Search

CN-122017722-A - Intelligent super-surface-assisted low-bit quantized array non-line-of-sight signal source DOA estimation method

CN122017722ACN 122017722 ACN122017722 ACN 122017722ACN-122017722-A

Abstract

The invention discloses an intelligent super-surface-assisted low-bit quantization array non-line-of-sight signal source DOA estimation method, which comprises the steps of processing received signals by adopting a low-precision ADC (analog to digital converter) after receiving echo signals scattered by targets reflected by RIS (radio-digital system), constructing RIS equivalent channels, estimating quantized signals by using a RIS equivalent channel matrix as an auxiliary variable, converting an angle estimation problem into a convex optimization problem, solving the convex optimization problem by using a CVX tool box in Matlab software, constructing two one-dimensional Toeplitz matrixes respectively containing target DOA and DOD information, solving signal angles by using one-dimensional vandermonde decomposition, and performing angle matching to realize DOA and DOD joint estimation of the non-line-of-sight signal source. The invention creatively merges the active regulation and control capability of RIS, the low-bit quantization technology and the atomic norm minimization algorithm, can realize high-precision and real-time DOA and DOD estimation in a non-line-of-sight environment while carrying out low-bit quantization, and has stronger engineering practical value.

Inventors

  • XI FENG
  • LV QIAN
  • CHEN YUANHUA
  • CHEN SIYU

Assignees

  • 南京理工大学

Dates

Publication Date
20260512
Application Date
20260122

Claims (9)

  1. 1. The intelligent super-surface-assisted low-bit quantization array non-line-of-sight signal source DOA estimation method is characterized by comprising the following steps of: The method comprises the steps of 1, transmitting signals by an array antenna, namely M-element transmitting antenna arrays transmitting M mutually orthogonal unit power waveforms, and initializing DOA (DOA) and DOD (DOD) information of K incoherent targets in an observable area; Step 2, RIS reflects the target echo signals, wherein the RIS consists of Q reflecting units, the reflection coefficient of each unit is regulated and controlled in real time through a phase control matrix v, the echo signals scattered by the target are subjected to directional reflection, and a complete 'transmitting end-target-RIS-receiving end' non-line-of-sight signal transmission link is constructed; step 3, receiving signals by the array antenna, namely receiving RIS reflected signals by the N-element receiving antenna array, accumulating L snapshots and constructing a receiving signal matrix Y; step 4, constructing an equivalent channel matrix, namely constructing an equivalent channel matrix S based on a known RIS phase control matrix v and a channel matrix G from the RIS to the receiving end; Step 5, carrying out low-bit quantization on the received signal, namely carrying out low-bit quantization on the received signal matrix Y by utilizing a low-precision analog-to-digital converter to obtain a quantized signal matrix Z; step 6, estimating angle parameters based on LQANM algorithm, namely converting an angle estimation problem into a semi-positive definite programming problem by using an atomic norm minimization algorithm under low-bit quantization on a quantization matrix Z, and solving to obtain a Toeplitz matrix and an optimization variable X; and 7, obtaining DOA and DOD estimation results, namely performing vandermonde decomposition on the estimated Toeplitz matrix to obtain an estimated angle, performing angle matching by using an optimized variable X, and finally obtaining DOA and DOD estimation results corresponding to K targets.
  2. 2. The method for estimating the non-line-of-sight signal source DOA of the low-bit quantized array assisted by the intelligent super-surface according to claim 1 is characterized in that in the step 1, the system adopts a MIMO array, a transmitting end is a uniform linear array with M antennas, an RIS is a uniform linear array with Q array elements, and a receiving end is a uniform linear array with N antennas; for signal wavelength, K incoherent targets in the same range gate of radar system are set, and the guiding vector is emitted Representing a direction vector between the transmitting end and the kth target; Representing the transpose of the matrix, for the kth target, its emission angle DOD is noted as The reception angle DOA is recorded as 。
  3. 3. The method for estimating DOA of low-bit quantized array non-line-of-sight signal source with assistance of intelligent super surface as claimed in claim 1, wherein in step 2, RIS is used as a reconfigurable intelligent super surface capable of being regulated in real time, echo signals scattered by targets are reflected by RIS and then directed to a receiving antenna through regulation, and a channel matrix from RIS to a receiving end is Representing the corresponding relation of the channel phases between the receiving array antenna and the RIS array element, wherein the phase control signal of the RIS is There are L configurations representing the active regulation of RIS array elements for each snapshot, and the RCS coefficient of radar cross-sectional area of the target remains unchanged between snapshots: Is kept unchanged, wherein Representing the magnitude of the RCS, Representing RCS phase, receiving direction vector 。
  4. 4. The method for estimating the non-line-of-sight signal source DOA of an intelligent super-surface assisted low-bit quantization array according to claim 1, wherein in the step 3, a complete signal link of a transmitting end-target-RIS-receiving end is constructed by reflecting signals from RIS to a receiving antenna array, and the signals received by the receiving antenna array are as follows: Wherein the method comprises the steps of Is the channel noise in the first snapshot.
  5. 5. The method for estimating a non-line-of-sight signal source DOA of an intelligent super-surface aided low-bit quantization array according to claim 1, wherein in step 4, a RIS channel relation matrix is constructed by using a channel matrix G of a known RIS and a receiving antenna and a phase control signal v, and in the case of no noise, a channel correlation part is defined first: Utilizing Kronecker product kronecker product binding: ; The received vector is deformed into: ; the channel matrix portion is extracted outside the accumulation term and renamed to a new matrix vector: Collecting L pieces of snapshot data to obtain a complete received signal: the channel portion is intended to be extracted from the received signal vector, vectorized: Defining a channel matrix The vectorized signal model is: And finally, constructing a complete channel matrix vector by using G and v: 。
  6. 6. the intelligent super-surface assisted low-bit quantization array non-line-of-sight signal source DOA estimation method as claimed in claim 1, wherein in step 5, each radio frequency link at the receiving end quantizes the real part and the imaginary part of the received signal using a pair of real-valued quantizers, respectively: Wherein the method comprises the steps of In order to quantify the function of the quantization, ; For the purpose of the quantization interval, The minimum value is taken to be the smallest value, Rounding upwards; b is the quantization bit number, and is low-bit quantization when b=1 to 4, and Z is the low-bit quantized received signal.
  7. 7. The method for estimating a low-bit-rate quantized array non-line-of-sight signal source DOA as defined in claim 1, wherein in step 6, the signal model is analyzed, The angle estimation problem is converted into a convex optimization problem by constructing an equivalent channel matrix S as an aid, and an optimization variable is set A matching matrix U is constructed, and the matching matrix U is constructed, ; Approximating the matching matrix U to the quantized received signal Z and satisfying the quantization constraint in the iterative process: The objective optimization function is: The constraint conditions are as follows: In the middle of For two one-dimensional Toeplitz matrices, Representing the first row of the picture A layer of Toeplitz matrix is determined, The same is true, and the concrete representation is as follows: In the middle of Is a non-coherent coefficient; the optimization variable X is: solving the convex optimization problem by using a CVX toolbox of Matlab to obtain 。
  8. 8. The method for estimating a non-line-of-sight signal source DOA with an intelligent super surface aided low bit quantization array as set forth in claim 7, wherein the method is characterized by performing one-dimensional Van der Waals decomposition on the Toeplitz matrix by utilizing a special structure of the Toeplitz matrix to obtain angles And However, the angle sequence is not matched with the real sequence value, so that we use the X estimated by the algorithm to perform angle matching, and the steps are as follows: first, a direction vector is constructed by using the estimation angle result And All possible permutations and combinations are enumerated by means of an exhaustive list; then, utilizing Kronecker product properties: constructing an estimation result matrix And calculating the residual error of the X, and taking the result corresponding to the minimum residual value as a matching result.
  9. 9. The intelligent super-surface-assisted low-bit quantization array non-line-of-sight signal source DOA estimation method according to claim 1, wherein in step 7, angle estimation results are obtained by utilizing LQANM algorithm processing, namely DOA and DOD estimation results of targets are obtained.

