CN-121984807-A - Multi-mode cooperative signal detection and on-demand frequency offset estimation method and system based on matrix parallel architecture

CN121984807ACN 121984807 ACN121984807 ACN 121984807ACN-121984807-A

Abstract

The invention relates to a multi-mode collaborative signal detection and on-demand frequency offset estimation method and system based on a matrix parallel architecture, and belongs to the technical field of satellite communication signal processing. The method solves the problems of unstable detection, high frequency estimation calculation, large occupation of hardware resources and insufficient instantaneity of the traditional technology in a high-frequency-offset high-dynamic scene. The method comprises the steps of constructing a sliding window signal matrix to generate a basic related data set in parallel, adopting three detection modes of a full-coherent path, a symbol differential path and a sample differential path to perform parallel operation, performing cooperative judgment based on a signal-to-noise ratio self-adaptive threshold and a path energy state, adaptively selecting an optimal path and performing frequency offset estimation on demand on the selected path. The invention realizes the high-robustness detection of a wide frequency offset range, and obviously reduces the operation complexity and the processing delay.

Inventors

ZHANG HONGSHENG
SHEN ZIANG
SUN WEIWEI
TANG ZHENG

Assignees

重庆邮电大学

Dates

Publication Date: 20260505
Application Date: 20260126

Claims (5)

1. A multi-mode collaborative signal detection and on-demand frequency offset estimation method based on a matrix parallel architecture is characterized by comprising the following steps: matrix constructing step, namely, received digital baseband signals Buffering, namely rearranging continuously input baseband sampling points into a sliding window matrix R with a dimension of L multiplied by N according to a time sequence based on the local word length L and the number N of single-row sampling points, wherein each row of the matrix consists of N continuous sampling points, and L rows of the matrix correspond to L symbol fragments of the local word; basic related data generating step, local unique word sequence is rearranged into matrix in same mode For each row of matrix R Performing conjugate inner product on the corresponding row to obtain a basic related data set Wherein Representing the local correlation result of the kth sampling point in the detection window, and calculating the average power of the signal of the current detection window ; The multi-mode operation step, inputting the basic related data set X into a parallel differential array, generating four groups of intermediate variables in parallel, including a Path1 full coherence Path Path2 symbol differential Path Wherein Path3 sample differential Path Wherein And Path4 estimated frequency Path Wherein And is also provided with ; Normalization decision step, respectively taking modular squares of intermediate variables of Path1, path2 and Path3 and dividing the modular squares by the intermediate variables Obtaining normalized detection measurement value And sets an adaptive threshold based on a Signal-to-Noise Ratio (SNR) ; A step of collaborative decision and frequency estimation according to need, which is to execute collaborative decision according to the state of M 1 、M 2 、M 3 , select the optimal detection Path, if M 1 >Th 1 and M 2 >Th 2 , select Path1 and activate Path4 to calculate If M 1 ≤Th 1 and M 2 >Th 2 use the complex output of Path3 Judging whether or not to meet Wherein K is a preset phase slope threshold, if yes, selecting Path2 or Path3, if yes And And if the Path3 is invalid or the Path3 is roughly judged to be a high frequency offset, selecting the Path3, and only calling a coordinate rotation digital computer (CORDIC) unit for phase extraction on the selected Path to obtain a frequency offset estimation value.
2. The method for detecting and estimating frequency offset as required by multi-mode cooperative signals based on matrix parallel architecture according to claim 1, wherein the detection measure of the Path1 full coherence Path is And the frequency offset estimation variable is Its frequency offset is not fuzzy and its upper limit is full Wherein Is the sampling rate.
3. The method for detecting and estimating frequency offset as required by multi-mode cooperative signal based on matrix parallel architecture as set forth in claim 1, wherein the Path2 symbol differential Path detection and frequency estimation variables are The upper limit of the applicable frequency deviation range is The detection and frequency estimation variables of the Path3 sample point differential Path are as follows The upper limit of the applicable frequency deviation range is The frequency offset estimation value is calculated by the complex variable phase angle of the selected Path, and if the selected Path is Path1, the frequency offset estimation value is If the selected Path is Path2, the frequency offset estimation value is If the selected Path is Path3, then the frequency offset estimation value 。
4. The method for detecting and estimating the frequency offset on demand of the multi-mode cooperative signal based on the matrix parallel architecture of claim 1, wherein the cooperative decision strategy comprises: If M 1 >Th 1 and M 2 >Th 2 are determined to be low frequency offset and Path1 is selected; If M 1 ≤Th 1 and M 2 >Th 2 are the same, the phase of Path3 is used for roughly distinguishing the medium frequency offset from the high frequency offset; Path3 is selected if M 1 and M 2 are both inactive or Path3 is roughly discriminated as a high frequency offset.
5. A multi-mode collaborative signal detection and on-demand frequency offset estimation system based on a matrix parallel architecture is characterized by comprising a matrix operation module, a parallel differential array, a normalization detection module, a collaborative decision module and an on-demand frequency estimation module, wherein the matrix operation module is used for constructing an LxN sliding window matrix R and generating a basic related data set X and a signal average power An output of which is connected to an input of the parallel differential array; The parallel differential array is used for executing the front and rear half-section accumulation and conjugate product of the Path1 full-coherence accumulation, path2 symbol difference, path3 sample point difference and Path1 frequency estimation Path in parallel, and the output of the parallel differential array is connected to the input of the normalization detection module; The normalization detection module is used for calculating a normalization detection metric value M 1 、M 2 、M 3 , the output of which is connected to the input of the collaborative decision module, the collaborative decision module is used for selecting an optimal path according to M 1 、M 2 、M 3 , the output of which is connected to the input of the on-demand frequency estimation module, and the on-demand frequency estimation module only performs phase calculation on complex accumulation results of the selected path to output a frequency offset estimation value.

