CN106301502B - Extensive MIMO downlink wireless communications method under radio circuit mismatch condition - Google Patents

Abstract

The present invention proposes a massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch, which is suitable for a TDD massive MIMO wireless communication system. The base station intermittently obtains the radio frequency circuit gain coefficients on the base station side and the user side. Each user transmits its own pilot signal, and the base station obtains the downlink channel estimation of each user and the statistical characteristics of the downlink channel estimation error according to the radio frequency circuit gain coefficient. The base station performs downlink robust precoding transmission according to the downlink channel estimation and the statistical characteristics of the downlink channel estimation error. The present invention can obtain downlink channel estimation in the presence of radio frequency circuit mismatch, reduce the loss of downlink precoding transmission performance caused by unsatisfied uplink and downlink channel reciprocity, and implement robust downlink precoding against downlink channel estimation errors accordingly. The transmission can further effectively improve the spectral efficiency of the wireless communication system, and the base station can also predict the downlink traversal rate of the user according to the radio frequency circuit gain coefficient and statistical channel information.

Description

Massive MIMO downlink wireless communication method in the case of RF circuit mismatch

technical field

The present invention relates to a MIMO wireless communication system using multiple antennas, and in particular, to a massive MIMO wireless communication system used in the presence of radio frequency circuit mismatch.

Background technique

With the popular application of intelligent mobile terminals and the continuous development of new mobile service requirements, the demand for mobile communication transmission rates continues to grow exponentially. In order to meet the needs of future mobile communication applications, it is necessary to deeply mine and utilize spatial wireless resources to greatly improve the spectrum utilization and power utilization of wireless communication. The MIMO wireless transmission technology using multi-antenna transmission and multi-antenna reception is the basic technology to improve the spectrum and power efficiency of wireless communication, and has been one of the mainstream technologies in the field of wireless communication research in the past ten years. Limited by the number of antennas (for example, in the LTE-A standard of 3GPP, a maximum of 8 antennas can be configured on the base station side), the spectrum and power efficiency of traditional MIMO technology are still low. It has become one of the development trends of future wireless communication to configure a large-scale antenna array (more than tens of antennas) on the base station side to deeply mine and utilize the spatial dimension resources.

In a time division duplex (TDD) massive MIMO wireless communication system, the uplink signal and the downlink signal share the same frequency band, so the propagation paths of the uplink and downlink signals are the same, and the uplink and downlink channels are reciprocal. However, in the actual system, each antenna RF front-end unit on the base station side and the user side has two sets of circuits to complete the signal transmission and reception respectively, as shown in Figure 1. Due to factors such as hardware process errors and different working environments of the radio frequency unit, the gain coefficient of the radio frequency circuit is different, resulting in a mismatch of the radio frequency circuit, which damages the channel reciprocity. The degree of mismatch is particularly severe for massive antenna arrays in massive MIMO systems. If the uplink channel estimation is directly used for downlink precoding transmission, the system performance will be obviously lost. The present invention provides a method for massive MIMO downlink wireless communication in the presence of radio frequency circuit mismatch, including the acquisition of radio frequency circuit gain coefficients on the base station side and the user side, and the use of radio frequency circuit gain coefficients to obtain downlink channel estimation and downlink channel estimation errors statistics, implement robust downlink precoding transmission against downlink channel estimation errors, and ergodic achievable rate prediction.

SUMMARY OF THE INVENTION

Technical problem: The purpose of the present invention is to provide a massive MIMO downlink wireless communication method in the presence of radio frequency circuit mismatch. The basic feature of the method is to obtain downlink channel estimation and estimation error statistical information based on RF circuit gain coefficient estimation, and implement robust downlink precoding transmission for estimation errors to improve system spectral efficiency.

