A non-reconstruction direct positioning method based on data compression for mobile receivers

Abstract

The present invention belongs to the field of signal processing technology and relates to a non-reconstruction direct positioning method based on mobile receiver data compression. The present invention aims to provide a non-reconstruction direct positioning method based on mobile receiver data compression. This method utilizes the characteristics of the Hadamard matrix to directly estimate the source position in the compressed measurement domain without the need for signal reconstruction and TDOA and FDOA parameter extraction. The present invention demonstrates that signals can still retain TDOA and FDOA information after Hadamard matrix compression, thereby reducing transmission throughput while maintaining positioning accuracy.

Description

Technical Field

[0001] The invention belongs to the technical field of signal processing and relates to a non-reconstruction direct positioning method based on mobile receiver data compression. Background Art

[0002] Passive positioning involves receiving, analyzing, and processing signals from non-cooperative emitters without transmitting detection signals, extracting the emitter's location parameters from the received signals. Compared to active positioning systems installed on active platforms like radar and sonar, passive positioning systems offer improved concealment, stronger survivability against anti-radiation attacks, and a longer detection range. It's no secret that electronic warfare systems play a crucial role in modern warfare, acting as force multipliers. As a crucial component of electronic support, the importance of passive positioning technology is self-evident. Consequently, countries around the world are prioritizing and vigorously developing this technology.

[0003] Passive positioning technology has long held broad application prospects and significant value. Specifically, tracking, monitoring, and reconnaissance and early warning of potentially threatening individual enemy targets and valuable platform equipment facilitate continuous perception and rapid assessment of battlefield situation, individual identification and threat assessment, and support and guidance for subsequent countermeasures. From this perspective, passive positioning is a crucial means of information acquisition in modern information-based warfare. Consequently, passive positioning has, and will continue to, receive widespread attention and ongoing research and exploration, with related technologies continuously evolving. In particular, multi-sensor passive positioning systems based on time difference of arrival (TDOA) and frequency difference of arrival (FDOA) information have broad application areas and require minimal equipment complexity.

[0004] As an improvement to the two-step localization method, direct localization has garnered widespread attention in recent years. Unlike the two-step method, while direct localization also utilizes the time and frequency information embedded in the signal, it does not estimate intermediate parameters. Instead, it directly maps the signal to the emitter's location, estimating the emitter's location directly from the signal observations. Specifically, direct localization first establishes a cost function for the observed signal with respect to position. It then selects the location candidate that optimizes the cost function as the emitter's location estimate. Compared to the two-step method, direct localization satisfies measurement-related constraints for any location candidate. As such, direct localization typically achieves better localization results than the two-step method and is more adaptable to weak observability conditions. However, since direct localization directly estimates the emitter's location from the signal, its processing strategy is a grid search. This involves dividing the localization area of ​​interest into a grid and iterating through all grid points one by one to find the optimal estimate. Consequently, the computational complexity increases exponentially with increasing signal length. In addition, centralized direct positioning requires all sensor raw data to be transmitted to a central receiving station, which will cause a large communication burden when the number of data packet transmission hops is large. Summary of the Invention

[0005] To address these existing issues, the present invention aims to provide a non-reconstruction direct positioning method based on mobile receiver data compression. This method leverages the properties of the Hadamard matrix to estimate the source position directly in the compressed measurement domain, eliminating the need for signal reconstruction and TDOA and FDOA parameter extraction. The present invention demonstrates that signals compressed using the Hadamard measurement matrix can still retain TDOA and FDOA information, thereby reducing transmission overhead while maintaining positioning accuracy.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A non-reconstruction direct positioning method based on mobile receiver data compression, which uses the Hadamard matrix characteristics to directly estimate the source position in the compressed measurement domain, includes the following steps:

[0008] S1. Observation signal model and data compression

[0009] Consider a direct positioning system model, which includes L mobile observation stations and 1 target transmitter. The data of the transmitter is intercepted in K observation intervals. The signal r received by the lth observation station in the kth observation interval is l,k (t) is:

[0010]

[0011] Where t is the time parameter, sk (t) is the signal waveform in the kth observation interval, τ l,k is the time delay of the signal arriving at the kth observation interval at the lth observation station, f l,k is the Doppler frequency shift caused by the relative displacement between the observation station and the target, n l,k (t) is the noise received by the observation station.