Description

Intelligent super-surface-assisted low-bit quantized array non-line-of-sight signal source DOA estimation method Technical Field The invention belongs to the technical field of array signal processing, and particularly relates to an intelligent super-surface-assisted low-bit quantized array non-line-of-sight signal source DOA estimation method. Background Direction of arrival (DOA, direction of Arrival) estimation is a core research direction in the field of array signal processing, aimed at determining the azimuth angle of a signal source by processing signals received by an array antenna. For DOA estimation of a non-line-of-sight signal source, the traditional methods such as a multipath radar technology and a method based on received signal strength (RSS, received Signal Strength) and Arrival time difference (TDOA, time Difference of Arrival) have the limitations that the effective information acquisition difficulty of the multipath radar technology is high, and the method based on RSS and TDOA is limited by hardware conditions, so that the estimation precision and resolution are low. The development of reconfigurable intelligent supersurfaces (RIS, reconfigurable Intelligent Surface) provides a new idea for solving the problem of non-line-of-sight propagation. RIS is an artificial electromagnetic surface composed of a large number of programmable electromagnetic units, and intelligent regulation and control of space electromagnetic waves are realized through electromagnetic characteristics (such as amplitude and phase) of dynamic control units. The method has the advantages of low cost, real-time programming and the like, can be used as a relay in the non-line-of-sight signal receiving process, and improves the signal transmission quality through directional reflection. Atomic norm minimization (ANM, atomic Norm Minimization) is a sparse signal processing method based on a compressed sensing framework. The method converts the DOA estimation problem into a sparse optimization problem by constructing an atomic set of a signal space, and solves the DOA estimation problem by using a semi-definite programming (SDP, SEMIDEFINITE PROGRAMMING). The ANM method does not need discrete meshing, avoids grid separation errors, and can realize super-resolution DOA estimation under the condition of limited snapshot number or low signal-to-noise ratio. The low-bit quantization technique refers to digitizing a received signal using a low-resolution Analog-to-Digital Converter (ADC), and generally adopts a 1-bit to 4-bit low-bit quantization mode to reduce hardware cost, power consumption and data processing complexity. In the DOA estimation field, low-bit quantization is applied in massive MIMO systems or Internet of things scenarios to cope with high data rates and energy efficiency requirements. However, low bit quantization can introduce significant quantization noise and nonlinear distortion, resulting in loss of signal information, affecting the accuracy of angle estimation. In the application of low-bit quantization in the direction of arrival estimation of a MIMO-OFDM system, the prior art proposes a DOA estimation method based on 1-bit quantization, which reduces quantization errors by compressed sensing theory, but has limited performance in non-line-of-sight environments and does not consider RIS-assisted path enhancement. The low-bit quantization DOA estimation method reduces hardware cost, but has the defects that quantization error amplification is serious in signal attenuation under a non-line-of-sight propagation environment, nonlinear distortion is further introduced into low-bit quantization, so that signal-to-noise ratio (SNR) is reduced, root Mean Square Error (RMSE) of angle estimation is obviously increased, estimation performance is rapidly deteriorated under the condition of low signal-to-noise ratio (such as SNR < -10 dB), and adaptability is insufficient, that the method does not consider RIS-assisted path reconstruction and cannot optimize signal quality by utilizing controllable reflection, so that estimation accuracy is limited in a complex non-line-of-sight scene. In "Low-Precision DOA Estimation via Atomic Norm Minimization", the prior art discusses the adaptability of atomic norms to minimize under Low bit reception, quantization noise is processed by regularization, but the computational complexity is high and there is a lack of optimization for non-line-of-sight scenes. However, the scheme has high calculation complexity, increases the solving load of the semi-definite programming problem because of introducing additional regularization term for compensating quantization noise, prolongs the average running time, is unfavorable for real-time application, and has poor robustness that quantization error and noise are superposed under the condition of small snapshot number (L < 20), so that the toeplitz matrix decomposition fails and the angle matching error rate is increased. These drawbacks