Distributed direct positioning method based on optimal interception node selection
By selecting the optimal receiving node in a large-scale distributed reconnaissance system and constructing a distributed direct positioning cost function, the problem of high communication volume and computational overhead caused by dense deployment of receiving nodes is solved, achieving high-precision positioning and improved system efficiency.
Patent Information
- Application Number
- CN202511373203.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-24
- Publication Date
- 2026-01-09
AI Technical Summary
The dense deployment of reconnaissance nodes in large-scale distributed reconnaissance systems leads to high communication and computational overhead, as well as reduced system reconnaissance efficiency.
A distributed direct localization method based on optimal receiver node selection is adopted. By constructing a local CRLB matrix and introducing a selection matrix, an optimal receiver node selection model is established, a distributed direct localization cost function is constructed, and a distributed optimization algorithm is performed to achieve high-precision localization.
It improves the robustness and fault tolerance of the positioning system, reduces communication overhead and computational burden, and enhances positioning accuracy and system reconnaissance efficiency.
Smart Images
Figure CN121299580A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of electronic reconnaissance technology and is applied to distributed reconnaissance systems. Specifically, it relates to a distributed direct location method based on the selection of the optimal reconnaissance node. Background Technology
[0002] Electronic reconnaissance is a key means of acquiring information about enemy radar, communications, and other electronic systems by intercepting, analyzing, and identifying electromagnetic signals. Passive positioning technology is a core method in electronic reconnaissance systems for obtaining the location information of target radiation sources. Compared to active positioning technology, passive positioning technology does not require the active transmission of detection signals, offering advantages such as better electromagnetic concealment and longer reconnaissance range. It is widely used in military and civilian fields such as wireless communication, navigation systems, and search and rescue.
[0003] Traditional passive localization relies on a two-step process. First, it estimates parameters related to the target's location from the intercepted signal, such as the Angle of Arrival (AOA), Time of Arrival (TOA), and Time Difference of Arrival (TDOA). Then, it calculates the location of the radiation source based on geometric relationships. Because of the parameter estimation step, this method introduces error accumulation, and the accuracy of parameter estimation significantly decreases under low signal-to-noise ratio (SNR) conditions, severely impacting the final reconnaissance and localization performance.
[0004] Direct Position Determination (DPD), also known as one-step positioning, eliminates the need for intermediate parameter estimation. It directly constructs a positioning cost function based on the signals intercepted by the receiving nodes and then achieves positioning by solving for the optimal value of the objective function. DPD fully utilizes the correlation between received signals, ensuring effective coherent accumulation and avoiding the propagation of position estimation errors and difficulties in correlating positioning parameters caused by the inaccuracy of intermediate parameter estimation. It offers advantages such as high positioning accuracy, strong resolution, and robustness.
[0005] In practical deployments, the Centralized Direct Position Determination (CDPD) algorithm requires transmitting the raw sample data from all reconnaissance nodes to a fusion center for joint processing. In large-scale reconnaissance systems, the dense deployment of reconnaissance nodes necessitates establishing communication links between each fusion center and a large number of nodes. As the number of communication links increases, the communication overhead and computational load of the fusion center rise sharply; furthermore, if the fusion center fails, the entire reconnaissance and positioning system will fail. Summary of the Invention
[0006] The purpose of this invention is to provide a distributed direct location method based on optimal reconnaissance node selection, which can solve the problems of high communication volume and computational overhead and reduced system reconnaissance efficiency caused by the dense deployment of reconnaissance nodes in current large-scale distributed reconnaissance systems.
[0007] To achieve the above objectives, the present invention employs the following technical solution: A distributed direct location method based on optimal detection node selection includes: In a distributed receiving system, each receiving node collects data from the target signal and constructs a joint received signal matrix. Construct a local CRLB matrix for each receiving node, and incorporate the selection matrix of the receiving node into the local CRLB matrix; by minimizing the local CRLB matrix with the selection matrix incorporated, establish a selection model for selecting multiple optimal receiving nodes from the set of adjacent receiving nodes and solve it to obtain the optimal receiving node corresponding to each receiving node; Construct a distributed direct location cost function based on the optimal detection node; Solve the cost function of distributed direct positioning to obtain the optimal location estimate of the target.
[0008] Furthermore, the process of establishing a selection model for choosing multiple optimal receiver nodes from a set of adjacent receiver nodes is as follows: First, based on the time difference between the arrival of the target signal at different receiving nodes, an arrival time difference vector is constructed; based on the path gain coefficient between the receiving nodes, a path gain coefficient vector is constructed; the arrival time difference vector and the path gain coefficient vector together constitute the first parameter vector, and the Fisher information matrix of the first parameter vector is established. Secondly, a second parameter vector is constructed based on the target position and the path gain coefficient vector. The Fisher information matrix of the second parameter vector is established and the Fisher information matrix of the first parameter vector is substituted into it. After sorting, the CRLB matrix based on the target position of all receiving nodes is obtained. Based on this, the local CRLB matrix of each receiving node in each frame is obtained. Finally, a selection matrix for the receiving nodes is introduced into the local CRLB matrix, and the constant terms are ignored. The selection model is constructed by minimizing the local CRLB matrix as the objective function.
[0009] Furthermore, the Fisher information matrix of the second parameter vector is:
[0010] in, The Fisher information matrix is the first parameter vector; the time difference of arrival vector is the second parameter vector. and path gain coefficient vector Form the first parameter vector From the path gain coefficient vector and target location Construct the second parameter vector .
[0011] Furthermore, the local CRLB matrix of each receiving node is constructed, as follows:
[0012] Among them, local coefficients , , All are constants. Indicates the detection node The noise covariance matrix of all adjacent receiving nodes; The local Jacobian matrix is represented as follows:
[0013] in, The location of the target. Represents the speed of light. For the first The location of each receiving node. , The number of surveillance nodes; Indicates the ability to communicate with the receiving node Communication No. The location of each adjacent receiving node, superscript Indicates matrix transpose; This represents the number of adjacent receiving nodes in the set of adjacent receiving nodes. .
[0014] Furthermore, the selection matrix of the receiving nodes is incorporated into the local CRLB matrix, as follows:
[0015] in, , For detection nodes The selection matrix, For detection nodes The selection vector; if the receiving node Selected the If there are 10 adjacent receiving nodes, then If the first option is not selected If there are 10 adjacent receiving nodes, then .
[0016] Furthermore, by minimizing the local CRLB matrix that introduces the selection matrix, a selection model for choosing multiple optimal receiver nodes from the set of adjacent receiver nodes is established, expressed as:
[0017] in, for A dimensional vector of all 1s. To find the trace of the matrix, Indicates from the detection node The number of optimal receiver nodes selected from the set of adjacent receiver nodes.
[0018] Furthermore, based on the optimal detection node, a distributed direct location cost function is constructed, expressed as:
[0019] in, Indicates the first l The local cost function of each detection node as the fusion center:
[0020] in, Indicates containing the first The delay response matrix of each receiving node and its selected optimal receiving node. Representation matrix The OK; Indicates the first Covariance matrix of each frequency point After eigenvalue decomposition, the noise subspace is formed by the eigenvectors corresponding to the smallest eigenvalue. This represents the number of frequency points.
[0021] Furthermore, the cost function of distributed direct positioning is solved to obtain the optimal location estimate of the target, including: The search area is divided into grids, and the time delay response matrix of each detection node for each grid position is calculated. Based on the time delay response matrix, the value of the local cost function with the detection node as the fusion center is solved, and then the value of the global cost function is obtained. The optimal position of the target is estimated to be the grid position corresponding to the maximum value of the global cost function.
[0022] A terminal device includes a processor, a memory, and a computer program stored in the memory; when the processor executes the computer program, it implements the distributed direct location method based on the selection of the optimal receiving node.
[0023] A computer-readable storage medium storing a computer program; when executed by a processor, the computer program implements the distributed direct location method based on the selection of the optimal receiving node.
[0024] Compared with the prior art, the present invention has the following technical features: This invention employs a distributed direct positioning architecture, which improves the robustness, reconfigurability, and fault tolerance of the positioning system, and expands the application scenarios of distributed reconnaissance systems. Furthermore, the direct positioning method solves the error propagation problem caused by the separation of parameter estimation and location calculation in traditional two-step positioning methods, thus improving the positioning accuracy of targets. Finally, this invention establishes a distributed direct positioning cost function based on an optimal receiver node selection model that minimizes the CRLB matrix trace, and solves for the radiation source location using a grid search method. This achieves high-precision positioning while significantly reducing communication overhead and computational burden. Attached Figure Description
[0025] Figure 1 This is a schematic flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the distribution of receiving nodes in an embodiment of the present invention; Figure 3 This refers to the spatial positioning spectrum in this embodiment of the invention. Figure 4 This is a performance comparison of the received signal-to-noise ratio under the embodiments of the present invention and under the random selection strategy. Detailed Implementation
[0026] This invention provides a Distributed Direct Position Determination (DDPD) method based on optimal receiver node selection. It employs a distributed cooperative receiver positioning framework, establishing an optimal receiver node selection model by optimizing the trace of the Cramér-Rao Lower Bound (CRLB) lower bound, and using this model to establish a distributed direct positioning cost function. Furthermore, a positioning spatial spectrum is generated based on the cost function to estimate target location information. In this architecture, each receiver node constructs a local cost function that fuses target location information based on locally received raw signal data and communication data from neighboring receiver nodes. Global location estimation is achieved through a distributed optimization algorithm. This method effectively reduces system communication overhead and computational burden while achieving high-precision positioning, thus improving system reconnaissance efficiency.
[0027] This invention provides a distributed direct location method based on optimal detection node selection, comprising the following steps: Step 1: In the distributed receiving system, each receiving node collects data from the target signal and constructs a joint received signal matrix.
[0028] In this distributed receiving system, each receiving node is equipped with a single antenna; the target may be, for example, a radiation source.
[0029] Assume it exists Each receiving node receives the target signal, and the target's two-dimensional position is... , No. The location of each detection node is Then the first The received signal of each receiving node can be described as follows: (1) in, For the first The channel parameters of the receiving node represent the distance from the target to the receiving node. The complex channel fading factor of each receiving node for The target signal at that moment, To achieve the goal of Signal transmission delay of each receiving node; Indicates the first Individual detection nodes Noise observed at any time, parameters This represents the sampling duration for each signal reception; superscript. T This indicates the matrix transpose, and the same applies below.
[0030] The target signal arrived at the latency of each detection node relative to the target position This information reflects the distance between the target and the receiving node. (2) in Represents the L2 norm, It represents the speed of light.
[0031] Perform on the received signal If point sampling is performed and a discrete Fourier transform is applied, then the th... The first detection node at the n The received signal at each frequency point is represented as: (3) in, Indicates the first One frequency component, It is a natural constant. Represents the imaginary unit. , They are respectively , The Each frequency point is sampled as a signal.
[0032] Then for the first Each receiving node, all Received signal at each frequency point It can be represented as: (4) in, For noise vectors, , Represents the diagonalization operator. The phase delay in the frequency domain is defined as: (5) Considering all detection nodes For each sampling point, the following expression can be constructed to cover all the data: (6) (7) (8) in, This represents the joint received signal matrix of all receiving nodes. The time-domain response matrix generated by each receiving node to the target. For the noise matrix, for The complex space.
[0033] The joint received signal from all receiving nodes can then be represented as a matrix: (9) Step 2: Construct the local CRLB matrix for each receiving node, and introduce the selection matrix of the receiving node into the local CRLB matrix; by minimizing the local CRLB matrix with the selection matrix introduced, establish a selection model for selecting multiple optimal receiving nodes from the set of adjacent receiving nodes and solve it to obtain the optimal receiving node corresponding to each receiving node.
[0034] The target signal arrived at the The first detection node and the first The arrival time difference between the receiving nodes is: (10) in, , These represent the arrival times of the target signal at the [number]th [number]. , The latency of each receiving node.
[0035] make Indicates the reference detection node. Then the vector consisting of the arrival time differences. It can be represented as: (11) Definition of the first The detection node is relative to the first Path gain coefficient of each receiving node for: (12) in, , They represent the target to the number of... , The multiple channel fading factor of each receiving node.
[0036] make , Then the vector composed of path gain coefficients Represented as: (13) From the time difference of arrival vector and path gain coefficient vector Form the first parameter vector Among them, the position relative to the target The relevant unknown is the arrival time difference vector. First parameter vector The Fisher information matrix is as follows: (14) Among them, coefficient , and They are all constants between 0 and 1. Represents the diagonalization operator. This represents finding the inverse of a matrix. This represents the noise covariance matrix of all receiving nodes.
[0037] To derive the lower bound of the Cramérault position for the target location, a new second parameter vector is constructed. Assuming all receiving nodes have the same attenuation coefficient of 1, the path gain coefficient is: , for The identity matrix; the second parameter vector The Fisher information matrix is as follows: (15) in: (16) in, To find the partial derivative operator, for A matrix of all zeros Let Jacobian matrix represent the target position. Location of the reconnaissance node The geometric relationship between them is expressed as follows: (17) in, Represents the speed of light. For the first The location of each receiving node. .
[0038] Substituting Equations 14, 16, and 17 into Equation 15 and rearranging, we obtain the CRLB matrix based on the target positions of all receiving nodes (this matrix is equivalent to the information matrix in Equation 15): (18) In distributed direct location, only the receiving nodes are used. Radiation source estimation is performed on the received signals of the receiver and its adjacent receiver nodes; then the receiver nodes The local CRLB matrix is represented as: (19) Among them, local coefficients , , All are constants between 0 and 1. Indicates the detection node The noise covariance matrix of all adjacent receiving nodes; denoted by the noise covariance matrix of the receiving nodes. Communication No. The location of each detection node is , The local Jacobian matrix is expressed as follows: (20) in, Indicates the ability to communicate with the first The set of neighboring receiver nodes for communication between receiver nodes The number of adjacent receiving nodes in the system .
[0039] Based on this, the receiving nodes selected by adjacent receiving nodes are derived. The local CRLB matrix is: (twenty one) in, For detection nodes The selection matrix, For detection nodes The selection vector.
[0040] (twenty two) Among them, if the receiving node Selected the If there are 10 adjacent receiving nodes, then If the first option is not selected If there are 10 adjacent receiving nodes, then .
[0041] neglect constant term in Construct from the detection node The set of adjacent receiving nodes In the middle, select The optimal detection node selection model: (twenty three) in, for A dimensional vector of all 1s. To find the trace of the matrix.
[0042] By solving the above selection model, we can obtain the results from the set of adjacent receiving nodes. China has determined The optimal detection node.
[0043] Step 3: Construct a distributed direct location cost function based on the optimal receiving node.
[0044] By the The received signals of each receiving node and its selected optimal receiving node are defined as follows: (twenty four) in, The representation is defined as follows: For the first The optimal set of receiving nodes selected by each receiving node, this set contains Adjacent detection nodes.
[0045] Including the The time delay response matrix of each receiving node and its selected optimal receiving node is represented as follows: (25) in, , For the first Channel parameters and phase delay of each receiving node.
[0046] No. The expression for calculating the covariance matrix of each frequency point is as follows: (26) in Indicates the first At the frequency point The received signals of each receiving node and its selected optimal receiving node, i.e., the matrix The Okay, superscript This indicates the conjugate transpose, and the same applies below.
[0047] Based on the Information about the first receiving node and its selected optimal receiving node is used to construct a system based on the information of the first receiving node. l The local cost function of each detection node as the fusion center : (27) in Indicates the first Detection nodes on each frequency point The time delay response of the selected optimal receiving node, i.e., the time delay response matrix. The OK; Indicates the first Covariance matrix of each frequency point After eigenvalue decomposition, the noise subspace is formed by the eigenvectors corresponding to the smallest eigenvalue.
[0048] After selecting the optimal receiving node, the global cost function of the distributed direct location algorithm It can be represented as: (28) Step 4: Solve the cost function of the distributed direct positioning algorithm to obtain the optimal location estimate of the target.
[0049] The target location can be determined as the maximum value of the cost function: (29) in The variable representing the maximum value of the global cost function. .
[0050] Solving using the two-dimensional grid search method can make the cost function... The optimal estimate of the target location can be obtained by finding the position where the maximum value is reached; this is achieved by dividing the search area where the target may exist into... For each of the grid points, calculate the relationship between each detection node and the first grid point. grid locations Delay response matrix , And solve according to equation (27). Substituting it into equation (28) yields For all The optimal location estimate of the target at grid points can be expressed as: (30) Example: In one embodiment of the present invention, it is assumed that the invention utilizes... Distributed detection nodes receive target signals, and their spatial locations are as follows: , , , , , , , , , , , The coordinates of the radiation source are The signal sampling time for each observation is Number of target signal sampling points From the detection node Select from adjacent receiving nodes The optimal detection nodes are divided into regions of interest as follows: Grid points.
[0051] The settings and related results of this embodiment are as follows: Figures 2 to 4 As shown, by comparing with existing methods, this scheme effectively improves the target positioning accuracy while significantly reducing communication overhead and computational burden.
[0052] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A distributed direct location method based on optimal receiver node selection, characterized in that, include: In a distributed receiving system, each receiving node collects data from the target signal and constructs a joint received signal matrix. Construct a local CRLB matrix for each receiving node, and incorporate the selection matrix of the receiving node into the local CRLB matrix; by minimizing the local CRLB matrix with the selection matrix incorporated, establish a selection model for selecting multiple optimal receiving nodes from the set of adjacent receiving nodes and solve it to obtain the optimal receiving node corresponding to each receiving node; Construct a distributed direct location cost function based on the optimal detection node; Solve the cost function of distributed direct positioning to obtain the optimal location estimate of the target.
2. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, The process of establishing a selection model for choosing multiple optimal receiver nodes from a set of adjacent receiver nodes is as follows: First, based on the time difference between the arrival of the target signal at different receiving nodes, an arrival time difference vector is constructed; based on the path gain coefficient between the receiving nodes, a path gain coefficient vector is constructed; the arrival time difference vector and the path gain coefficient vector together constitute the first parameter vector, and the Fisher information matrix of the first parameter vector is established. Secondly, a second parameter vector is constructed based on the target position and the path gain coefficient vector. The Fisher information matrix of the second parameter vector is established and the Fisher information matrix of the first parameter vector is substituted into it. After sorting, the CRLB matrix based on the target position of all receiving nodes is obtained. Based on this, the local CRLB matrix of each receiving node in each frame is obtained. Finally, a selection matrix for the receiving nodes is introduced into the local CRLB matrix, and the constant terms are ignored. The selection model is constructed by minimizing the local CRLB matrix as the objective function.
3. The distributed direct location method based on optimal detection node selection according to claim 2, characterized in that, The Fisher information matrix of the second parameter vector is: in, The Fisher information matrix is the first parameter vector; the time difference of arrival vector is the second parameter vector. and path gain coefficient vector Form the first parameter vector From the path gain coefficient vector and target location Construct the second parameter vector .
4. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, Construct the local CRLB matrix for each receiving node, as follows: Among them, local coefficients , , All are constants. Indicates the detection node The noise covariance matrix of all adjacent receiving nodes; The local Jacobian matrix is represented as follows: in, The location of the target. Represents the speed of light. For the first The location of each receiving node. , The number of surveillance nodes; Indicates the ability to communicate with the receiving node No. of communication The location of each adjacent receiving node, superscript Indicates matrix transpose; This represents the number of adjacent receiving nodes in the set of adjacent receiving nodes. .
5. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, The selection matrix of the receiving nodes is incorporated into the local CRLB matrix, and expressed as follows: in, , For detection nodes The selection matrix, For detection nodes The selection vector; if the receiving node Selected the If there are 10 adjacent receiving nodes, then If the first option is not selected If there are 10 adjacent receiving nodes, then .
6. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, By minimizing the local CRLB matrix that introduces the selection matrix, a selection model for choosing multiple optimal receiver nodes from a set of adjacent receiver nodes is established, expressed as: in, for A dimensional vector of all 1s. To find the trace of the matrix, Indicates from the detection node The number of optimal receiver nodes selected from the set of adjacent receiver nodes.
7. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, Based on the optimal detection node, a distributed direct location cost function is constructed, expressed as: in, Indicates the first l The local cost function of each detection node as the fusion center: in, Indicates containing the first The delay response matrix of each receiving node and its selected optimal receiving node. Representation matrix The OK; Indicates the first Covariance matrix of each frequency point After eigenvalue decomposition, the noise subspace is formed by the eigenvectors corresponding to the smallest eigenvalue. This represents the number of frequency points.
8. The distributed direct location method based on optimal detection node selection according to claim 1, characterized in that, Solving the cost function of distributed direct positioning yields the optimal location estimate of the target, including: The search area is divided into grids, and the time delay response matrix of each detection node for each grid position is calculated. Based on the time delay response matrix, the value of the local cost function with the detection node as the fusion center is solved, and then the value of the global cost function is obtained. The optimal position of the target is estimated to be the grid position corresponding to the maximum value of the global cost function.
9. A terminal device, comprising a processor, a memory, and a computer program stored in the memory; characterized in that, When the processor executes a computer program, it implements the distributed direct location method based on the selection of the optimal receiving node as described in any one of claims 1-8.
10. A computer-readable storage medium storing a computer program; characterized in that, When the computer program is executed by a processor, it implements the distributed direct location method based on the selection of the optimal receiving node as described in any one of claims 1-8.