Description

Multi-mode cooperative signal detection and on-demand frequency offset estimation method and system based on matrix parallel architecture Technical Field The invention belongs to the technical field of satellite communication signal processing, and relates to a multi-mode collaborative signal detection and on-demand frequency offset estimation method and system based on a matrix parallel architecture. Background In the technical field of satellite communications, a receiver needs to overcome various adverse factors such as doppler frequency offset, noise interference, symbol timing offset, and the like, accurately detect a synchronization sequence such as Unique Word (UW) or preamble, and accurately estimate carrier frequency offset. Doppler frequency offset is a key factor influencing receiving performance, and particularly in high dynamic communication scenes such as low-orbit satellites, the frequency offset is wide in range and quick in change, and a serious challenge is provided for robustness and instantaneity of a detection algorithm. The traditional signal detection method generally adopts a shift register structure for processing, and the serial processing mode has low efficiency and is difficult to meet the low delay requirement of burst communication. In addition, in order to cover a wide frequency offset range, the existing receiver often adopts a full path calculation mode, that is, a complete phase extraction arctangent function (ARCTANGENT, atan 2) and frequency offset calculation are executed on all algorithm branches in parallel, which results in serious waste of calculation resources, increases hardware complexity and power consumption, and limits the application of the system in a resource limited environment. Therefore, developing a signal detection and frequency estimation method capable of adapting to a wide frequency offset range, with high robustness and low computational complexity is an urgent need in the art. Disclosure of Invention In view of the above, the present invention aims to provide a method and a system for detecting and estimating on-demand frequency offset of a multi-mode cooperative signal based on a matrix parallel architecture. In order to achieve the above purpose, the present invention provides the following technical solutions: A multi-mode collaborative signal detection and on-demand frequency offset estimation method based on a matrix parallel architecture comprises the following steps: matrix constructing step, namely, received digital baseband signals Buffering, namely rearranging continuously input baseband sampling points into a sliding window matrix R with a dimension of L multiplied by N according to a time sequence based on the local word length L and the number N of single-row sampling points, wherein each row of the matrix consists of N continuous sampling points, and L rows of the matrix correspond to L symbol fragments of the local word; basic related data generating step, local unique word sequence is rearranged into matrix in same mode For each row of matrix RPerforming conjugate inner product on the corresponding row to obtain a basic related data setWhereinRepresenting the local correlation result of the kth sampling point in the detection window, and calculating the average power of the signal of the current detection window; The multi-mode operation step, inputting the basic related data set X into a parallel differential array, generating four groups of intermediate variables in parallel, including a Path1 full coherence PathPath2 symbol differential PathWhereinPath3 sample differential PathWhereinAnd Path4 estimated frequency PathWhereinAnd is also provided with; Normalization decision step, respectively taking modular squares of intermediate variables of Path1, path2 and Path3 and dividing the modular squares by the intermediate variablesObtaining normalized detection measurement valueAnd sets an adaptive threshold based on a Signal-to-Noise Ratio (SNR); A step of collaborative decision and frequency estimation according to need, which is to execute collaborative decision according to the state of M 1、M2、M3, select the optimal detection Path, if M 1>Th1 and M 2>Th2, select Path1 and activate Path4 to calculateIf M 1≤Th1 and M 2>Th2 use the complex output of Path3Judging whether or not to meetWherein K is a preset phase slope threshold, if yes, selecting Path2 or Path3, if yesAndAnd if the Path3 is invalid or the Path3 is roughly judged to be a high frequency offset, selecting the Path3, and only calling a coordinate rotation digital computer (CORDIC) unit for phase extraction on the selected Path to obtain a frequency offset estimation value. Further, the detected value of the Path1 full coherence Path is as followsAnd the frequency offset estimation variable isIts frequency offset is not fuzzy and its upper limit is fullWhereinIs the sampling rate. Further, the Path2 symbol differential Path has a detection and frequency estimation variable