Technical scheme: In order to realize the above-mentioned purpose of the invention, the technical scheme adopted in the present invention is:

A massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch, suitable for a time division duplex (TDD) massive MIMO wireless communication system, a base station side is equipped with a massive array antenna, the number of antennas is more than dozens, and the base station is in the Perform wireless communication with multiple users on the same time-frequency resource; the method specifically includes:

(1) RF circuit gain coefficient acquisition: by setting a gain coefficient measurement unit in the RF circuit at the front end of the transceiver antenna, the measurement unit sends a reference signal to the RF circuit and receives the signal feedback to estimate the absolute RF circuit gain coefficient on the base station side and the user side; or , obtain two-way channel estimation by sending and receiving reference signals between the reference antenna on the base station side and other antennas, and then obtain the relative RF circuit gain coefficient on the base station side, and obtain the uplink and downlink channel parameters by sending and receiving reference signals between the reference antenna and the user, and then obtain the relative frequency of the user side. RF circuit gain factor;

(2) User downlink channel estimation: each user sends their respective uplink pilot signals on the same time-frequency resource, and the base station performs downlink of each user according to the received pilot signal and the acquired radio frequency circuit gain coefficients on the base station side and the user side. Channel estimation, and obtain estimation error statistical information; wherein the downlink channel estimation is calculated and obtained based on the absolute RF circuit gain coefficient, or is calculated and obtained based on the relative RF circuit gain coefficient combined with the uplink channel estimation;

(3) Downlink robust precoding data transmission: In the downlink data transmission stage, the base station uses the acquired downlink channel estimation and the statistical information of the downlink channel estimation error to obtain the robust precoding matrix required to transmit data to each user signal, and implement the Robust precoding sends data signals to each user simultaneously on the same time-frequency resource.

In the TDD massive MIMO wireless communication system, the antenna array on the base station side includes more than ten antenna units, and the spacing between the antenna units is smaller than the wavelength of the carrier. For sector antennas, the spacing between antennas is 1/2 wavelength or wavelength or 1 wavelength; each antenna unit can use a single-polarized or multi-polarized antenna; the communication process includes three stages of radio frequency circuit gain coefficient acquisition on the base station side and user side, downlink channel estimation, and downlink robust precoding data transmission.

The absolute RF circuit gain coefficient in step (1) is calculated by the following formula:

in, sending the reference signal power for the gain factor measurement unit, represents the reference signal sent by the gain coefficient measurement unit to the RF circuit module at the front end of the antenna m in the t-th cycle, receiving a reference signal for the gain factor measurement unit, Indicates the estimated value of the gain coefficient of the radio frequency circuit at the front end of the transmit antenna on the base station side or the user side or the front end of the receive antenna.

The relative RF circuit gain coefficient on the base station side in step (1) is calculated by the following formula:

in, is the estimation of the relative RF circuit gain coefficient of the mth antenna on the base station side, is the channel estimation from the mth antenna on the base station side to the reference antenna, is the channel estimate from the reference antenna to the mth antenna.

The relative gain coefficient of the user side radio frequency circuit in step (1) is obtained by the following formula:

in, is an estimate of the relative RF circuit gain factor for user k, represents the downlink channel estimate from the reference antenna to user k, represents the uplink channel estimate from user k to the reference antenna.

In step (2), the downlink channel estimation based on the absolute radio frequency circuit gain coefficient is specifically: based on the base station side receiving antenna absolute radio frequency circuit gain coefficient estimation Transmitting Antenna Absolute RF Circuit Gain Coefficient Estimation User k receiving antenna absolute RF circuit gain factor Transmitting antenna absolute RF circuit gain factor And the pilot signal Y ^tr sent by the user received by the base station obtains the downlink channel estimate by any one of the following three methods:

Method 1: First obtain the partial estimate of the wireless channel of user k by the following formula:

Then the downlink channel estimate is obtained by the following formula:

in, is the wireless channel estimation for user k, is the covariance matrix of the wireless channel of user k, ρ ^tr is the pilot training signal-to-noise ratio, is the orthogonal pilot sequence sent by user k, transmit pilot sequence power for user k, where K is the number of users;

Method 2: First obtain the uplink channel estimate of user k by the following formula:

Then the downlink channel estimate is obtained by the following formula:

In the formula, is the covariance matrix of the uplink channel of user j;

Method 3: Obtain the downlink channel estimate of user k directly from the following formula

In the formula, is the covariance matrix of the downlink channel of user j;

The covariance matrix of the downlink channel estimation error is calculated as follows:

In step (2), the downlink channel estimation based on the relative radio frequency circuit gain coefficient is specifically: estimating based on the relative radio frequency circuit gain coefficient at the base station side User k relative RF circuit gain factor estimation And the uplink channel estimation obtains the downlink channel estimate, and the downlink channel estimate of user k is calculated according to the following formula:

in is the uplink channel estimation for user k, which is calculated by the following formula:

The downlink channel estimation error covariance matrix is obtained by the following formula:

in is the estimation error of the uplink channel for user k, which is calculated by the following formula:

The robust precoding matrix in step (3) is calculated by:

in, is the downlink channel estimation, ρ ^dl is the average transmit signal-to-noise ratio of each user’s downlink transmission, γ is the transmit power constraint parameter on the base station side, which is calculated by the following formula:

where tr{.} represents the matrix trace operation.

Beneficial effects: The massive MIMO downlink wireless communication method provided by the present invention has the following advantages under the condition of radio frequency circuit mismatch:

1. When the channel reciprocity is not satisfied due to the mismatch of the radio frequency circuits, obtain the downlink channel estimation and the statistical characteristics of the estimation error to ensure the downlink precoding transmission performance.

2. The channel estimation error is considered during downlink data transmission, which improves the robustness and efficiency of system data transmission.

3. The base station can predict the downlink traversal rate of the user based on the absolute RF circuit gain coefficient.

Description of drawings

FIG. 1 is a schematic diagram of the antenna front-end transceiver system on the base station side and the user side.

FIG. 2 is a schematic diagram of a gain measurement unit of a radio frequency circuit.

Detailed ways

In order for those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only These are some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

1. System configuration and communication process

Figure 1 is a schematic diagram of the configuration of a large-scale antenna array on the base station side. Considering the case of a single-cell base station, the base station side is configured with an antenna array containing more than tens of antenna elements. other array structures. Each antenna unit can use an omnidirectional antenna or a sector antenna. When each antenna unit uses an omnidirectional antenna, a 120-degree sector antenna, and a 60-degree sector antenna, the spacing between the antennas can be configured as 1/2 wavelength, wavelength and 1 wavelength. Each antenna unit can use a single-polarized or multi-polarized antenna.

In this embodiment, only the narrowband channel is considered, there is only a single composite path in the considered narrowband channel, and the considered narrowband channel can be regarded as a single subcarrier channel in the conventional wideband OFDM system. Consider the time division duplex (TDD) transmission mode, and assume that the number of antennas equipped on the base station side is M, the number of users is K, and each user is equipped with a single antenna.

The communication process based on the above massive MIMO system includes the following three stages:

i. Obtaining the gain coefficient of the radio frequency circuit on the base station side and the user side: The absolute gain coefficient of the radio frequency circuit can be obtained by adding a measuring device to the radio frequency circuit at the front end of the transceiver antenna. The model is shown in Figure 2. The measuring device sends a reference signal to the radio frequency circuit and receives the signal. Feedback to estimate the gain coefficient of the radio frequency circuit. When the user accesses the base station, the user sends the absolute gain coefficient of the radio frequency circuit at the user side to the base station through a special feedback link; the base station side can select a reference antenna, and transmit and receive with other antennas through the reference antenna. The reference signal obtains two-way channel estimation. The base station obtains the relative gain coefficient of the radio frequency circuit on the base station side according to the channel estimation between the reference antenna and other antennas. When the user accesses the base station, the base station refers to the reference signal sent and received between the antenna and the user to obtain the reference antenna and the user. The uplink and downlink channel parameters between users, and then obtain the relative gain coefficient of the radio frequency circuit on the user side.

ii. User downlink channel estimation: each user sends an uplink pilot signal, and the base station side uses the received pilot signal, combined with the acquired RF circuit gain coefficient, to calculate and obtain downlink channel estimation and downlink channel estimation error statistical information.

iii. Downlink robust precoding data transmission: The base station uses the downlink channel parameter estimation and the statistical information of the estimation error to obtain the robust precoding matrix required to transmit data to each user signal, thereby generating a downlink transmission signal, which is sent by the base station to each user. At the same time, each user performs reception processing according to the received signal, and obtains the downlink transmission bit data stream. In addition, the base station can also predict the user traversal rate based on the absolute RF circuit gain coefficient.

2. Obtaining the gain coefficient of the radio frequency circuit on the base station side and the user side:

(1) Obtaining the absolute gain coefficient

The gain coefficient of the radio frequency circuit can be obtained by adding a measuring device to the radio frequency circuit at the front end of the transceiver antenna. The model is shown in Figure 2. The measuring device estimates the gain coefficient of the radio frequency circuit by sending a reference signal to the radio frequency circuit and receiving the signal feedback. When the user accesses the base station, The user sends the gain coefficient of the radio frequency circuit on the user side to the base station through a special feedback link.

The radio frequency gain coefficient measuring unit intermittently sends a reference signal to the radio frequency module of the antenna front end, and estimates the radio frequency circuit gain coefficient according to the received feedback signal of the radio frequency module. by represents the reference signal sent by the RF gain coefficient measurement unit to the RF circuit module at the front end of the antenna m in the t-th cycle, Indicates that the RF gain coefficient measurement unit receives the feedback signal from the RF circuit module at the front end of the antenna m in the t-th cycle, then the RF circuit gain coefficient measurement unit estimates the RF circuit gain coefficient according to the following formula:

in, is the reference signal power, Indicates the estimated value of the gain coefficient of the front-end RF circuit of the transmitting antenna or the receiving antenna.

(2) Relative gain coefficient acquisition

The base station intermittently sends detection signals through the internal antenna to obtain the relative gain coefficient of the radio frequency circuit on the base station side. One antenna needs to be selected as the reference antenna, or a separate additional antenna needs to be set up as the reference antenna.

Assuming that the i-th antenna on the base station side is selected as the reference antenna, represents the reference signal sent by the reference antenna to other antennas in the t-th cycle, represents the reference signal received by the mth antenna in the tth cycle, represents the reference signal sent by other antennas to the reference antenna in turn in the t-th cycle, represents the reference signal sent by the mth antenna received by the reference antenna in t cycles, h _0→m represents the channel parameter from the reference antenna to the mth antenna, where h _0→m =r _bm v _m t _b0 , r _bm represents the gain coefficient of the radio frequency circuit in the receiving channel of the mth antenna, v _m represents the wireless channel parameter from the reference antenna to the mth antenna, and t _b0 represents the gain coefficient of the radio frequency circuit in the transmitting channel of the reference antenna. The reference signal received by the mth antenna can be expressed as:

in, is additive white Gaussian noise with a mean of zero and a variance of The reference signal from the mth antenna received by the reference antenna can be expressed as:

Among them, h _m→0 =r _b0 v _m t _bm is the channel parameter from the mth antenna to the reference antenna, r _b0 represents the gain coefficient of the radio frequency circuit in the receiving channel of the reference antenna, and t _bm represents the radio frequency circuit in the transmitting channel of the mth antenna gain factor, is additive white Gaussian noise with a mean of zero and a variance of

It is assumed that the reference signal power is Then the channel estimation between the reference antenna and the mth antenna is:

Then the relative gain coefficient of the radio frequency circuit of the mth antenna is obtained by the following formula:

