CN110632584B - Passive target external radiation source positioning method based on parallel radial basis network - Google Patents

Passive target external radiation source positioning method based on parallel radial basis network Download PDF

Info

Publication number
CN110632584B
CN110632584B CN201910892008.6A CN201910892008A CN110632584B CN 110632584 B CN110632584 B CN 110632584B CN 201910892008 A CN201910892008 A CN 201910892008A CN 110632584 B CN110632584 B CN 110632584B
Authority
CN
China
Prior art keywords
training
radial basis
network
training set
received signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910892008.6A
Other languages
Chinese (zh)
Other versions
CN110632584A (en
Inventor
程永强
吴昊
王宏强
陈茜茜
杨政
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
National University of Defense Technology
Original Assignee
National University of Defense Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by National University of Defense Technology filed Critical National University of Defense Technology
Priority to CN201910892008.6A priority Critical patent/CN110632584B/en
Publication of CN110632584A publication Critical patent/CN110632584A/en
Application granted granted Critical
Publication of CN110632584B publication Critical patent/CN110632584B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
    • G01S13/06—Systems determining position data of a target
    • G01S13/46—Indirect determination of position data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
    • G01S13/06—Systems determining position data of a target
    • G01S13/46—Indirect determination of position data
    • G01S2013/466—Indirect determination of position data by Trilateration, i.e. two antennas or two sensors determine separately the distance to a target, whereby with the knowledge of the baseline length, i.e. the distance between the antennas or sensors, the position data of the target is determined

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention provides a target external radiation source passive positioning method based on a parallel radial basis network. The mapping relation between the target position and the received signal is learned through training a group of radial basis networks connected in parallel and through training of sample data, and passive positioning of the target external radiation source is achieved. The method provided by the invention does not depend on radiation source information in an electromagnetic environment, can realize target positioning only by training data, and can solve the problem that the radial basis network is not suitable for the radial basis network fitting in a training set. The method has the characteristics of simple positioning process and high efficiency, has good positioning performance, and has the potential of completing target positioning by utilizing the communication base station.

Description

Passive target external radiation source positioning method based on parallel radial basis network
Technical Field
The invention relates to the technical field of radar target positioning, in particular to a passive target positioning method by using an external radiation source.
Background
The passive positioning of the target external radiation source is different from the traditional radar positioning technology, the target is positioned by means of a third-party radiation source, and the passive positioning method has the advantages of low cost, interference resistance, strong hiding capability and the like, and has wide application prospect. Nowadays, the coverage range of external radiation sources is wide, such as mobile communication base stations, television broadcast signal base stations, navigation satellites and the like, and favorable objective conditions are provided for the development of passive positioning technologies of the external radiation sources. Therefore, the passive target external radiation source positioning technology has multiple application scenes, wide deployment of external radiation sources and great development potential.
In the development of passive target positioning technology of external radiation source, the passive target positioning method of external radiation source is the key technology. Existing passive localization methods typically rely on information such as the position of the external radiation source or the signal waveform. However, when such information is not accurately obtained, positioning cannot be achieved using existing methods, and it is necessary to study a positioning method that relies only on received signal data.
Disclosure of Invention
Aiming at the problems in the prior art, the invention aims to provide a target external radiation source passive positioning method based on a parallel radial basis network.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
a passive positioning method of a target external radiation source based on a parallel radial basis network comprises the following steps
Firstly, training a target external radiation source passive positioning network;
(1.1) acquiring a target position-received signal data pair of the region of interest.
(1.2) processing the target position-received signal data pair, and acquiring a feature vector of the received signal data and a data pair of the target position as training data to form a total training set;
and (1.3) dividing the total training set into two sub-training sets through the characteristic vector norm, and training respective radial basis networks by utilizing the divided two sub-training sets.
And (1.4) taking the output values corresponding to the two sub-training sets obtained in the step (1.3) and the target position coordinates in the total training set as input and output to form a new training set, and training the cascade reverse propagation network to combine the output of the dual-radial basis network. And the trained cascade reverse propagation network is the final target external radiation source passive positioning network.
And secondly, calculating a characteristic vector of received signal data received in real time, and inputting the characteristic vector into the passive positioning network of the target external radiation source obtained by the training in the first step to obtain the target position.
In the invention (1.1), the region of interest is uniformly divided into N units, the standard scattering balls are sequentially placed in the center of each unit, and the position coordinates of the N standard scattering balls are recorded as X ═ X1,x2,...,xN}。
When the standard scattering ball is placed in the ith unit, the position coordinate x of the ith standard scattering ball is recordediReceived signal data with M receivers is(s)i,1,...,si,M) And forming a target position-received signal data pair, wherein i is more than or equal to 1 and less than or equal to N.
The position coordinates of the N standard scattering balls and the received signal data of the M receivers are S:
S＝{(s1,1,...,s1,M),(s2,1,...,s2,M),...,(sN,1,...,sN,M)}。
in the invention (1.2), for received signal data(s)i,1,…,si,M) Each s ini,jCalculating a first moment and a second moment, wherein j is more than or equal to 1 and less than or equal to M; reducing the dimension of training data by utilizing the statistical divergence of the received signal data and the noise statistical moment of the receiver, and acquiring a characteristic vector of the received signal data; and taking the data pairs of the feature vectors and the target positions as training data to form a total training set.
For each received signal data
Figure BDA0002209034770000021
Wherein l is the data length, and the first moment and the second moment are calculated. In order to save computing resources and storage resources, the first order moment and the second order moment are both reduced to LcAnd (5) maintaining.
Figure BDA0002209034770000031
Figure BDA0002209034770000032
Wherein,
Figure BDA0002209034770000033
superscript denotes conjugating the complex number. p is an index in the summation formula, where represents the pair si,jListing each item in (a).
The receiver measures the noise variance sigma, and KL (Kullback-Leibler) divergence (KLD) between the received signal data and the noise of the receiver is obtained as a characteristic vector
Figure BDA0002209034770000034
Wherein I represents an identity matrix and KLD is calculated by
Figure BDA0002209034770000035
The superscript H of the vector represents the Hermite transpose, tr (-) represents the trace of the matrix, | - | represents the determinant of the matrix, and ln (-) represents the natural logarithm function.
Obtaining a feature vector gamma of received signal dataiAnd target position xkAs training data, form a total training set { (γ)1,x1),(γ2,x2),…,(γN,xN)}。
In the invention (1.3), since the feature vector grows sharply, the components in the feature vector are first mathematically transformed
g(γi)＝10log10γi
Slowing down the growth rate of the feature vector.
Get the total training set after transformation E { (g (γ) { (r) }1),x1),(g(γ2),x2),…,(g(γN),xN) Dividing the total training set into two sub-training sets E1And E2。
Establishing two radial basis function neural networks which are respectively a sub-training set E1Corresponding radial basis function neural network and sub-training set E2A corresponding radial basis function neural network.
Set of child training sets E1The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp1(ii) a In the sub-training set E1Under, use wsp1Training child training set E1Corresponding radial basis function neural network, output
Figure BDA0002209034770000041
Set of child training sets E2Radial basis function of corresponding radial basis function neural networkHas a width parameter of wsp2(ii) a In the sub-training set E2Under, use wsp2Training child training set E2Corresponding radial basis function neural network, output
Figure BDA0002209034770000042
The radial basis function neural network established by the invention consists of three layers. The first layer is input layer with M points, the hidden layer with N points, and the output layer with two points. Wherein, the connection state of full connection is between two adjacent layers, the input layer and the output layer are linear layers, and the activation function of the hidden layer is
Figure BDA0002209034770000043
Wherein, wspIs the width parameter of the radial basis function.
In the invention (1.4), according to
Figure BDA0002209034770000044
And
Figure BDA0002209034770000045
construct new training set
Figure BDA0002209034770000046
And establishing a cascade reverse propagation network, and training the cascade reverse propagation network by using a training set E'.
The structure of the cascaded counter-propagating network established therein is shown in fig. 2. It comprises an input layer, an output layer and DNThe layer is hidden from the layer. The input layer comprises four nodes, the output layer comprises two nodes, and each hidden layer comprises HNAnd (4) each node. Wherein D isNAnd HNThe value of (c) depends on the situation. The input layer and the output layer are linear layers, and the activation function of the hidden layer is
Figure BDA0002209034770000051
In the second step of the invention, for the real-time receiving signal of the receiver, firstly, the characteristic vector gamma of the receiving signal is calculated, and after the function transformation g (gamma), the characteristic vector gamma is input into the cascade reverse propagation network obtained by training, and then the target position coordinate can be obtained.
A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the parallel radial basis network based passive localization method of target external radiation sources.
A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for passive localization of target external radiation sources based on a parallel radial basis network.
The invention has the beneficial technical effects that:
the relationship between the target position and the received signal can be described by a mathematical mapping, which is essentially obtained by the positioning of the target, from which the target position is deduced from the received signal. The data-driven method is different from the traditional method based on a signal model, and the mapping relation can be obtained directly through the training of data.
The invention utilizes a group of radial basis networks connected in parallel, learns the mapping relation between the target position and the received signal through the training of sample data, and realizes the passive positioning of the target external radiation source. The method provided by the invention does not depend on radiation source information, and can realize target positioning only through data training.
Drawings
FIG. 1 is a schematic diagram of a passive target external radiation source positioning method based on a parallel radial basis network in one embodiment;
FIG. 2 is a schematic diagram of the structure of a cascaded backscatter network used in one embodiment;
fig. 3 is a test error chart of the method of the present invention when the cascaded counter-propagation network is 10 layers.
Detailed Description
In order to facilitate the practice of the invention, further description is provided below with reference to specific examples.
The invention provides a target external radiation source passive positioning method based on a parallel radial basis network, which adopts a data driving method to realize a target external radiation source passive positioning technology. Specifically, the problem that the radial basis network is not suitable for the training set but the radial basis network is fitted can be solved by training the dual-radial basis network. The method has the characteristics of simple positioning process and high efficiency, and has good positioning performance. In addition, the method does not depend on radiation source information in an electromagnetic environment, can complete positioning only by receiving signals, and has the potential of completing target positioning by utilizing a communication base station.
Referring to fig. 1, a schematic diagram of a passive positioning method of a target external radiation source based on a parallel radial basis network in an embodiment is shown, and the present invention is described in detail below with reference to specific embodiments.
Firstly, training a target external radiation source passive positioning network;
(1.1) acquiring a target position-received signal data pair of the region of interest.
Uniformly dividing the region of interest into N units, sequentially placing standard scattering balls in the center of each unit, and recording the position coordinates of the N standard scattering balls as X ═ X1,x2,…,xN}。
When the standard scattering ball is placed in the ith unit, the position coordinate x of the ith standard scattering ball is recordediReceived signal data with M receivers is(s)i,1,…,si,M) And forming a target position-received signal data pair, wherein i is more than or equal to 1 and less than or equal to N.
The position coordinates of the N standard scattering balls and the received signal data of the M receivers are S:
S＝{(s1,1,…,s1,M),(s2,1,...,s2,M),...,(sN,1,…,sN,M)}。
and (1.2) processing the target position-received signal data pair, and acquiring a feature vector of the received signal data and a data pair of the target position as training data to form a total training set.
For received signal data(s)i,1,...,si,M) Each s ini,jAnd calculating a first moment and a second moment, wherein j is more than or equal to 1 and less than or equal to M. Reducing the dimension of training data by utilizing the statistical divergence of the received signal data and the noise statistical moment of the receiver, and acquiring a characteristic vector of the received signal data; and taking the data pairs of the feature vectors and the target positions as training data to form a training set.
For each received signal data
Figure BDA0002209034770000071
Calculating the first and second moments
Figure BDA0002209034770000072
Figure BDA0002209034770000073
Wherein,
Figure BDA0002209034770000074
superscript denotes conjugating the complex number. p is an index in the summation formula, where represents the pair si,jListing each item in (a). The receiver measures the noise variance sigma, and KL (Kullback-Leibler) divergence (KLD) between the received signal data and the noise of the receiver is obtained as a characteristic vector
Figure BDA0002209034770000075
Wherein I represents an identity matrix and KLD is calculated by
Figure BDA0002209034770000081
The superscript H of the vector represents the Hermite transpose, tr (-) represents the trace of the matrix, | - | represents the determinant of the matrix, and ln (-) represents the natural logarithm function.
Obtaining a feature vector gamma of received signal dataiAnd target position xkAs training data, form a total training set { (γ)1,x1),(γ2,x2),…,(γN,xN)}。
And (1.3) dividing the total training set through the characteristic vector norm, and training respective radial basis networks by utilizing the divided two sub-training sets.
Since the feature vector grows dramatically, the components in the feature vector are first mathematically transformed
g(γi)＝10log10γi
Slowing down the growth rate of the feature vector.
Get the total training set after transformation E { (g (γ) { (r) }1),x1),(g(γ2),x2),…,(g(γN),xN) Dividing the total training set into two sub-training sets E1And E2。
Two sub-training sets E1And E2The dividing method of (2) is not limited, and for example, the dividing method can be randomly divided according to a certain proportion. The partitioning method adopted in the embodiment is as follows: will satisfy g (gamma)k)＞gb(g (γ) ofk),xk) Division into E1And others to E2，gbThe selection of (a) is determined empirically.
A radial basis function neural network is established, which is composed of three layers. The first layer is input layer with M points, the hidden layer with N points, and the output layer with two points. Wherein, the connection state of full connection is between two adjacent layers, the input layer and the output layer are linear layers, and the activation function of the hidden layer is
Figure BDA0002209034770000082
Wherein, wspIs the width parameter of the radial basis function.
The invention establishes two radial basis function neural networks which are respectively a sub-training set E1Corresponding radial basis function neural network and sub-training set E2A corresponding radial basis function neural network. Set of child training sets E1The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp1(ii) a In the sub-training set E1Under, use wsp1Training child training set E1Corresponding radial basis function neural network, output
Figure BDA0002209034770000091
Set of child training sets E2The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp2(ii) a In the sub-training set E2Under, use wsp2Training child training set E2Corresponding radial basis function neural network, output
Figure BDA0002209034770000092
And (1.4) taking the output values corresponding to the two sub-training sets obtained in the previous step and the target position coordinates in the total training set as input and output to form a new training set, and training the cascade reverse propagation network to combine the output of the dual-radial basis network. And the trained cascade reverse propagation network is the final target external radiation source passive positioning network.
According to
Figure BDA0002209034770000093
And
Figure BDA0002209034770000094
construct new training set
Figure BDA0002209034770000095
And establishing a cascade reverse propagation network, and training the cascade reverse propagation network by using a training set E'.
The established cascaded counter-propagating network structure is shown in fig. 2. It comprises an input layer, an output layer and DNThe layer is hidden from the layer. The input layer comprises four nodes, the output layer comprises two nodes, and each hidden layer comprises HNAnd (4) each node. Wherein D isNAnd HNThe value of (c) depends on the situation. The output input layer and the output layer are linear layers, and the activation function of the hidden layer is
Figure BDA0002209034770000096
And secondly, for real-time received signals of the receiver, firstly calculating a characteristic vector gamma of the received signals, and inputting the characteristic vector gamma into a cascade reverse propagation network obtained by training after function transformation g (gamma) is carried out, so that the target position coordinates can be obtained.
The effectiveness of the method is verified through simulation experiments. The positioning scene is a 500m x 500m square area containing five radiation sources and four receivers. Wherein, the radiation source information is as follows:
Figure BDA0002209034770000101
the positions of the four receivers are (450 ), (450,1050), (1050,450), (1050), the receiver thermal noise power is-80 dBm, and the radiation source signal power is 10 dB. Signal length L10000, feature vector dimension Lc21. The size of the training set is 128 × 128, and the coordinate data in the training set is uniformly distributed in the positioning scene. The test set was 10000 in size and was randomly distributed in the positioning scenario. For the cascaded back-propagation network, two scales are considered, respectively DN＝5,HN10 and DN＝10,HN10. The mean square error and the maximum square error for a single radial basis network and both network systems are shown in the following table.
Figure BDA0002209034770000102
When the cascaded counter-propagation network is 10 layers, the test error of the method provided by the invention is shown in fig. 3.
The experimental result shows that the method provided by the invention has better positioning precision, and the mean square error and the maximum square error are reduced by a certain order of magnitude. And when the number of layers of the cascade reverse propagation network is increased, the positioning precision is gradually improved, and the method has obvious improvement compared with the method only using a single radial basis network.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (7)

1. A target external radiation source passive positioning method based on a parallel radial basis network is characterized by comprising the following steps:
firstly, training a target external radiation source passive positioning network;
(1.1) acquiring a target location-received signal data pair of the region of interest;
uniformly dividing the region of interest into N units, sequentially placing standard scattering balls in the center of each unit, and recording the position coordinates of the N standard scattering balls as X ═ X1,x2,...,xN}；
When the standard scattering ball is placed in the ith unit, the position coordinate x of the ith standard scattering ball is recordediReceived signal data with M receivers is(s)i,1,...,si,M) Forming a target position-received signal data pair, wherein i is more than or equal to 1 and less than or equal to N;
the position coordinates of the N standard scattering balls and the received signal data of the M receivers are S:
S＝{(s1,1,…,s1,M),(s2,1,…,s2,M),...,(sN,1,…,sN,M)}
(1.2) processing the target position-received signal data pair, and acquiring a feature vector of the received signal data and a data pair of the target position as training data to form a total training set;
for received signal data(s)i1,...,siM) Each of which receives signal data
Figure FDA0003000812290000011
Wherein l is the data length, j is more than or equal to 1 and less than or equal to M, and the first moment and the second moment are calculated:
Figure FDA0003000812290000012
Figure FDA0003000812290000013
wherein,
Figure FDA0003000812290000021
superscript denotes conjugating the complex number;
the receiver is set to measure the noise variance sigma, and KL divergence between the received signal data of the receiver and noise is obtained as a feature vector:
Figure FDA0003000812290000022
wherein I represents an identity matrix and KLD is calculated by
Figure FDA0003000812290000023
The superscript H of the vector represents the Hermite transposition, tr (-) represents the trace of the matrix, | - | represents the determinant of the matrix, and ln (-) represents the natural logarithm function;
obtaining a feature vector gamma of received signal dataiAnd target position xkAs training data, form a total training set { (γ)1,x1),(γ2,x2),...,(γN,xN)}；
(1.3) dividing the total training set into two sub-training sets, and training respective radial basis networks by using the two sub-training sets;
(1.4) taking the output values corresponding to the two sub-training sets obtained in the step (1.3) and the target position coordinates in the total training set as input and output to form a new training set, training a cascade reverse propagation network, and combining the output of the biradial basis network; the trained cascade reverse propagation network is the final target external radiation source passive positioning network;
and secondly, calculating a characteristic vector of received signal data received in real time, and inputting the characteristic vector into the passive positioning network of the target external radiation source obtained by the training in the first step to obtain the target position.
2. The passive target external radiation source positioning method based on the parallel radial basis network as claimed in claim 1, wherein in (1.3), the components in the feature vector are first mathematically transformed
g(γi)＝10log10γi
Get the total training set after transformation E { (g (γ) { (r) }1),x1),(g(γ2),x2),...,(g(γN),xN) Dividing the total training set into two sub-training sets E1And E2；
Establishing two radial basis function neural networks which are respectively a sub-training set E1Corresponding radial basis function neural network and sub-training set E2A corresponding radial basis function neural network;
set of child training sets E1The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp1(ii) a In the sub-training set E1Under, use wsp1Training child training set E1Corresponding radial basis function neural network, output
Figure FDA0003000812290000031
Set of child training sets E2The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp2(ii) a In the sub-training set E2Under, use wsp2Training child training set E2Corresponding radial basis function neural network, output
Figure FDA0003000812290000032
3. The method for passively positioning target external radiation sources based on parallel radial basis networks according to claim 2, wherein in (1.4), the method is based on
Figure FDA0003000812290000033
And
Figure FDA0003000812290000034
construct new training set
Figure FDA0003000812290000035
And establishing a cascade reverse propagation network, and training the cascade reverse propagation network by using a training set E'.
4. A training method of a passive positioning network of a target external radiation source is characterized by comprising the following steps
(1.1) acquiring a target location-received signal data pair of the region of interest;
uniformly dividing the region of interest into N units, sequentially placing standard scattering balls in the center of each unit, and recording the position coordinates of the N standard scattering balls as X ═ X1,x2,...,xN}；
When the standard scattering ball is placed in the ith unit, the position coordinate x of the ith standard scattering ball is recordediReceived signal data with M receivers is(s)i,1,...,si,M) Forming a target position-received signal data pair, wherein i is more than or equal to 1 and less than or equal to N;
the position coordinates of the N standard scattering balls and the received signal data of the M receivers are S:
S＝{(s1,1,...,s1,M),(s2,1,...,s2,M),...,(sN,1,...,sN,M)}
(1.2) processing the target position-received signal data pair, and acquiring a feature vector of the received signal data and a data pair of the target position as training data to form a total training set;
for received signal data(s)i,1,...,si,M) Each of which receives signal data
Figure FDA0003000812290000041
Wherein l is the data length, j is more than or equal to 1 and less than or equal to M, and the first moment and the second moment are calculated:
Figure FDA0003000812290000042
Figure FDA0003000812290000043
wherein,
Figure FDA0003000812290000044
superscript denotes conjugating the complex number;
the receiver is set to measure the noise variance sigma, and KL divergence between the received signal data of the receiver and noise is obtained as a feature vector:
Figure FDA0003000812290000051
wherein I represents an identity matrix and KLD is calculated by
Figure FDA0003000812290000052
The superscript H of the vector represents the Hermite transposition, tr (-) represents the trace of the matrix, | - | represents the determinant of the matrix, and ln (-) represents the natural logarithm function;
obtaining a feature vector gamma of received signal dataiAnd target position xkAs training data, form a total training set { (γ)1,x1),(γ2,x2),...,(γN,xN)}；
(1.3) dividing the total training set into two sub-training sets through the characteristic vector norm, and training respective radial basis networks by using the two divided sub-training sets;
and (1.4) taking the output values corresponding to the two sub-training sets obtained in the step (1.3) and the target position coordinates in the total training set as input and output to form a new training set, training the cascade reverse propagation network to combine the output of the dual-radial basis network, wherein the cascade reverse propagation network obtained by training is the final target external radiation source passive positioning network.
5. The method for training the passive target external radiation source positioning network according to claim 4,
(1.3) first, the components in the feature vector are mathematically transformed
g(γi)＝10log10γi
Get the total training set after transformation E { (g (γ) { (r) }1),x1),(g(γ2),x2),...,(g(γN),xN) Dividing the total training set into two sub-training sets E1And E2；
EstablishingTwo radial basis function neural networks, each being a sub-training set E1Corresponding radial basis function neural network and sub-training set E2A corresponding radial basis function neural network;
set of child training sets E1The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp1(ii) a In the sub-training set E1Under, use wsp1Training child training set E1Corresponding radial basis function neural network, output
Figure FDA0003000812290000061
Set of child training sets E2The width parameter of the radial basis function of the corresponding radial basis function neural network is wsp2(ii) a In the sub-training set E2Under, use wsp2Training child training set E2Corresponding radial basis function neural network, output
Figure FDA0003000812290000062
(1.4) according to
Figure FDA0003000812290000063
And
Figure FDA0003000812290000064
construct new training set
Figure FDA0003000812290000065
And establishing a cascade reverse propagation network, and training the cascade reverse propagation network by using a training set E'.
6. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor when executing the computer program performs the steps of the parallel radial basis network based passive localization method for target external radiation sources of any of claims 1 to 3.
7. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the parallel radial basis network-based passive localization method of an external source of radiation of an object of any of claims 1 to 3.
CN201910892008.6A 2019-09-20 2019-09-20 Passive target external radiation source positioning method based on parallel radial basis network Active CN110632584B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910892008.6A CN110632584B (en) 2019-09-20 2019-09-20 Passive target external radiation source positioning method based on parallel radial basis network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910892008.6A CN110632584B (en) 2019-09-20 2019-09-20 Passive target external radiation source positioning method based on parallel radial basis network

Publications (2)

Publication Number Publication Date
CN110632584A CN110632584A (en) 2019-12-31
CN110632584B true CN110632584B (en) 2021-05-14

Family

ID=68971996

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910892008.6A Active CN110632584B (en) 2019-09-20 2019-09-20 Passive target external radiation source positioning method based on parallel radial basis network

Country Status (1)

Country Link
CN (1) CN110632584B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112418245B (en) * 2020-11-04 2024-04-26 武汉大学 Electromagnetic emission point positioning method based on urban environment physical model

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101893704B (en) * 2010-07-20 2012-07-25 哈尔滨工业大学 Rough set-based radar radiation source signal identification method
RU2560934C1 (en) * 2014-06-30 2015-08-20 Общество с ограниченной ответственностью "Интеллектуальные радиооптические системы" Method of real time reception of radar information and system for its implementation
CN107730001A (en) * 2017-09-28 2018-02-23 深圳市易成自动驾驶技术有限公司 Radial basis function network parametric solution method, apparatus and computer-readable recording medium
CN108090412A (en) * 2017-11-17 2018-05-29 西北工业大学 A kind of radar emission source category recognition methods based on deep learning
CN109188417B (en) * 2018-09-04 2022-11-15 同方电子科技有限公司 Method for single-station passive positioning of scanning radiation source by adopting unmanned aerial vehicle platform
CN109856621A (en) * 2018-12-07 2019-06-07 中国船舶重工集团公司第七二四研究所 A kind of radar self-adaption Passive Location based on satellite detecting information

Also Published As

Publication number Publication date
CN110632584A (en) 2019-12-31

Similar Documents

Publication Publication Date Title
Wang et al. CSI-based fingerprinting for indoor localization: A deep learning approach
Noroozi et al. Weighted least squares target location estimation in multi‐transmitter multi‐receiver passive radar using bistatic range measurements
WO2021175344A2 (en) Method and device for dynamic measurement and reconstruction of wireless channel impulse response
Luo et al. On the accuracy of passive source localization using acoustic sensor array networks
CN110515038B (en) Self-adaptive passive positioning device based on unmanned aerial vehicle-array and implementation method
CN110536245A (en) A kind of indoor wireless positioning method and system based on deep learning
CN107171703B (en) Method capable of simulating propagation characteristics of fading signals in indoor multi-antenna communication system
CN110596668B (en) Target external radiation source passive positioning method based on reciprocal deep neural network
Yin et al. Millimeter wave wireless assisted robot navigation with link state classification
Steen et al. Acoustic source tracking in long baseline microphone arrays
CN113949439A (en) Geometric random channel modeling method for unmanned aerial vehicle air-to-air communication
CN114386321A (en) AOA and TOF joint estimation method and device for indoor positioning and storage medium
Zheng et al. Convex optimization algorithms for cooperative RSS-based sensor localization
CN110632584B (en) Passive target external radiation source positioning method based on parallel radial basis network
Griffin et al. Real-time localization of multiple audio sources in a wireless acoustic sensor network
Ma et al. Super-resolution time delay estimation using exponential kernel correlation in impulsive noise and multipath environments
Chatelier et al. Influence of dataset parameters on the performance of direct ue positioning via deep learning
CN108614235A (en) A kind of single snap direction-finding method of more dove group information exchanges
CN116930963A (en) Through-wall imaging method based on wireless communication system
Al-Sadoon et al. New and less complex approach to estimate angles of arrival
CN111263295A (en) WLAN indoor positioning method and device
Serbetci et al. Simple and Effective Augmentation Methods for CSI Based Indoor Localization
Chen et al. TDOA positioning method based on Taylor series expansion based on cuckoo search algorithm
CN115397012A (en) Realization method of UWB positioning tracking system based on TWR-TDOA estimation and MPGA layout optimization
CN108414973A (en) A kind of direct localization method of multiple target based on neural computing

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant