A channel estimation method for intelligent reflector-assisted cloud access network

Abstract

A channel estimation method for a cloud access network assisted by an intelligent reflecting surface belongs to the field of wireless communication technology. It is characterized in that the direct channel and the reflected channel from the user UE to the radio remote head RRH and IRS providing services are estimated in three stages, and a pilot allocation method is applied in the channel estimation in each stage; a single-antenna UE communicates with a baseband processing unit BBU pool through the RRH, and multiple IRSs are deployed between the UE and the RRH; the RRH compresses the received signal and then transmits it to the BBU pool through a fronthaul link. The present invention is aimed at an IRS-assisted C-RAN uplink transmission system, and estimates the user's direct channel and reflected channel in stages to obtain more accurate channel state information CSI with less pilot resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communications, and in particular relates to a channel estimation method for an intelligent reflecting surface IRS-assisted cloud access network C-RAN. Background Art

[0002] With the rapid growth of mobile users and wireless devices, as well as the rapid emergence of new wireless applications such as augmented reality / mixed reality / virtual reality (AR / MR / VR), industrial automation, and tactile Internet, 5G may encounter various challenges in supporting the high demand for large-scale / ubiquitous connectivity in the upcoming Internet of Things (IoE) era. Therefore, traditional wireless communication systems can no longer meet current communication needs.

[0003] In response to the above challenges, the cloud access network C-RAN is a wireless communication system that is expected to alleviate current communication needs. The cloud access network is mainly composed of a baseband processing unit BBU, a radio remote head RRH and an antenna. The BBU is integrated backward into a BBU pool. The BBU pool and the RRH are connected by a fronthaul link, and the RRH is closer to the user. The main feature of the cloud access network is that the baseband signal processing can be migrated to the BBU pool, so that the RRH is only responsible for simple transmission and reception. Since some users are far away from the RRH, the intelligent reflection surface IRS can be used to assist users in accessing the RRH. The IRS is composed of passive reflection units, and the reflection characteristics (amplitude, phase) of each unit are digitally controllable, so the desired reflection beam can be formed. Unlike traditional relays, it can intelligently reconstruct the wireless network environment and effectively improve the performance of all aspects of the wireless network.

[0004] Through the IRS-assisted C-RAN access link, the user transmits the signal to the RRH through the direct link and the reflection link. The RRH receives the signal by compression and then transmits the signal to the BBU pool through the fronthaul link. The performance of this system depends on the acquisition of the early channel state information CSI, so an effective channel estimation method is urgently needed. Summary of the invention

[0005] The purpose of the present invention is to provide a channel estimation method for an intelligent reflecting surface IRS-assisted cloud access network C-RAN, which estimates the direct channel and reflected channel from the user to the remote radio head RRH and IRS providing services in three stages, and uses a pilot allocation method in the channel estimation of each stage to obtain higher-precision channel state information CSI with fewer pilot resources.

[0006] The technical solution of the present invention is:

[0007] A channel estimation method for a cloud access network assisted by an intelligent reflection surface, wherein a C-RAN access link is assisted by an IRS, and with the purpose of obtaining more accurate channel state information with fewer pilot resources, the direct channel and the reflected channel from the user to the RRH and IRS providing services are estimated in three stages, and a pilot allocation method is used in the channel estimation in each stage. The specific steps are as follows:

[0008] 1.1) In the access link transmission system of C-RAN assisted by IRS, users communicate with the BBU pool through RRH, and multiple IRSs are deployed between users and RRH to assist users in accessing RRH. In the link transmission system, there are K single-antenna user UEs, I RRHs, each RRH has M receiving antennas, J IRSs are deployed between users and RRHs, and each IRS has N reflection units. RRH compresses the received signal and then transmits it to the BBU pool through the fronthaul link.

[0009] 1.2) At time t in the channel estimation phase, user k sends a signal to the RRH Each RRH receives the signal sent by the user through the direct link and the reflection link of the IRS. The signal received by the i-th RRH, RRHi,i=1,…,I at time t is expressed as:

[0010]

[0011] in represents the direct channel from the kth user to RRHi, h i,k The covariance matrix of i,k I, α i,k Indicates h i,k p represents the same transmission power of the user. φ represents the set of IRSs serving RRHi. j,n,t represents the reflection coefficient of the nth reflection unit of the jth IRS at the tth time for the considered coherent block, satisfying g k,j,n =u k,j,n r i,j,n represents the UE-IRS-RRH reflection channel from the kth user to the i-th RRH through the nth IRS reflection unit of the jth IRS. is the overall channel from user k to the jth IRS, where u k,j The covariance matrix of k,j I, α k,j Indicates u k,j of path loss. is the overall path from the jth IRS to the ith RRH, where R i,j The covariance matrix of i,jNI,α i,j Represents R i,j The path loss is z (t) represents the additive white Gaussian noise (AWGN) at RRHi at time t, and its covariance matrix is ​​σ 2 I.

[0012] Furthermore, all IRS components are turned off, i.e., φ j,n,t = 0, estimate the direct channel from all users to the RRH providing service. Definition is the pilot sequence sent by user k in the first stage, then the RRHi received signal at time t (t=1,···,τ1) in the first stage is

[0013]

[0014] 2.1) The signal received at RRHi during the entire phase is

[0015]

[0016] in

[0017] 2.2) For different users with orthogonal pilot sequences, we have where q 2 =-1, then τ1≥K. Then we can get

[0018] 2.3) According to steps 2.1) and 2.2) we can get

[0019]

[0020] in Reuse The user set of the pilot.

[0021] 2.4) According to step 2.3), the one-stage LMMSE channel estimator can be obtained as

[0022]

[0023] in express The cth element of .

[0024] Further, all IRS reflection units are turned on, and there is only one typical user group Send non-zero pilot symbols to RRHi, that is, Estimate the reflected channel to the serving RRH and IRS. Definition is the pilot sequence of the second-stage user k, then the received signal at RRHi in the second-stage time slot t (t = τ1+1,…,τ1+τ2) is

[0025]

[0026] 3.1) In the second stage, due to h i,k The estimation is imperfect. The effective received signal at RRHi in time slot t (t = τ1+1, ..., τ1+τ2) given by the above formula can be expressed as

[0027]

[0028] 3.2) At this time, we use time-sharing for the second stage. Assume that in τ1-t k +1≤t≤τ1+t k In the second stage, only user k in the second stage user set is allowed to send non-zero pilot symbols. Then in τ1-t k +1≤t≤τ1+t k In the stage, the effective received signal at RRHi is expressed as

[0029]

[0030] in j k Existing and unique;

[0031] Reuse The user set of the pilot,

[0032] 3.3) Now let a k =a k' =[1,…,1] T , and construct a rank(Φ k )=N's Φ k , this Φ k The construction of is based on the DFT matrix, that is,

[0033]

[0034] in q 2 = -1. In this case, t k ≥N, it follows that Φ k (Φ k ) H =t k I.

[0035] 3.4) Re-order

[0036] As the total noise,

[0037]

[0038] 3.5) Redefine Then the LMMSE channel estimator in the second stage is

[0039]

[0040] Furthermore, some reflection units of the IRS are turned on according to time slot t to estimate the channels from the remaining users to the IRS providing services, and then the reflection channels from the remaining users to the RRH and IRS providing services can be estimated by using the channel correlation. is the pilot sequence of user k, then the received signal at RRHi in the time slot t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) of the third stage is

[0041]

[0042] in And k2,

[0043] 4.1) In the third stage, due to the h i,k and The estimation is imperfect, and the effective received signal at each time t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) at RRHi is

[0044]

[0045] 4.2) We assume that the channel estimation error make u k,j,n No self-interference is generated, then the formula of claim 4.1) becomes

[0046]

[0047] 4.3) At this point, we design an orthogonal transmission and reflection strategy: in each time slot t, only the user number is The users can transmit their pilot symbols to the RRH, and each IRS element contains only M t ≤M, using the set Δ t Indicates that M t =|Δ t |≤M, turned on to reflect pilot symbols so that RRH can estimate g k,j,n , Specifically, when t ≥ τ1 + τ2 + 1, we define in Therefore, for user number k t Users, assigned time slots to estimate its reflected channel. In each time slot, M IRS elements are turned on, and in the last time slot, the remaining In general, under this orthogonal transmission and reflection strategy, time slot to estimate g k,j,n ,

[0048] 4.4) At time

[0049] set up At this time, the effective received signal of RRHi at time slot t is reduced to

[0050]

[0051] in j k Existing and unique, Reuse The set of pilot users,

[0052]

[0053]

[0054] 4.5) It can be obtained from claim 4.4)

[0055]

[0056] 4.6) Redefine Then the LMMSE channel estimator in the third stage is

[0057]

[0058] 4.7) The corresponding reflection channel from the user to the RRH and IRS providing service is:

[0059] Further, the pilots of each user are allocated according to the following method: the relative positions of different users are analyzed according to certain rules, and then different pilots are allocated to users in each stage with the purpose of using as few pilot resources as possible, which specifically includes the following steps:

[0060] 5.1) In one stage, orthogonal pilots should be allocated to UEs served by the same RRH. This constraint can be represented by a K×K binary matrix B1, where the kth row and kth column of the matrix B1 are given by the following formula: in represents the set of RRHs serving user k, 1 means that any two UEs are served by at least one RRH and need to be assigned different pilots. Based on the matrix B1, we can construct an undirected Figure 1 To describe the relationship between any two UEs, if the b k,k′ If it is 1, they should be connected to each other, otherwise they should not be connected. And use the Dsatur algorithm to Figure 1 Dyeing is performed to obtain the number of colors for one stage of dyeing. and each user's coloring number C 1,k ,k=1,···,K,

[0061] 5.2) In the second stage, corresponding to the same constraint as in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B2 are given by:

[0062]

[0063] Based on the matrix B2, we can construct an undirected Figure 2 To describe the relationship between any two UEs, and use the Dsatur algorithm to Figure 2 Dyeing to get the number of colors required for the second stage and the serial number of each user's coloring in It is the user set of the second stage.

[0064] 5.3) In the third stage, corresponding to the same constraint in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B3 are given by the following formula:

[0065] Based on the matrix B3, we can construct an undirected graph 3 to describe the relationship between any two UEs, and use the Dsatur algorithm to color the undirected graph 3 to obtain the number of colors required for the three stages. and the serial number of each user's coloring in It is a set of users in three stages.

[0066] The beneficial effects of the present invention are:

[0067] The present invention uses an access link based on an intelligent reflecting surface IRS to assist a cloud access network C-RAN communication system, estimates the direct channel and reflected channel from the user to the remote radio head RRH and IRS providing services in three stages, and uses a pilot allocation method in the channel estimation at each stage, so that the required pilot resources are small and the estimation error is low. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 Schematic diagram of the access link system of the cloud access network assisted by intelligent reflective surface.

[0069] Figure 2 The relative error of channel estimation in the access link system of the cloud access network assisted by the intelligent reflector after adopting the method of the present invention. The figure shows the relationship between the relative error of channel estimation and the user's transmission power p: the three-stage channel estimation method is the method proposed by this patent. The traditional two-stage channel estimation method means that in the first stage, the IRS is turned off to estimate the direct channel from the user to the radio remote head (RRH) providing service; in the second stage, all IRSs are turned on to estimate the reflected channel from the user to the RRH and IRS providing service. DETAILED DESCRIPTION

[0070] The communication system transmission process based on IRS-assisted C-RAN access link is as follows: Figure 1 Before the transmission starts, the relative position of the user is used to determine which stage the user belongs to, and the user group in each stage is colored to determine the coloring sequence of the user. Then, the direct channel and reflection channel from the user to the RRH and IRS providing services are estimated in three stages in sequence, and finally the relative error of the estimation is obtained. The specific optimization process is as follows:

[0071] 1.1) In the access link transmission system of C-RAN assisted by IRS, users communicate with the BBU pool through RRH, and multiple IRSs are deployed between users and RRH to assist users in accessing RRH. In the link transmission system, there are K single-antenna user UEs, I RRHs, each RRH has M receiving antennas, J IRSs are deployed between users and RRHs, and each IRS has N reflection units. RRH compresses the received signal and then transmits it to the BBU pool through the fronthaul link.

[0072] 1.2) At time t in the channel estimation phase, user k sends a signal to the RRH Each RRH receives the signal sent by the user through the direct link and the reflection link of the IRS. The signal received by the i-th RRH, RRHi,i=1,…,I at time t is expressed as:

[0073]

[0074] in represents the direct channel from the kth user to RRHi, h i,k The covariance matrix of i,k I, α i,k Indicates h i,k p represents the same transmission power of the user. φ represents the set of IRSs serving RRHi. j,n,t represents the reflection coefficient of the nth reflection unit of the jth IRS at the tth time for the considered coherent block, satisfying g k,j,n =u k,j,n r i,j,n represents the UE-IRS-RRH reflection channel from the kth user to the i-th RRH through the nth IRS reflection unit of the jth IRS. is the overall channel from user k to the jth IRS, where u k,j The covariance matrix of k,j I, α k,j Indicates u k,j of path loss. is the overall path from the jth IRS to the ith RRH, where R i,j The covariance matrix of i,j NI,α i,j Represents R i,j The path loss is z (t) represents the additive white Gaussian noise (AWGN) at RRHi at time t, and its covariance matrix is ​​σ 2 I.

[0075] Specifically, all IRS components are turned off, i.e., φ j,n,t = 0, estimate the direct channel from all users to the RRH providing service. Definition is the pilot sequence sent by user k in the first stage, then the RRHi received signal at time t (t=1,···,τ1) in the first stage is

[0076]

[0077] 2.1) The signal received at RRHi during the entire phase is

[0078]

[0079] in

[0080] 2.2) For different users with orthogonal pilot sequences, we have where q 2 =-1, then τ1≥K. Then we can get

[0081] 2.3) According to steps 2.1) and 2.2) we can get

[0082]

[0083] in Reuse The user set of the pilot.

[0084] 2.4) According to step 2.3), the one-stage LMMSE channel estimator can be obtained as

[0085]

[0086] in express The cth element of .

[0087] Specifically, open all IRS reflection units, only one typical user group Send non-zero pilot symbols to RRHi, that is, Estimate the reflected channel to the serving RRH and IRS. Definition is the pilot sequence of the second-stage user k, then the received signal at RRHi in the second-stage time slot t (t = τ1+1,…,τ1+τ2) is

[0088]

[0089] 3.1) In the second stage, due to h i,k The estimation is imperfect. The effective received signal at RRHi in time slot t (t = τ1+1, ..., τ1+τ2) given by the above formula can be expressed as

[0090]

[0091] 3.2) At this time, we use time-sharing for the second stage. Assume that in τ1-t k +1≤t≤τ1+t k In the second stage, only user k in the second stage user set is allowed to send non-zero pilot symbols. Then in τ1-t k +1≤t≤τ1+t k In the stage, the effective received signal at RRHi is expressed as

[0092]

[0093] in j k Existing and unique.

[0094] Reuse The user set of the pilot,

[0095] 3.3) Now let a k =a k' =[1,…,1] T , and construct a rank(Φ k )=N's Φ k , this Φ k The construction of is based on the DFT matrix, that is,

[0096]

[0097] in q 2 = -1. In this case, t k ≥N, it follows that Φ k (Φ k ) H =t k I.

[0098] 3.4) Re-order

[0099] As the total noise,

[0100]

[0101] 3.5) Redefine Then the LMMSE channel estimator in the second stage is

[0102]

[0103] Specifically, some reflection units of the IRS are turned on in time slot t to estimate the channels from the remaining users to the IRS providing services, and then the reflection channels from the remaining users to the RRH and IRS providing services are estimated by using the channel correlation. is the pilot sequence of user k, then the received signal at RRHi in the time slot t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) of the third stage is

[0104] τ1+τ2+1≤t≤τ1+τ2+τ3

[0105] in And k2,

[0106] 4.1) In the third stage, due to the h i,k and The estimation is imperfect, and the effective received signal at each time t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) at RRHi is

[0107]

[0108] 4.2) We assume that the channel estimation error make u k,j,n No self-interference is generated, then the formula of claim 4.1) becomes

[0109]

[0110] 4.3) At this point, we design an orthogonal transmission and reflection strategy: in each time slot t, only the user number is The users can transmit their pilot symbols to the RRH, and each IRS element contains only M t ≤M, using the set Δ t Indicates that M t =|Δ t |≤M, turned on to reflect pilot symbols so that RRH can estimate g k,j,n , Specifically, when t ≥ τ1 + τ2 + 1, we define in Therefore, for user number k t Users, assigned time slots to estimate its reflected channel. In each time slot, M IRS elements are turned on, and in the last time slot, the remaining In general, under this orthogonal transmission and reflection strategy, time slot to estimate g k,j,n ,

[0111] 4.4) At time

[0112] set up At this time, the effective received signal of RRHi at time slot t is reduced to

[0113]

[0114] in j k Existing and unique, Reuse The set of pilot users,

[0115]

[0116]

[0117]

[0118] 4.5) From step 4.4) we can get

[0119]

[0120] 4.6) Redefine Then the LMMSE channel estimator in the third stage is

[0121]