When user k accesses the base station, the base station first sends a downlink reference signal to user k through the reference antenna Then user k feeds back the received reference signal to the base station through a dedicated link to Represents the feedback signal of user k received by the reference antenna and assumes that the feedback signal received by the base station is accurate, g _0→k represents the downlink channel parameter from the reference antenna to user k, where g _0→k =r _uk v _k t _b0 , r _uk represents the gain of the radio frequency circuit in the receiving channel of user k, and v _k represents the wireless channel parameter from the reference antenna to user k. Let the downlink reference signal power be The downlink channel estimation between the reference antenna and user k is:

in, is additive white Gaussian noise with a mean of zero and a variance of

Then user k transmits an uplink reference signal to the base station by represents the reference signal received by the reference antenna, g _k→0 represents the uplink channel parameter from user k to the reference antenna, where g _k→0 =r _b0 v _k t _uk , and t _uk represents the gain of the radio frequency circuit in the transmission channel of user k of the reference antenna . Let the uplink reference signal power be The uplink channel estimation between user k and the reference antenna is:

in, is additive white Gaussian noise with a mean of zero and a variance of

The relative gain coefficient of the RF circuit on the user side is obtained by the following formula:

3. User downlink channel estimation

(1) Downlink channel estimation based on absolute RF circuit gain coefficients

Based on the acquired absolute RF circuit gain coefficient, the acquisition of the statistical channel information of each user is completed by the uplink channel detection process. In the uplink, each user intermittently transmits a sounding signal, and the sounding signals of each user are orthogonal to each other.

by represents the sounding signal sent by the kth user in the cell in the tth sounding cycle, represents the sounding signal received by the mth antenna on the base station side in the tth sounding period, g _t,k→m represents the channel parameter between the kth user and the mth antenna on the base station side in the tth sounding period, represents the channel vector between the kth user and the M antennas on the base station side, The mth element of is g _t,k→m . Assume where T represents the matrix transpose. Assume where H represents the conjugate transpose of the matrix, is the power of the transmitted signal, I is the identity matrix, and L is the length of the detection signal. The sounding signal received by the base station can be expressed as:

in is an additive white Gaussian noise matrix, the mean of each element is zero, and the variance is

Assume The statistical model for where U is a fixed matrix (called an eigenmode matrix, for example, when the base station side antenna adopts a uniform linear array, the eigenmode matrix is a DFT matrix), which depends on the antenna configuration at the base station side, R _b =diag{r _b1 ,r _b2 ,...,r _bM } is the gain coefficient matrix of the radio frequency circuit of the receiving antenna on the base station side, r _bm is the gain coefficient of the radio frequency circuit of the mth antenna on the base station side, t _uk is the gain coefficient of the radio frequency circuit of the transmitting antenna of user k, and m _k is the radio frequency circuit of the kth user A vector composed of all unique channel statistical parameters (each element is a positive value), Each element of is subject to the assumption of independent and identical distribution (the mean of each element is zero and the variance is 1), Represents an element-wise product. say is the eigenmode domain channel vector of the kth user in the tth detection period, and set When the eigenmode matrix U is known, r _k is the required statistical channel information of the kth user, which is called the eigenmode domain channel energy coupling vector.

In the t-th detection period, the received signal is first Obtain the estimated value of the channel vector in the characteristic mode domain of each user, and the calculation formula is as follows:

Where * indicates that each element is conjugated, is the estimated value of the absolute RF circuit gain coefficient of the base station side antenna, Transmit antenna absolute RF circuit gain factor estimate for user k. then use and the sample-enhanced averaging method, the estimated value of the channel energy coupling vector e _k in the eigenmode domain can be obtained. The calculation formula is as follows:

In the formula, α _t' is the weighting factor, which satisfies _Ns is the window size. From _ek and U, the spatial correlation matrix of each user's wireless channel in the t-th detection period can be obtained:

In the formula, diag( _ek ) represents a diagonal matrix, and the vector formed by its diagonal elements is _ek . The spatial correlation matrix of the upstream channel:

The spatial correlation matrix of the downlink channel:

In the formula, is the estimation of the absolute RF circuit gain coefficient of the transmit antenna on the base station side, Estimated absolute RF circuit gain factor for user k's receive antenna.

The uplink channel estimation of each user is completed in the uplink training phase. Suppose there are K single-antenna users in the cell, and the pilot signals sent by the users are orthogonal to each other, that is, in is the power of the transmitted pilot signal, by represents the pilot signal received by the mth antenna on the base station side, g _k→m =r _bm w _km t _uk represents the uplink channel parameters between the kth user and the mth antenna on the base station side in the current training period, where w _km represents the wireless channel from user k to the mth antenna on the base station side, represents the uplink channel vector between the kth user and the M antennas on the base station side, and the mth element of g _k is g _k→m . Assume The pilot signal received by the base station can be expressed as:

Y ^tr =G ^ul X ^tr +Z ^tr (16)

where Z ^tr is an additive white Gaussian noise matrix, the mean of each element is zero, and the variance is

The base station makes a minimum mean square error (MMSE) estimate based on the received pilot signal, and can obtain the downlink channel estimate in the following three ways:

(1) Obtain the partial estimate of the wireless channel of user k by the following formula:

Then the downlink channel estimate is obtained by the following formula:

(2) First obtain the user k uplink channel estimate by the following formula:

Then the downlink channel estimate is obtained by the following formula:

(3) Obtain the downlink channel estimate of user k directly from the following equation

The downlink channel estimates obtained in the above three methods are the same. The estimated mean square error can be calculated as follows:

in,

(2) Downlink channel estimation based on relative RF circuit gain coefficient

In the uplink training phase, the uplink channel estimation of each user is completed. Suppose there are K single-antenna users in the cell, and the pilot signals sent by the users are orthogonal to each other, that is, in is the power of the transmitted pilot signal, by represents the pilot signal received by the mth antenna on the base station side, g _k→m =r _bm v _km t _uk represents the uplink channel parameters between the kth user and the mth antenna on the base station side in the current training period, where v _km represents the wireless channel from user k to the mth antenna on the base station side, represents the uplink channel vector between the kth user and the M antennas on the base station side, and the mth element of g _k is g _k→m . Assume The pilot signal received by the base station can be expressed as:

Y ^tr =G ^ul X ^tr +Z ^tr (23)

where Z ^tr is an additive white Gaussian noise matrix, the mean of each element is zero, and the variance is

Assume that the base station side knows the uplink channel statistics of each user The base station performs minimum mean square error (MMSE) channel estimation according to the received pilot signal, and obtains the estimated value of each user's uplink channel and its mean square error. The channel estimation value of the kth user is calculated as follows:

The mean squared error of its estimate is calculated as:

in is the transmit signal-to-noise ratio of each user in the uplink training phase.

Using the obtained uplink channel estimation value of each user and the mean square error of the channel estimation, the downlink channel estimation value can be obtained by the following formula:

In the formula, The mean square error of downlink channel estimation is obtained by the following formula:

4. Downlink robust precoding

Let x ^dl represent the data signal before precoding sent by the base station to the K users in the cell at the current moment, where the kth element is the data signal sent by the kth user, and its mean value is set to zero and the variance is The transmitted data signal of each user is a data signal obtained by channel coding, interleaving and modulation symbol mapping of the transmitted information bit stream. The base station precoding matrix is represented by B, and the signal actually sent by the base station side is Bx ^dl . The data signal received by the K users is represented by y ^dl , wherein the kth element is the data signal received by the kth user. The downlink channel is denoted by G ^dl . The received signal can be expressed as:

y ^dl =G ^dl Bx ^dl +z ^dl (28)

where z ^dl is an additive white Gaussian noise vector, the mean of each element is zero, and the variance is

Under the average minimum mean square error criterion, the robust precoding matrix on the base station side is calculated by the following formula:

in, is the average transmit signal-to-noise ratio of each user's downlink transmission, and γ is the transmit power constraint parameter on the base station side, which can be calculated by the following formula:

where tr{.} represents the matrix trace operation.

Using the received signal, each scheduled user can obtain the estimated value of the downlink transmitted information bit stream through processes such as demodulation, deinterleaving, and channel decoding.

5. Predict user traversal rate based on absolute RF circuit gain coefficient

If the absolute RF circuit gain coefficient is obtained, the downlink SNR of user k can be approximated by the following formula:

in,

in,

In the formula, Ψ is calculated by the following formula,

The elements in e[e ₁ ,e ₂ ,…,e _K ] are defined as For t=1,2,…

and initial value when When it is less than the given iteration termination value, the iteration stops, and the function and are defined as:

In the formula, Ψ' _I = Ψ'(I), where Ψ'(D) is a function of D defined as:

where e'(D) = [e ₁ '(D)...e' _K (D)] ^T is calculated by:

e'(D)=[In _-J ] ^-1 q(D) (43)

where J and q(D) are calculated by:

Then the traversal rate of user k can be approximated by the following formula:

The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited to this. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention. should be included within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims (9)

1. a massive MIMO downlink wireless communication method under a radio frequency circuit mismatch situation, applicable to a TDD massive MIMO wireless communication system, is characterized in that, the method comprises the steps: (1) RF circuit gain coefficient acquisition: by setting a gain coefficient measurement unit in the RF circuit at the front end of the transceiver antenna, the measurement unit sends a reference signal to the RF circuit and receives the signal feedback to estimate the absolute RF circuit gain coefficient on the base station side and the user side; or , obtain two-way channel estimation by sending and receiving reference signals between the reference antenna on the base station side and other antennas, and then obtain the relative RF circuit gain coefficient on the base station side, and obtain the uplink and downlink channel parameters by sending and receiving reference signals between the reference antenna and the user, and then obtain the relative frequency of the user side. RF circuit gain factor; (2) User downlink channel estimation: each user sends their respective uplink pilot signals on the same time-frequency resource, and the base station performs downlink of each user according to the received pilot signal and the acquired radio frequency circuit gain coefficients on the base station side and the user side. Channel estimation, and obtain estimation error statistics; wherein the downlink channel estimation is calculated and obtained based on the absolute radio frequency circuit gain coefficient combined with the uplink channel estimation, or calculated based on the relative radio frequency circuit gain coefficient combined with the uplink channel estimation; (3) Downlink robust precoding data transmission: In the downlink data transmission stage, the base station uses the acquired downlink channel estimation and the statistical information of the downlink channel estimation error to obtain the robust precoding matrix required to transmit data to each user signal, and implement the Robust precoding sends data signals to each user simultaneously on the same time-frequency resource. 2. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, wherein: in the TDD massive MIMO wireless communication system, the base station side antenna array comprises more than ten antenna units, The spacing between the antenna elements is smaller than the wavelength of the carrier, and each antenna element can use a single-polarized or multi-polarized antenna. 3. the massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, is characterized in that: the absolute radio frequency circuit gain coefficient in step (1) is calculated by following formula: in, sending the reference signal power for the gain factor measurement unit, represents the reference signal sent by the gain coefficient measurement unit to the RF circuit module at the front end of the antenna m in the t-th cycle, receiving a reference signal for the gain factor measurement unit, Indicates the estimated value of the gain coefficient of the radio frequency circuit at the front end of the transmit antenna on the base station side or the user side or the front end of the receive antenna. 4. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, is characterized in that: the relative radio frequency circuit gain coefficient of the base station side in step (1) is calculated by the following formula: in, is the estimation of the relative RF circuit gain coefficient of the mth antenna on the base station side, For the channel estimation from the reference antenna to the mth antenna, it is calculated by the following formula: in, To transmit the reference signal power, is the reference signal sent by the reference antenna in the t-th cycle, the reference signal sent by the reference antenna received by the mth antenna; is the channel estimation from the mth antenna on the base station side to the reference antenna, which is calculated by the following formula: in, is the reference signal sent by the mth antenna to the reference antenna in the tth cycle, is the reference signal sent by the mth antenna received by the reference antenna in the tth cycle. 5. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, is characterized in that: the user side radio frequency circuit relative gain coefficient in step (1) is obtained by the following formula: in, is an estimate of the relative RF circuit gain factor for user k, Represents the downlink channel estimate from the reference antenna to user k, which is calculated by the following formula: in, is the reference signal power, is the reference signal received by user k, the downlink reference signal sent by the reference antenna; Represents the uplink channel estimate from user k to the reference antenna, which is calculated by the following formula: in, is the reference signal received by the reference antenna, The uplink reference signal is transmitted to the base station for user k. 6. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, characterized in that: in step (2), performing downlink channel estimation based on the absolute radio frequency circuit gain coefficient is specifically: based on the base station side receiving antenna Absolute RF Circuit Gain Factor Estimation Transmitting Antenna Absolute RF Circuit Gain Coefficient Estimation User k receiving antenna absolute RF circuit gain factor Transmitting antenna absolute RF circuit gain factor And the pilot signal Y ^tr sent by the user received by the base station obtains the downlink channel estimate by any one of the following three methods: Method 1: First obtain the partial estimate of the wireless channel of user k by the following formula: Then the downlink channel estimate is obtained by the following formula: in, is the wireless channel estimation for user k, is the covariance matrix of user k wireless channel, is the covariance matrix of the wireless channel of user j, K is the number of users, ρ ^tr is the pilot training signal-to-noise ratio, is the orthogonal pilot sequence sent by user k, transmit pilot sequence power for user k; Method 2: First obtain the uplink channel estimate of user k by the following formula: Then the downlink channel estimate is obtained by the following formula: In the formula, is the covariance matrix of the uplink channel of user j, is the covariance matrix of the uplink channel of user k; Method 3: Obtain the downlink channel estimate of user k directly from the following formula In the formula, is the covariance matrix of the downlink channel of user j, is the covariance matrix of the downlink channel of user k; The covariance matrix of the downlink channel estimation error is calculated as follows: 7. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, characterized in that: in step (2), the downlink channel estimation based on the relative radio frequency circuit gain coefficient is specifically: based on the relative radio frequency on the base station side Circuit Gain Factor Estimation User k relative RF circuit gain factor estimation And the uplink channel estimation obtains the downlink channel estimate, and the downlink channel estimate of user k is calculated according to the following formula: in is the uplink channel estimation for user k, which is calculated by the following formula: where ρ ^tr is the pilot training signal-to-noise ratio, is the orthogonal pilot sequence sent by user k, Y ^tr is the pilot signal sent by the user received by the base station, transmit pilot sequence power for user k, is the covariance matrix of the uplink channel of user j, and K is the number of users; The downlink channel estimation error covariance matrix is obtained by the following formula: in is the estimation error of the uplink channel for user k, which is calculated by the following formula: 8. The massive MIMO downlink wireless communication method according to claim 1, wherein the robust precoding matrix in step (3) is calculated by the following formula: in, For downlink channel estimation, is the downlink channel estimation for user k, is the downlink channel estimation error covariance matrix, ρ ^dl is the average transmit signal-to-noise ratio of each user’s downlink transmission, γ is the transmit power constraint parameter on the base station side, which is calculated by the following formula: where tr{.} represents the matrix trace operation. 9. The massive MIMO downlink wireless communication method in the case of radio frequency circuit mismatch according to claim 1, it is characterized in that: also comprises the step that base station predicts user traversal rate based on absolute radio frequency circuit gain coefficient, and described user traversal rate is measured by: Calculated as follows: in, is the downlink SINR prediction of user k, which is calculated by the following formula: in, in, In the formula, Ψ is calculated by the following formula, The elements in e[e ₁ ,e ₂ ,…,e _K ] are defined as For t=1,2,… and initial value when When it is less than the given iteration termination value, the iteration stops, and the function and are defined as: in, Ψ' _I = Ψ'(I), where Ψ'(D) is a function of D defined as: where e′(D)=[e′ ₁ (D)…e′ _K (D)] ^T is calculated by: e'(D)=[In _-J ] ^-1 q(D) where J and q(D) are calculated by: