Search

CN-121984543-A - Improved channel estimation method based on combination of RIP criterion and SOMP algorithm

CN121984543ACN 121984543 ACN121984543 ACN 121984543ACN-121984543-A

Abstract

The invention provides an improved channel estimation method based on combination of RIP criteria and SOMP algorithm, relates to the technical field of channel estimation of RIS auxiliary systems, is suitable for the channel estimation algorithm of RIS auxiliary millimeter wave large-scale MIMO systems, aims at solving the problems that the traditional SOMP algorithm lacks sparsity prior information and atom selection strategy is stiff, designs a sparsity estimation method based on the RIP criteria when the SOMP algorithm sets the atom number selection of initial iteration, takes the result as the initial atom selection number, and utilizes the residual energy ratio of two iterations before and after the SOMP algorithm to dynamically adjust the atom number of the subsequent iteration process.

Inventors

  • LV RUIHONG
  • SHEN HONGBO

Assignees

  • 沈阳工业大学

Dates

Publication Date
20260505
Application Date
20260210

Claims (8)

  1. 1. An improved channel estimation method based on the combination of RIP criterion and SOMP algorithm is characterized in that, The method comprises the following steps: Step S1, a signal model of an RIS auxiliary uplink multi-user millimeter wave MIMO system is established, a channel transmission process is divided into Q time slots, and signals from K users received by a base station end in the Q time slots are obtained; S2, extracting a received signal of each user in each time slot based on the signal model to obtain a cascade channel between the base station and the RIS users; S3, converting the cascade channel model into a virtual angle domain channel by utilizing the angle domain sparse characteristic of the millimeter wave channel for analysis; S4, combining the received signals of all time slots, constructing a compressed sensing observation model in a multi-measurement vector form, and converting a channel estimation problem into an MMV problem with common column sparsity; S5, performing sparsity pre-estimation on the observation model based on RIP criteria to obtain a channel sparsity estimation value; And S6, taking the channel sparsity estimation value as an initial atomic number, adopting an improved SOMP algorithm, and iteratively recovering a sparse angular domain channel matrix through a dynamic atomic selection and backtracking screening mechanism to finish channel estimation.
  2. 2. The improved channel estimation method based on the combination of RIP criteria and SOMP algorithm according to claim 1, wherein in step S1, the specific method for establishing a signal model of an RIS-assisted uplink multiuser millimeter wave MIMO system, dividing a channel transmission process into Q time slots, and obtaining signals from K users received by a base station in the Q time slot includes: The signal model building process comprises the following steps: ; wherein K is the number of users, For the channel between the base station BS and the RIS, For the channel between the kth user and the RIS, For the pilot signal transmitted by the kth user in the qth slot, In the case of additive white gaussian noise, 。
  3. 3. The improved channel estimation method based on the combination of RIP criteria and SOMP algorithm as claimed in claim 1, wherein the specific method for extracting the received signal of each user in each time slot based on the signal model in step S2 to obtain the cascade channel between base station-RIS-users comprises: ; Where y k,q is the received signal of the kth user in the qth slot, due to Then it can be further expressed as ; Thus, the cascade channel between base station-RIS-users can be expressed as: 。
  4. 4. The method for preparing the ceramic particle reinforced FeCoNiCrAl high-entropy alloy coating according to claim 1, wherein the specific method for converting the cascade channel model into a virtual angle domain channel for analysis by utilizing the angle domain sparse characteristic of the millimeter wave channel in the step S3 comprises the following steps: description of the BS-RIS channel Using the widely applied Saleh-Valenzuela channel model With RIS-UE channels ; BS-RIS channel The description is as follows: ; Wherein, the Indicating the number of antennas at the base station side, The number of reflection units representing the RIS, The number of paths is indicated and, Is the first The complex gain of the strip path is used, And The departure angle and the arrival angle of the first path of the BS-RIS channel respectively; And Array guide vectors of the RIS end UPA and the BS end ULA respectively; RIS-UE channel The description is as follows: ; Wherein, the Is the number of paths for the kth user, Is the complex gain of the gain amplifier, AoD representing the p-th path of the user-RIS; The concatenated channel is further denoted as: ; Wherein, the Is the equivalent complex gain after combination; projecting physical channels into virtual corner domains by building an overcomplete dictionary matrix And The column vector of which is composed of discretized guide vectors covering the whole angle domain Subsequent analysis considered using virtual angular fields to represent the concatenated channel as: ; Wherein, the Is the angular domain channel matrix of the kth user.
  5. 5. The method for improved channel estimation based on RIP criterion and SOMP algorithm as claimed in claim 1, wherein the step S4 is characterized in that the specific method for constructing a compressed sensing observation model in the form of multi-measurement vector by combining the received signals of all time slots to convert the channel estimation problem into the MMV problem with common column sparsity comprises the following steps: signals combining Q slots: ; Wherein, the Representing the number of Q consecutive time slots, the superimposed signal vectors received by the base station at each time slot from all users, arranged in columns into a matrix, And And the same is done; the joint observation model for Q slots is: ; Substituting the cascade channel model represented by the angle domain to obtain: ; To be the Converting into a compressed sensing form, performing conjugate transposition on the equation, wherein, Representing a conjugate transpose of the matrix; For all users, due to the shared BS-RIS channel, their non-zero elements are identical in column direction, i.e. there is common column sparsity, let Ignoring the common column sparsity, reducing the algorithm complexity, and obtaining: ; To convert the upper mode into the applicable compressed sensing mode, let I.e. ; In this case with the multi-measurement vector MMV problem that aims at observing matrices from a set of observation matrices sharing the same sparse support set In which the sparse matrix is jointly recovered 。
  6. 6. The improved channel estimation method based on the combination of RIP criteria and SOMP algorithm according to claim 1, wherein in step S5, the specific method for performing sparsity pre-estimation on the observation model based on RIP criteria to obtain the channel sparsity estimation value comprises: initializing a sparsity candidate value S, a residual matrix R and a candidate original subset; in the iterative process, screening a group of candidate atoms according to the correlation between the current residual error and each column of the dictionary, and calculating updated residual error through least square reconstruction; And when the residual error is not reduced any more and is not iterated for the first time, ending the iteration and outputting a final sparsity estimation value.
  7. 7. The improved channel estimation method based on the combination of RIP criteria and SOMP algorithm as claimed in claim 1, wherein in step S6, the specific method for performing channel estimation by using the channel sparsity estimation value as an initial number of atoms, using an improved synchronous orthogonal matching pursuit algorithm, and iteratively recovering a sparse angular domain channel matrix through a dynamic atom selection and backtracking screening mechanism comprises: taking the sparsity estimation value as the atom selection number of the first iteration; in each iteration, selecting a plurality of atoms to be added into a support set according to the correlation between the residual error and the dictionary matrix; updating residual errors through least square, and calculating the energy ratio of the residual errors before and after twice; If the residual error is obviously reduced, dynamically increasing the selection number of atoms in the next round; if the residual error is reduced and slowed down, entering a backtracking screening stage, and eliminating redundant or error atoms; And stopping iteration when the residual error converges to a set threshold value or does not fall any more, and outputting an estimated channel matrix.
  8. 8. A method according to claim 7 based on RIP criteria an improved channel estimation method combined with the SOMP algorithm, the backtracking screening stage is characterized by comprising the following steps: Temporarily removing each atom in the current support set, and calculating the residual variation after removal; if the residual error does not increase or decrease after removing an atom, the atom is judged to be a redundant atom and is rejected.

Description

Improved channel estimation method based on combination of RIP criterion and SOMP algorithm Technical Field The invention relates to the technical field of channel estimation of RIS auxiliary systems, in particular to an improved channel estimation method based on combination of RIP criteria and SOMP algorithm. Background As wireless communication evolves toward higher frequency bands, larger capacity, and lower latency, millimeter wave (millimeter-wave) communication becomes one of the key technologies of fifth generation (5G) and future sixth generation (6G) mobile communication systems by virtue of its abundant spectrum resources. However, millimeter wave signals face serious path loss and penetration loss in the propagation process, and are easily shielded by obstacles, so that the coverage range of the millimeter wave signals is limited, and the link stability is poor. To effectively overcome this physical layer limitation, reconfigurable smart surfaces (Reconfigurable Intelligent Surface, RIS) have received widespread attention in recent years as an emerging passive smart reflection technology in academia and industry. The RIS is composed of a large number of low-cost programmable electromagnetic units, and the amplitude-phase response of each unit can be regulated and controlled in real time through software, so that the wireless propagation environment is intelligently reconstructed, and the directional enhancement, beam forming and interference suppression of signals are realized. In a millimeter wave large-scale Multiple-Input Multiple-Output (MIMO) system assisted by an RIS, the RIS can construct an adjustable line-of-sight or strong reflection path, so that the coverage, spectral efficiency and energy efficiency of the system are remarkably improved, and a very promising solution is provided for practical deployment of high-frequency band communication. However, the introduction of RIS also significantly increases the complexity of the system channel structure and difficulty of Channel State Information (CSI) acquisition. In a typical downlink or uplink RIS-assisted communication scenario, the overall channel is formed by concatenating base station-RIS segments with RIS-user segments, with the dimension of the channel matrix increasing linearly or even secondarily with the number of RIS units. Millimeter wave channels exhibit significant sparsity in physical propagation. Due to the enhanced spatial selectivity of high frequency electromagnetic waves, multipath components in the real environment are mainly concentrated on limited scattering paths, so that the channel has sparse representation in the angle domain or the beam domain. The structural characteristic provides a theoretical basis for a channel estimation method based on compressed sensing (Compressive Sensing, CS). The CS technology utilizes sparse prior of signals, and can recover original signals with high probability of observation data far lower than Nyquist rate, so that pilot frequency overhead is greatly reduced. In recent years, a series of CS methods based on greedy tracking algorithms have been introduced into the channel estimation of RIS-aided systems, such as orthogonal matching Pursuit (Orthogonal Matching Pursuit, OMP), segmented OMP (STAGEWISE OMP, stOMP), compressed sample matching Pursuit (Compressive SAMPLING MATCHING burst, coSaMP), and synchronous orthogonal matching Pursuit (Simultaneous OMP, SOMP). The SOMP algorithm can use joint sparsity among multiple measurement vectors to share a common support set among multiple users or time slots, so that estimation efficiency is further improved, and the SOMP algorithm becomes a hot spot of current research. However, the existing CS-based RIS channel estimation scheme still faces a plurality of key challenges, namely, most greedy tracking algorithms need to take channel sparsity as a priori parameter or iteration termination condition, in an actual wireless environment, the effective path number is time-varying and unknown due to user movement, barrier change and dynamic scattering, so that the algorithm performance is seriously dependent on inaccurate sparsity assumption, and secondly, the traditional method generally adopts a fixed number of atom selection strategies in an iteration process, lacks self-adaptive capacity on dynamic change of a channel sparse structure, easily causes atom misselection, under-fitting or over-fitting, further influences estimation accuracy and algorithm robustness, and in addition, the existing scheme often causes estimation performance to be rapidly deteriorated due to observation matrix correlation enhancement under the condition of low signal-to-noise ratio or pilot frequency resource limitation. Therefore, for the RIS-assisted millimeter wave large-scale MIMO system, research on a high-efficiency channel estimation method which does not depend on priori sparseness knowledge and has a dynamic atom selection mechanism and strong r