[0122] 4.7) The corresponding reflection channel from the user to the RRH and IRS providing service is:

[0123] Specifically, the pilots of each user are allocated according to the following method: the relative positions of different users are analyzed according to certain rules, and then different pilots are allocated to users in each stage with the purpose of using as few pilot resources as possible, which specifically includes the following steps:

[0124] 5.1) In one stage, orthogonal pilots should be allocated to UEs served by the same RRH. This constraint can be represented by a K×K binary matrix B1, where the kth row and kth column of the matrix B1 are given by the following formula: in represents the set of RRHs serving user k, 1 means that any two UEs are served by at least one RRH and need to be assigned different pilots. Based on the matrix B1, we can construct an undirected Figure 1 To describe the relationship between any two UEs, if the b k,k′ If it is 1, they should be connected to each other, otherwise they should not be connected. And use the Dsatur algorithm to Figure 1 Dyeing is performed to obtain the number of colors for one stage of dyeing. and each user's coloring number C 1,k ,k=1,···,K,

[0125] 5.2) In the second stage, corresponding to the same constraint as in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B2 are given by:

[0126] Based on the matrix B2, we can construct an undirected Figure 2 To describe the relationship between any two UEs, and use the Dsatur algorithm to Figure 2 Dyeing to get the number of colors required for the second stage and the serial number of each user's coloring in It is the user set of the second stage.

[0127] 5.3) In the third stage, corresponding to the same constraint in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B3 are given by the following formula:

[0128] Based on the matrix B3, we can construct an undirected graph 3 to describe the relationship between any two UEs, and use the Dsatur algorithm to color the undirected graph 3 to obtain the number of colors required for the three stages. and the serial number of each user's coloring in It is a set of users in three stages.

[0129] Computer simulation shows that the relative error of channel estimation of the communication system based on IRS-assisted C-RAN access link is lower than that of the traditional two-stage channel estimation method after adopting the method of this patent.

Claims

1. A channel estimation method for a cloud access network assisted by an intelligent reflective surface, characterized in that: The IRS assists the C-RAN access link to estimate the direct channel and reflected channel from the user to the RRH and IRS providing services in stages, including the following steps: 1.1) In the access link transmission system based on IRS-assisted C-RAN, users communicate with the BBU pool through RRH, and multiple IRSs are deployed between users and RRH to assist users in accessing RRH; There are K single-antenna user UEs in the link transmission system, I RRHs, each RRH has M receiving antennas, J IRSs are deployed between the user and the RRH, and each IRS has N reflection units; the RRH compresses the received signal and then transmits it to the BBU pool through the fronthaul link; 1.2) At time t in the channel estimation phase, user k sends a signal to the RRH Each RRH receives the signal sent by the user through the direct link and the reflection link of the IRS. The signal received by the i-th RRH, RRHi,i=1,…,I at time t is expressed as: in represents the direct channel from the kth user to RRHi, h i,k The covariance matrix of i,k I, α i,k Indicates h i,k The path loss of p represents the same transmission power of the user. represents the set of IRSs served by RRHi, φ j,n,t represents the reflection coefficient of the nth reflection unit of the jth IRS at the tth time for the considered coherent block, satisfying g k,j,n =u k,j,n r i,j,n represents the UE-IRS-RRH reflection channel from the kth user to the i-th RRH through the n-th IRS reflection unit of the j-th IRS, where is the overall channel from user k to the jth IRS, where u k,j The covariance matrix of k,j I, α k,j Indicates u k,j The path loss is is the overall path from the jth IRS to the ith RRH, where R i,j The covariance matrix of i,j NI,α i,j Represents R i,j The path loss, z (t) represents the additive white Gaussian noise (AWGN) at RRHi at time t, and its covariance matrix is ​​σ 2 I; Turn off all IRS components, i.e. j,n,t = 0, estimate the direct channel from all users to the RRH providing service, and define is the pilot sequence sent by user k in the first stage, then the RRHi received signal at time t (t=1, ···, τ1) in the first stage is: 2.1) The signal received at RRHi during the entire phase is: in 2.2) For different users with orthogonal pilot sequences, we have where q 2 =-1, then τ1≥K, then we can get 2.3) According to steps 2.1) and 2.2) we can get in Reuse A user set of pilot signals; 2.4) According to step 2.3), the one-stage LMMSE channel estimator can be obtained as in express the cth element of ; Open all IRS reflector units, only one typical user group Send non-zero pilot symbols to RRHi, that is, Estimate the reflected channel to the serving RRH and IRS; define is the pilot sequence of the second-stage user k, then the received signal at RRHi in the second-stage time slot t (t = τ1+1, ..., τ1+τ2) is: 3.1) In the second stage, due to h i,k The estimation is imperfect. The effective received signal at RRHi in time slot t (t = τ1+1, ..., τ1+τ2) given by the above formula can be expressed as 3.2) At this time, we use time-sharing for the second stage, assuming that in τ1-t k +1≤t≤τ1+t k In the second stage, only user k in the second stage user set is allowed to send non-zero pilot symbols. Then in τ1-t k +1≤t≤τ1+t k In the stage, the effective received signal at RRHi is expressed as in j k Existing and unique, Reuse The user set of the pilot, 3.3) Now let a k =a k '=[1,…,1] T , and construct a rank(Φ k )=N's Φ k , Φ k The construction of is based on the DFT matrix, namely in q 2 = -1, in this case, t k ≥N, it follows that Φ k (Φ k ) H =t k I; 3.4) Re-order As the total noise, 3.5) Redefine Then the LMMSE channel estimator in the second stage is Turn on some reflection units of the IRS at time slot t, estimate the channels from the remaining users to the IRS providing services, and then use the channel correlation to estimate the reflection channels from the remaining users to the RRH and IRS providing services; define is the pilot sequence of user k, then the received signal at RRHi in the time slot t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) of the third stage is τ1+τ2+1≤t≤τ1+τ2+τ3 in and 4.1) In the third stage, due to the h i,k and The estimation is imperfect, and the effective received signal at each time t (t = τ1 + τ2 + 1, ..., τ1 + τ2 + τ3) at RRHi is 4.2) We assume that the channel estimation error make u k,j,n No self-interference is generated, then the formula of claim 4.1) becomes 4.3) At this point, we design an orthogonal transmission and reflection strategy: in each time slot t, only the user number is The users can transmit their pilot symbols to the RRH, and each IRS element contains only M t ≤M, using the set Δ t Indicates that M t =|Δ t |≤M, turned on to reflect pilot symbols so that RRH can estimate Specifically, when t ≥ τ1 + τ2 + 1, define in Therefore, for user number k t Users, assigned time slots to estimate its reflection channel. In each time slot, M IRS elements are turned on, and in the last time slot, the remaining IRS elements are turned on. Under this orthogonal transmission and reflection strategy, Time slot estimation 4.4) At time set up At this time, the effective received signal of RRHi at time slot t is reduced to in j k Existing and unique, Reuse The set of pilot users, 4.5) From step 4.4) we can get 4.6) Redefine Then the LMMSE channel estimator in the third stage is 4.7) The corresponding reflection channel from the user to the RRH and IRS providing service is: The pilots of each user are allocated according to the following method: the relative positions of different users are analyzed, and different pilots are allocated to the users in each stage, which specifically includes the following steps: 5.1) In one stage, orthogonal pilots should be allocated to UEs served by the same RRH. This constraint can be represented by a K×K binary matrix B1, where the kth row and kth column of the matrix B1 are given by the following formula: in represents the set of RRHs serving user k, 1 means that any two UEs are served by at least one RRH and need to be assigned different pilots; based on the matrix B1, we can construct an undirected graph 1 to describe the relationship between any two UEs. If the b of any two UEs k, k′ k,k′ If the two graphs are 1, they should be connected to each other, otherwise they should not be connected. The Dsatur algorithm is used to color the undirected graph 1 to obtain the color number of the first-stage coloring. and each user's coloring number C 1,k ,k=1,···,K, 5.2) In the second stage, corresponding to the same constraint as in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B2 are given by: Based on the matrix B2, we can construct an undirected graph 2 to describe the relationship between any two UEs, and use the Dsatur algorithm to color the undirected graph 2 to obtain the number of colors required for the second stage. and the serial number of each user's coloring in is the user set of the second stage; 5.3) In the third stage, corresponding to the same constraint in the first stage - orthogonal pilots should be allocated to UEs served by the same RRH, the kth row and kth column of the binary matrix B3 are given by the following formula: Based on the matrix B3, we can construct an undirected graph 3 to describe the relationship between any two UEs, and use the Dsatur algorithm to color the undirected graph 3 to obtain the number of colors required for the three stages. and the serial number of each user's coloring in It is a set of users in three stages.

Images