[0012] N=2 n (n=1,2,3...) order Hadamard matrix (H N ), is an N×N matrix whose elements are only +1 or -1 and whose rows are orthogonal to each other, that is, it satisfies I N is the identity matrix of order N. If the inner product (sum of its own squares) of any row of an N-order matrix is ​​N, and the inner product of any two rows is 0, then the matrix is ​​a Hadamard matrix.

[0013] For example, a 4th order Hadamard matrix is:

[0014]

[0015] Select H N The first M rows of the matrix can be used to obtain an M×N partial Hadamard matrix, which is expressed as H M×N .

[0016] In order to reduce the communication burden, H M×N The original sampled signal is compressed, where N is the original signal length, M is the compressed signal length, and the compression ratio is α = M / N, where M < < N. The compressed signal received by the central receiving station is:

[0017] y l,k =H M×N r l,k +ε l,k

[0018] r l,k =[r l,k (t1),...,r l,k (t N )] T is the collected signal vector, ε l,k is the noise vector generated in the kth communication gap between the mobile observation station l and the central receiving station. Since the linear processing in the above formula does not change the signal-to-noise ratio of the compressed signal, y l,k It can be approximated as:

[0019] y l,k ≈H M×N r l,k

[0020] S2. Compressed signal processing and direct positioning problem model

[0021] The central receiving station multiplies the received compressed signal with the transpose of the partial Hadamard matrix:

[0022]

[0023] (·) T is the transpose, r l ' ,k and r l,k The time shift and frequency shift relationship between reserve.

[0024] The target search range is divided into grids according to the preset steps, and the distance between each grid point and each moving receiving station at each sampling point is determined. The time-shift and frequency-shift matrix of each grid point is determined based on the distance. Then, the cost function value of each grid point is determined based on the time-shift and frequency-shift matrix of each grid point and the received signal of each moving receiving station. Then, the position information of the target radiation source is located based on the cost function of each grid point. The central receiving station can use the processed signal r l ' ,k Calculate the cost function:

[0025]

[0026] in(·) H is the conjugate transpose, p is the coordinate of the grid point, is a matrix containing delay information, diag(·) represents a diagonal matrix with diagonal elements (·), The unit matrix is ​​shifted downward cyclically by floor(Tf l,k )+1 rows, floor(·) means rounding down, T is the observation time in the observation gap, Λ is a diagonal matrix whose main diagonal is the signal spectrum, σ 2 is the noise power, (·) -1 is the inverse matrix.

[0027] The estimation of the target position can be transformed into the maximum likelihood estimation of the cost function, that is:

[0028]

[0029] argmax(·) represents the set of independent variables that makes the function (·) reach its maximum value. is an estimate of the target source location.

[0030] The present invention provides a non-reconstruction direct positioning method based on data compression for mobile receivers. This method utilizes the properties of the Hadamard matrix to estimate the source position directly in the compressed measurement domain, without the need for signal reconstruction and TDOA and FDOA parameter extraction. Under high signal-to-noise ratio conditions, the direct positioning method based on data compression performs comparable to the classical direct positioning method without compression. At reasonable compression rates, the direct positioning method based on data compression also performs well under low signal-to-noise ratio conditions, reducing the amount of transmission by ζ = L(1-α)N. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A flowchart of the process for implementing the present invention;

[0032] Figure 2 Schematic diagram of a non-reconstruction direct positioning system model based on mobile receiver data compression of the present invention;

[0033] Figure 3 A distribution diagram of the fixed transmitting sources, mobile receiving stations, and average positioning points of the present invention;

[0034] Figure 4 Comparison of the grid cost function graphs obtained by the method proposed in the present invention and the original mobile receiving station direct positioning method. (a) is the grid cost function graph obtained by the method proposed in the present invention, and (b) is the grid cost function graph obtained by the original mobile receiving station direct positioning method. DETAILED DESCRIPTION

[0035] The effectiveness of the present invention will be described below with reference to the accompanying drawings and simulation examples.

[0036] Simulation example:

[0037] The feasibility and effectiveness of the non-reconstruction direct positioning method based on mobile receiver data compression are verified through simulation. The example process of the non-reconstruction direct positioning method based on mobile receiver data compression is shown in the attached figure. Figure 1 The schematic diagram of the non-reconstruction direct positioning system model based on mobile receiver data compression is shown in the attached figure. Figure 2 As shown in the example, the distribution diagram of fixed transmitting sources, mobile receiving stations and average positioning points is shown in the attached figure. Figure 3 In this example, the simulation conditions are set to the static transmitter position coordinates (2000m, 2000m), and the transmitted signal is the carrier frequency f c =20MHz, sampling frequency f s=100MHz, BPSK signal with bandwidth of 33.72MHz, 1024 snapshots are collected at each source position, SNR = 0dB, compression ratio α = 1 / 2, and the compressed signal is 512 snapshots. The observation interval time is selected as T = 0.972ms, and the observation interval between the two observation stations is 3.33s; the location distribution is as follows Figure 3 The observation station shown moves at a speed of 300 m / s, and the search area is limited to ±500 m of the actual position. Each receiving station transmits the compressed signal to the central processing station, where it is multiplied by the transpose of the compression matrix to generate 1024 snapshots.

[0038] The comparison diagram of the grid cost function obtained by the method proposed in this invention and the original mobile receiving station direct positioning method is shown in the figure below. Figure 4 As shown. It can be found that the grid cost function graph obtained by the present invention achieves the desired purpose of the grid cost function graph, and the position of the transmitting source can be located by searching for the peak of the cost function. The present invention experiments on different situations where the transmitting source is on the grid point and not on the grid point. The cost function reaches a peak at the transmitting source. Therefore, the grid cost function graph obtained by the present invention meets expectations. The method proposed by the present invention can perform well under high signal-to-noise ratio even at low compression ratio. In addition, under low signal-to-noise ratio conditions, a reasonable high compression ratio value can be selected. Therefore, the method proposed by the present invention can reduce the communication burden while maintaining reasonable positioning accuracy.

Claims

1. A non-reconstruction direct positioning method based on mobile receiver data compression, which uses the Hadamard matrix characteristics to estimate the source position directly in the compressed measurement domain, characterized by: The following steps are involved: S1. Definition: In the direct positioning system model, there are L mobile observation stations and 1 target transmitter. The data of the transmitter is intercepted within K observation intervals. The signal r received by the lth observation station in the kth observation interval is l,k (t) is: Where t is the time parameter, s k (t) is the signal waveform in the kth observation interval, t l,k is the time delay of the signal arriving at the kth observation interval at the lth observation station, f l,k is the Doppler frequency shift caused by the relative displacement between the observation station and the target, n l,k (t) is the noise received by the observing station; Select N = 2 n The Hadamard matrix H of order N The first M rows of the matrix are used to obtain an M×N partial Hadamard matrix, denoted as H M×N , n=1,2,3...; Using H M×N The original sampled signal is compressed, N is the original signal length, M is the compressed signal length, the compression ratio is α = M / N, M < < N, the compressed signal received by the central receiving station is: the l,k =H M×N r l,k +ε l,k r l,k =[r l,k (t1),...,r l,k (t N )] T is the collected signal vector, ε l,k is the noise vector generated in the kth communication gap between the mobile observation station l and the central receiving station. Since the linear processing in the above formula does not change the signal-to-noise ratio of the compressed signal, y l,k Approximately: y l,k ≈H M×N r l,k S2. The central receiving station multiplies the received compressed signal by the transpose of the partial Hadamard matrix: (·) T is the transpose, r′ l,k and r l,k The time shift and frequency shift relationship between reserve; The target search range is divided into grids according to the preset steps, and the distance between each grid point and each moving receiving station at each sampling point is determined. The time-shift and frequency-shift matrix of each grid point is determined based on the distance. Then, the cost function value of each grid point is determined based on the time-shift and frequency-shift matrix of each grid point and the received signal of each moving receiving station. Then, the position information of the target radiation source is located according to the cost function of each grid point. The central receiving station uses the processed signal r′ l,k Calculate the cost function: in(·) H is the conjugate transpose, p is the coordinate of the grid point, is a matrix containing delay information, diag(·) represents a diagonal matrix with diagonal elements (·), The unit matrix is ​​shifted downward cyclically by floor(Tf l,k )+1 rows, floor(·) means rounding down, T is the observation time in the observation gap, Λ is a diagonal matrix whose main diagonal is the signal spectrum, σ 2 is the noise power, (·) -1 is the inverse matrix; The estimation of the target position is converted into the maximum likelihood estimation of the cost function, that is: argmax(·) represents the set of independent variables that makes the function (·) reach its maximum value. is an estimate of the target source location.

Images