = [H] _mn · [β _a ] _mn (8)

[G] _mn is the downlink channel state information of the mnth channel, [H] _mn is the uplink channel state information of the mnth channel, [β _a ] _mn is the channel calibration coefficient of the mnth channel, which can be based on the uplink channel information Estimate the downlink channel state information, (since the coherence time of the antenna transmitting circuit and the receiving circuit is longer than the coherence time of the electromagnetic channel, it can be calibrated multiple times in the coherence time of the antenna transmitting circuit and the receiving circuit, and the multiple channels can be estimated according to formula (8) Secondary downlink channel state information.

Claims (1)

1. A channel calibration method for a large-scale multiple-input multiple-output time division duplex system is characterized by comprising the steps of dividing the large-scale multiple-input multiple-output system into three subsystems, wherein the three subsystems comprise a single-input single-output system and two single-input multiple-output systems; calibrating uplink and downlink channels of the three subsystems respectively to obtain channel calibration coefficients; then, obtaining the calibration coefficient of the whole multi-input multi-output system according to the channel calibration coefficients of the three subsystems, and obtaining the accurate downlink channel state information of the whole system according to the uplink channel state information;

the method specifically comprises the following steps:

1) determining a reference channel, a reference antenna, and dividing the subsystem: dividing a multi-input multi-output system a of a large-scale base station and a user into three subsystems, wherein a user B end sends an uplink pilot signal, a base station A end estimates uplink channel state information according to the pilot signal sent by the B end, one channel with the highest channel gain is selected as a reference channel, antennas at two ends of the reference channel are used as reference antennas, and the A end is provided with M antennas which are respectively A₁～A_MThe reference antenna selected by the base station A is A₁，A₂～A_M，A_i∈ C, i is 2,3, M represents other antennas of the base station A, C represents a set formed by other antennas of the base station A, the base station A sends information containing reference antennas to the base station B, and N antennas at the B end are respectively B₁～B_NThen the base station B selects the reference antenna as B according to the information sent by the base station A₁，B₂～B_N，B_j∈ D, j 2,3, N, for other antennas of base station B, D for the set of other antennas of base station B, antenna a₁And an antenna B₁Form a single-input single-output SISO system b, a reference antenna A₁And an antenna A₂～A_MForming a single input multiple output sub SIMO system c and a reference antenna B₁And an antenna B₂～B_NForming another single-input multi-output sub SIMO system d;

2) estimating β channel calibration coefficients for system b_b: estimating uplink and downlink channel state information of SISO system b according to formula G ═ P_B·H·P_A(3) Wherein G is downlink channel state information, H is uplink channel state information, T_Aequivalent filter for A-terminal transmission circuit, T_BEquivalent filter for B-terminal transmission circuit, R_AEquivalent filters for A-terminal receiving circuits, R_BIs terminated by BAn equivalent filter of the receiving circuit, and T without considering the problem of crosstalk between the antennas_A，T_B，R_A，R_BIs a diagonal matrix;

obtaining: g_b＝H_b·β_b(4)，

Wherein G is_b,H_bFor estimating the downlink and uplink channel state information of SISO system b,multiple estimation of G_b,H_bGenerating a total least squares problem, solving the least squares problem, and finding the calibration coefficients β_b；

3) Estimating β channel calibration coefficients for system c_c：A₁And A₂～A_MSIMO system c composed of obtaining downlink and uplink channel state information G of SIMO system c_c,H_cAccording to the formula G ═ P_B·H·P_A(3) Obtaining:

G_c＝β_c·H_c(5)

wherein,multiple estimation of G_c,H_cGenerating M-1 total least squares problems, solving the least squares problem, and finding the calibration coefficients β_c：diag(β_c) Is β_cA main diagonal element;

4) estimating β channel calibration coefficients for system d_d: obtained by B₁，B₂～B_NComposed SIMO system d, obtaining the status information G of the downlink and uplink channels of the SIMO system d_d,H_dG is obtained from the formula (3)_d＝β_d·H_d(6)，

WhereinMultiple estimation of G_d,H_dGenerating N-1 total least squares, solving the least squares problem, and finding the calibration coefficients β_d：diag(β_d) Is β_dA main diagonal element;

5) calculating β calibration coefficients for massive MIMO system a_aComprises the following steps:

from steps 2),3),4) it is possible to derive β_b,β_c,β_cContaining β_aAll of the information of (1), therefore β_aBy β_b,β_c,β_cIs represented as follows:

<mrow> <msub> <mrow> <mo>&amp;lsqb;</mo> <msub> <mi>&amp;beta;</mi> <mi>a</mi> </msub> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mi>m</mi> <mi>n</mi> </mrow> </msub> <mo>=</mo> <mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mrow> <mfrac> <mrow> <msub> <mi>&amp;beta;</mi> <mi>b</mi> </msub> <mo>&amp;CenterDot;</mo> <msub> <mrow> <mo>&amp;lsqb;</mo> <mi>d</mi> <mi>i</mi> <mi>a</mi> <mi>g</mi> <mrow> <mo>(</mo> <msub> <mi>&amp;beta;</mi> <mi>d</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> </mrow> <msub> <mrow> <mo>&amp;lsqb;</mo> <mi>d</mi> <mi>i</mi> <mi>a</mi> <mi>g</mi> <mrow> <mo>(</mo> <msub> <mi>&amp;beta;</mi> <mi>c</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mi>m</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> </mfrac> <mo>,</mo> <mn>2</mn> <mo>&amp;le;</mo> <mi>m</mi> <mo>&amp;le;</mo> <mi>M</mi> <mo>,</mo> <mn>2</mn> <mo>&amp;le;</mo> <mi>n</mi> <mo>&amp;le;</mo> <mi>N</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mfrac> <msub> <mi>&amp;beta;</mi> <mi>b</mi> </msub> <msub> <mrow> <mo>&amp;lsqb;</mo> <mi>d</mi> <mi>i</mi> <mi>a</mi> <mi>g</mi> <mrow> <mo>(</mo> <msub> <mi>&amp;beta;</mi> <mi>c</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mi>m</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> </mfrac> <mo>,</mo> <mn>2</mn> <mo>&amp;le;</mo> <mi>m</mi> <mo>&amp;le;</mo> <mi>M</mi> <mo>,</mo> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>&amp;beta;</mi> <mi>b</mi> </msub> <mo>&amp;CenterDot;</mo> <msub> <mrow> <mo>&amp;lsqb;</mo> <mi>d</mi> <mi>i</mi> <mi>a</mi> <mi>g</mi> <mrow> <mo>(</mo> <msub> <mi>&amp;beta;</mi> <mi>d</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> <mo>,</mo> <mi>m</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>&amp;le;</mo> <mi>n</mi> <mo>&amp;le;</mo> <mi>N</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>&amp;beta;</mi> <mi>b</mi> </msub> <mo>,</mo> <mi>m</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>

6) estimating downlink channel state information of the massive multiple-input multiple-output system a according to formula (8):

[G]_mn＝[H]_mn·[β_a]_mn(8)

[G]_mnfor downlink channel state information of the nth channel, [ H ]]_mnFor the uplink channel state information of the mn-th channel,

[β_a]_mnand if the channel calibration coefficient is the channel calibration coefficient of the nth channel, estimating the state information of the downlink channel according to the uplink channel information.