CN106355245A - Method for integrating array antenna directional images on basis of neural network algorithms - Google Patents
Method for integrating array antenna directional images on basis of neural network algorithms Download PDFInfo
- Publication number
- CN106355245A CN106355245A CN201610817682.4A CN201610817682A CN106355245A CN 106355245 A CN106355245 A CN 106355245A CN 201610817682 A CN201610817682 A CN 201610817682A CN 106355245 A CN106355245 A CN 106355245A
- Authority
- CN
- China
- Prior art keywords
- array
- antenna
- network
- sample
- training
- 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.)
- Granted
Links
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density 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
Abstract
The invention provides a method for integrating array antenna directional images on the basis of neural network algorithms. By the aid of the method, problems of high computational complexity, long computation time and extreme difficulty in finding the optimal weight values of the traditional technologies for integrating array antenna directional images can be solved. The method includes steps of building antenna array models; acquiring the optimal weight values; generating antenna array directional images. The method has the advantages that the neural network algorithms are applied to technologies for integrating the array antenna directional images, trained RBF (radial basis function) neural networks can be effectively used for integrating the array antenna directional images, in other words, the array antenna directional images can be integrated by the aid of the neural network algorithms, the method is specially high in synthesis speed, low in zero-trap level and outstanding in main lobe, and the directional images are in stable shapes.
Description
Technical field
The present invention relates to mobile communication antenna technical field.
Background technology
Pattern Synthesis of Antenna Array technology is applied to any antenna array with sophisticated signal processor, and it can adjust
Or its beam pattern of self adaptation is it is therefore an objective to strengthening signal interested and reducing interference signal additionally it is possible to mitigate multipath effect
Adverse effect, it is one of research field of smart antenna.With social information exchange sharply increase, PMC
Rapid popularization, frequency spectrum has become more and more valuable resource, and smart antenna adopts space division multiple access technique, is propagated using signal
Difference on direction, changes array aerial direction figure by the weighted amplitude and phase place adjusting each array element signals, thus suppressing
Interference, improves the signal bandwidth of signal to noise ratio, power system capacity and permission, the resource such as effectively save frequency spectrum and power.
The directional diagram of antenna array system is to rely on the superposition of directional diagram of each oscillator and realize, if each oscillator
The amplitude of voltage drive and phase place are varied from, then this antenna system is wanted corresponding directional diagram also can change.
In mobile communication application, many times base station is relatively-stationary, and terminal is mobile；Good between them in order to maintain
Wireless channel it is ensured that link normal work it is desirable to the main lobe of smart antenna be capable of the moment track terminal that is to say, that
The directional diagram of antenna system will do the adjustment of self with the change in location of user.The present invention is using in the time the shortest
Find a vector so that the main lobe of its corresponding antenna radiation pattern is directed at the direction of user.
The present invention mainly realizes linear array antenna Pattern Synthesis using intelligent algorithm, and neural network algorithm has very strong appearance
Mistake and Function approximation capabilities, using multigroup preferable directional diagram sample data train RBF Neural Network, non-during training network
The independent variable of Linear Mapping and dependent variable, are the snapshot of signal direction of arrival and the optimum weight coefficient value of array element respectively, signal snapshot
It is randomly selecting, substantial amounts of.After network training is good, when the direction of arrival at the different angle of input, the adaptive life of antenna array system
Become directional diagram.
Content of the invention
It is an object of the invention to overcoming shortcoming and the deficiency of prior art, provide a kind of battle array based on neural network algorithm
Array antenna Pattern Synthesis method, the method aggregate velocity is especially fast, and null level is low, and main lobe projects, and pattern shapes are steady
Fixed.
The purpose of the present invention is achieved through the following technical solutions: a kind of array aerial direction figure based on neural network algorithm
Integrated approach, comprising:
Antenna array model establishment step: n coco antenna forms array, and described array comprises n2Individual monopole antenna, sky
Linear array adopts uniform straight line array form, antenna spacing d≤λ, and the unit section length of side is 1/2 medium wavelength:
According to the operating center frequency f=1.8ghz of antenna, the DIELECTRIC CONSTANT ε of substrate materialr=2.56, for making positive and negative
Micro-strip section is in picturesque disorder, produces the transmission mode being suitable for and radiation mode, needs a ≈ b；Dielectric-slab length l is dielectric-slab width simultaneously
6.5 times of degree w, wherein b is spaced for chip unit；
Best initial weights obtaining step: include information determining step, network parameter selects step, learning training algorithms selection step
Suddenly, adjustment and obtaining step；
Information determining step: problem is converted into mode the treatable form that network can be expressed, determines each node
Input and output；Calculate each center of a sample with k means Method, n be center of a sample's number it may be assumed that first clustering, then by value
Method asks the center of a sample of every class, with required center of a sample for implicit layer unit；
Network parameter selects step: determines input, the number of output neuron, the number of plies of multitiered network and hidden neuron
Number, snapshot number of times is k, and training sample number is a, and wherein a is plural number, is divided into real part and imaginary part and calculates it is desirable to be output as
The real part of plural weight coefficient of array element and imaginary part, training sample input is to be reached by taking identical with number of training group of ripple at random
Angle obtains, and plural weight coefficient is obtained using conjugate gradient algorithms；Choosing of anticipation error is missed by the expectation different to two
The network of difference is trained, and therefrom determines a network；
Learning training algorithms selection step: first initialized, then complete function approximation using radial basis function network, determine
Input sample and desired output, wherein, the snapshot a of antenna array k × n matrix is set to input sample, weight vector ω of each array element
It is set to desired output；Antenna array array number is configured, determines arrival bearing, after normalization, obtain input vector signal, real
Existing input vector signal is to the mapping of weight vector；
Adjustment and obtaining step: to clustering method, Hidden nodes, hidden node data center learning coefficient, hidden node extension
Constant learning coefficient and hidden node output weights learning coefficient are adjusted, and after completing training, network connection power is determined, can
The best initial weights of desired signal are directly obtained using network；
Antenna array pattern generation step: using the best initial weights obtaining, antenna array model is verified, thus raw
Become antenna array pattern.
Described snapshot number of times k is set to 200, and training sample number a is set to 800, and gained input sample data is 1600,
800 training sample inputs are obtained by taking 800 groups of direction of arrival at random, obtain 800 groups of plural numbers using conjugate gradient algorithms
Weight coefficient, with these data training networks.
Described hidden node data center learning coefficient are 0.001, and hidden node extension constant learning coefficient are 0.002, hidden section
Point output weights learning coefficient is 0.005.
Described antenna array array number is set to 8, and arrival bearing is arranged between [- π, π].
Described every group of direction of arrival includes a useful signal and three interference signals, and adds Gauss in each signal
Noise.
The present invention can easily find out best initial weights, and aggregate velocity is especially fast, and null level is low, and main lobe projects, and directional diagram
Dimensionally stable.
Brief description
Fig. 1 is that the robust detection value of three simulation results compares.
Fig. 2 is radial base neural net structural representation.
Fig. 3 is the array antenna figure general flowchart of neural network algorithm.
Fig. 4 is antenna model schematic diagram.
Fig. 5 is collinear array schematic diagram.
Specific embodiment
Specific embodiment one: a kind of Pattern Synthesis of Antenna Array method based on neural network algorithm, specifically include
Following steps,
Antenna array model establishment step: n coco antenna forms array, and described array comprises n2Individual monopole antenna, sky
Linear array adopts uniform straight line array form, antenna spacing d≤λ, and the unit section length of side is 1/2 medium wavelength:
According to the operating center frequency f=1.8ghz of antenna, the DIELECTRIC CONSTANT ε of substrate materialr=2.56, for making positive and negative
Micro-strip section is in picturesque disorder, produces the transmission mode being suitable for and radiation mode, needs a ≈ b；Dielectric-slab length l is dielectric-slab width simultaneously
6.5 times of degree w, wherein b is spaced for chip unit；
Best initial weights obtaining step: include information determining step, network parameter selects step, learning training algorithms selection step
Suddenly, adjustment and obtaining step；
Information determining step: problem is converted into mode the treatable form that network can be expressed, determines each node
Input and output；Calculate each center of a sample with k means Method, n be center of a sample's number it may be assumed that first clustering, then by value
Method asks the center of a sample of every class, with required center of a sample for implicit layer unit；
Network parameter selects step: determines input, the number of output neuron, the number of plies of multitiered network and hidden neuron
Number, snapshot number of times is k, and training sample number is a, and wherein a is plural number, is divided into real part and imaginary part and calculates it is desirable to be output as
The real part of plural weight coefficient of array element and imaginary part, training sample input is to be reached by taking identical with number of training group of ripple at random
Angle obtains, and plural weight coefficient is obtained using conjugate gradient algorithms；Choosing of anticipation error is missed by the expectation different to two
The network of difference is trained, and therefrom determines a network；
Learning training algorithms selection step: first initialized, then complete function approximation using radial basis function network, determine
Input sample and desired output, wherein, the snapshot a of antenna array k × n matrix is set to input sample, weight vector ω of each array element
It is set to desired output；Antenna array array number is configured, determines arrival bearing, after normalization, obtain input vector signal, real
Existing input vector signal is to the mapping of weight vector；
Adjustment and obtaining step: to clustering method, Hidden nodes, hidden node data center learning coefficient, hidden node extension
Constant learning coefficient and hidden node output weights learning coefficient are adjusted, and after completing training, network connection power is determined, can
The best initial weights of desired signal are directly obtained using network；
Antenna array pattern generation step: using the best initial weights obtaining, antenna array model is verified, thus raw
Become antenna array pattern.
Specific embodiment two: specific embodiment two is differred primarily in that with specific embodiment one, snapshot first
Number of times k is set to 200, and training sample number a is set to 800, because a is plural number, so real part and imaginary part calculating need to be separated into, so
Input sample data 1600 is it is desirable to be output as the real part of plural weight coefficient and the imaginary part of array element.This 800 training sample inputs
Obtained by taking 800 groups of direction of arrival at random.Obtain 800 groups of plural weight coefficients using conjugate gradient algorithms, with these data
Training network.
Maximum frequency of training is 5000 times in a network.Generally, the selection of learning rate tend to choose less
The stability to ensure system for the learning rate, the selection range of learning rate is between 0.01～0.8.Too big learning rate
Lead to the unstable of study, too little value leads to the training time extremely grown again.Learning rate is adjusted with frequency of training, but each
In training, the learning rate of whole network is identical.Specifically there are two schemes, one is to train with larger learning rate, so when starting
Afterwards learning rate is adjusted according to training.Two is with the training of less learning rate when starting, and is then adjusted according to training and learns
Habit rate.Final hidden node data center learning coefficient are 0.001, and hidden node extension constant learning coefficient are 0.002, and hidden node is defeated
Going out weights learning coefficient is 0.005.
The selection of anticipation error: generally as a comparison, the net of expected error value that can be simultaneously different to two
Network is trained, and the consideration finally by composite factor to determine one of network, determines that expectation target error is 0.09.
Secondly, antenna array array number is set to 8.Arrival bearing between [- π, π], 800 groups of direction of arrival randomly selecting,
Every group of direction of arrival includes a useful signal and three interference signals, also adds Gaussian noise in each signal.
Operation principle:
Radial basis function neural network by up of three layers, structure chart such as Fig. 2, input layer only transmits input signal to hidden
Layer, hidden node is made up of various transmission functions, and the most frequently used radial function as Gaussian function is constituted, and
Output node layer is typically simple linear function.
When n is 0, RBF exports maximum 1, when that is, between weight vector ω and input vector p, distance reduces,
Output will increase.That is, RBF produces response to input signal in local.Input signal n of function is close
During the central range of function, hidden node will produce larger output.It can be seen that this network has partial approximation energy
Power, so radial primary function network also becomes local sensing field network.Threshold value b is used for adjusting the sensitivity of radial direction base neuron.
Assume that obtaining the input vector signal after normalization is f, when data by input layer near hidden layer when, hidden layer rbf quilt
Activation, outfan needs the weight vector that the amount of nonlinearity approached is time-varying, thus realizing f to the mapping of weight vector ω.Can represent
For:
Wherein, l, m are hidden layer and output layer nodes respectively, hi,lIt is l-th node of hidden layer and i-th node of output layer
Connection weight, clIt is Gaussian function center, | | | | it is Euclidean Norm.Available l-th unit is corresponding to be output as:
In formula, ziK () is the output of l-th hidden unit, xnT () is n-th component of t-th input sample vector, ct,n
For n-th component of l-th transform center vector in hidden layer, σlIt is the control parameter of corresponding l-th center vector.
Omnidirectional antenna is emulated:
By the great advantages that coco antenna forms array antenna it is exactly: the array being made up of n coco antenna is suitable
In containing n2Individual monopole antenna (assuming that each coco antenna is also to be made up of n section microstrip line section), but its only n feedback
Electric, comparatively speaking, if an array is made up of n monopole antenna array, it is it is necessary to have n2Individual distributing point is (false
If each one pole subarray is also to be made up of n monopole).Certainly, because coco antenna structure is simple, have price advantage and
Performance advantage.
The unit section length of side is 1/2 medium wavelength:
According to the operating center frequency f=1.8ghz of antenna, dielectric constant=2.56 of substrate material, for making positive and negative
Micro-strip section is in picturesque disorder, produces the transmission mode being suitable for and radiation mode, needs a ≈ b；Dielectric-slab length l is dielectric-slab width simultaneously
6.5 times about of degree w, can obtain a=52mm, b=58.5mm, l=520mm, w=80mm, wa=69.5mm, wb=
11.4mm, h=2.5mm, punch on antenna is away from the paster of feed end, and hole is connected to upper and lower surface, and the radius in hole is 3mm,
At feed end 480mm.Wherein h is substrate thickness, and b is spaced for chip unit.Antenna schematic diagram is as shown in Figure 4.Antenna
The input impedance of distributing point requires as 50 ω.Observe the input impedance z parameter of the feed port under mid frequency.Adjustment paster
Size and feed position all can change the input impedance value of antenna, it is 50 ω and void that purpose makes the real part of input impedance
Portion is 0 ω.
Emulated using feko software, model is created according to data above and adjusts mesh-density, by adjusting medium
Control card, output control card, Electromagnetic Calculation card are it is considered to becoming to taking effect, calculating far field, calculate standing wave etc..The feed electricity of antenna
The amplitude of pressure is 1v, and phase place is 0 °.Simulation establishes 8 array element linear array antenna, in the case of known direction of arrival, by using
Amplitude and phase place to control each antenna element voltage for the weights that algorithm obtains, by the way of each array element individually feeds.Depositing
In the case of coupling mutual inductance between less antenna array element, the reliability and stability of the directional diagram that inspection generates.As Fig. 5 is
The schematic diagram that 8 array element collinear arrays are placed.Linear array is designed with uniform straight line array form, prevents appearance on the h face of antenna
Graing lobe, needs to ensure bay center distance d≤λ, and when antenna spacing d≤λ/2 it may appear that Mutual Inductance Coupling is existing
As so selecting d ≈ 0.6 λ.Each bay enjoys a feed voltage source.
The beneficial effect of actual moving process:
The present invention draws each group of data average by many experiments emulation it can be seen that the average of null level and side everywhere
The average of lobe level all meets desired indicator and requires (referring to table 1).
The meansigma methodss of table 1 each group partial parameters
Every group of emulation, averagely occurs in that 3.5～3.8 nulls being less than -30db, illustrates that neural network algorithm is easier
Produce low level null, but in these more low level nulls, some appear on unexpected direction, but certainly not
Appear in useful signal principal direction, the probability of appearance accounts for deep null occurs 20%～50%, illustrates side
Null is occurred in that in the angle of lobe.
Although unexpected null occurs, have no effect on neural network algorithm adaptive solution array aerial direction figure comprehensive
Conjunction problem, it still has practicality.For different array antennas, as long as choosing appropriate neural network algorithm, be given enough
Training sample, according to actual requirement adjust network parameter it is possible to obtain accurate network weight.Neural network algorithm is realized
Pattern Synthesis of Antenna Array, aggregate velocity is especially fast, and null level is low, and main lobe projects, and pattern shapes are stable.
At same group, under the interference signal effect in three directions, the directional diagram generating every time is very identical, tentatively illustrates to calculate
The robustness of method is good.The data obtained will be emulated and calculate robust detection value, acquired results will be shown in Table 2.
Table 2 neural network algorithm robust detection value
Plotted block diagram, such as Fig. 1.Can more intuitively find out, the detection value changes of its each robustness are little, so warp
Interference signal direction of arrival direction be varied multiple times, the directional diagram of this algorithm gained remain to adaptive convert therewith, obtain preferably
Antenna reception.
In terms of the depth of null, null level is all relatively low so that main lobe is very full.Form directional diagram using neutral net
When, efficiency is very high, and robustness is good, and gained directional diagram is stable.
After completing training, network connection power determines that, and directly can obtain the optimum power of desired signal using network
Value, flow chart such as Fig. 3 of whole process, using multigroup preferable directional diagram sample data train RBF Neural Network, trains
The independent variable of network non-linear hour mapping and dependent variable, are the snapshot of signal direction of arrival and the optimum weight coefficient value of array element respectively,
Signal snapshot is randomly selecting, substantial amounts of.After network training is good, when the direction of arrival at the different angle of input, antenna array system is certainly
The generation directional diagram adapting to.
Above to a kind of Pattern Synthesis of Antenna Array method based on neural network algorithm provided by the present invention, carry out
It is discussed in detail, specific case used herein is set forth to the principle of the present invention and embodiment, above example
Explanation be only intended to help and understand the method for the present invention and its core concept；Simultaneously for one of ordinary skill in the art,
According to the thought of the present invention, all will change in specific embodiments and applications, in sum, in this specification
Hold and should not be construed as limitation of the present invention.
Claims (5)
1. a kind of Pattern Synthesis of Antenna Array method based on neural network algorithm it is characterised in that
Antenna array model establishment step: n coco antenna forms array, and described array comprises n2Individual monopole antenna, aerial array
Using uniform straight line array form, antenna spacing d≤λ, the unit section length of side is 1/2 medium wavelength:
According to the operating center frequency f=1.8ghz of antenna, the DIELECTRIC CONSTANT ε of substrate materialr=2.56, for making positive and negative micro-strip
Section is in picturesque disorder, produces the transmission mode being suitable for and radiation mode, needs a ≈ b；Dielectric-slab length l is dielectric-slab width w simultaneously
6.5 times, wherein b is spaced for chip unit；
Best initial weights obtaining step: include information determining step, network parameter select step, learning training algorithms selection step,
Adjustment and obtaining step；
Information determining step: problem is converted into mode the treatable form that network can be expressed, determines the defeated of each node
Enter output；Calculate each center of a sample with k means Method, n be center of a sample's number it may be assumed that first clustering, then by value method
Ask the center of a sample of every class, with required center of a sample for implicit layer unit；
Network parameter selects step: determines input, the number of the number of output neuron, the number of plies of multitiered network and hidden neuron
Mesh, snapshot number of times is k, and training sample number is a, and wherein a is plural number, is divided into real part and imaginary part calculates it is desirable to be output as array element
The real part of plural weight coefficient and imaginary part, training sample input be to be obtained by the direction of arrival taking identical with number of training group at random
Arrive, plural weight coefficient is obtained using conjugate gradient algorithms；The expected error value different to two are passed through in choosing of anticipation error
Network be trained, therefrom determine a network；
Learning training algorithms selection step: first initialized, then complete function approximation using radial basis function network, determine input
Sample and desired output, wherein, the snapshot a of antenna array k × n matrix is set to input sample, and weight vector ω of each array element is set to
Desired output；Antenna array array number is configured, determines arrival bearing, after normalization, obtain input vector signal, realize letter
Number input vector is to the mapping of weight vector；
Adjustment and obtaining step: to clustering method, Hidden nodes, hidden node data center learning coefficient, hidden node extension constant
Learning coefficient and hidden node output weights learning coefficient are adjusted, and after completing training, network connection power is determined, can be direct
Obtain the best initial weights of desired signal using network；
Antenna array pattern generation step: using the best initial weights obtaining, antenna array model is verified, thus generating sky
Linear array directional diagram.
2. it is characterised in that snapshot number of times k is set to 200, training sample number a is set to 800 to the method for claim 1,
Gained input sample data is 1600, and 800 training sample inputs are obtained by taking 800 groups of direction of arrival at random, utilize
Conjugate gradient algorithms obtain 800 groups of plural weight coefficients, with these data training networks.
3. method as claimed in claim 2 it is characterised in that described hidden node data center learning coefficient be 0.001, hidden section
Point extension constant learning coefficient are 0.002, and hidden node output weights learning coefficient is 0.005.
4. it is characterised in that described antenna array array number is set to 8, arrival bearing is arranged on method as claimed in claim 3
Between [- π, π].
5. method as claimed in claim 4 is it is characterised in that every group of direction of arrival includes a useful signal and three interference
Signal, and add Gaussian noise in each signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817682.4A CN106355245B (en) | 2016-09-12 | 2016-09-12 | A kind of Pattern Synthesis of Antenna Array method based on neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610817682.4A CN106355245B (en) | 2016-09-12 | 2016-09-12 | A kind of Pattern Synthesis of Antenna Array method based on neural network algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106355245A true CN106355245A (en) | 2017-01-25 |
| CN106355245B CN106355245B (en) | 2018-09-21 |
Family
ID=57859704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610817682.4A Active CN106355245B (en) | 2016-09-12 | 2016-09-12 | A kind of Pattern Synthesis of Antenna Array method based on neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106355245B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107623691A (en) * | 2017-09-29 | 2018-01-23 | 长沙市智为信息技术有限公司 | A kind of ddos attack detecting system and method based on reverse transmittance nerve network algorithm |
| CN107729694A (en) * | 2017-11-17 | 2018-02-23 | 电子科技大学 | A kind of multi-parameter electromagnetic field modeling and simulating method based on neutral net |
| CN107729627A (en) * | 2017-09-26 | 2018-02-23 | 西安电子科技大学 | The unit fast selecting method of chance array antenna |
| CN108957406A (en) * | 2018-05-18 | 2018-12-07 | 电子科技大学 | A kind of radar main lobe interference suppression method neural network based |
| CN109271695A (en) * | 2018-09-07 | 2019-01-25 | 中南大学 | Multiple target antenna design method neural network based |
| CN109631997A (en) * | 2018-12-20 | 2019-04-16 | 树根互联技术有限公司 | Engineering equipment health detecting method, device and electronic equipment |
| CN110427590A (en) * | 2019-08-01 | 2019-11-08 | 电子科技大学 | The efficient integrated approach of Large Scale Sparse array antenna based on adaptive probability study |
| CN110427669A (en) * | 2019-07-20 | 2019-11-08 | 中国船舶重工集团公司第七二四研究所 | A kind of neural network model calculation method of phase-array scanning radiation beam |
| CN110554352A (en) * | 2019-09-11 | 2019-12-10 | 哈尔滨工业大学 | Method for estimating direction of arrival of interference source of aerospace measurement and control system based on VGG16 neural network |
| CN110596668A (en) * | 2019-09-20 | 2019-12-20 | 中国人民解放军国防科技大学 | Target external radiation source passive positioning method based on reciprocal deep neural network |
| CN111276822A (en) * | 2020-02-24 | 2020-06-12 | 西安电子科技大学 | Antenna pattern main lobe controllable antenna array design method |
| CN112013833A (en) * | 2020-07-23 | 2020-12-01 | 中国人民解放军战略支援部队信息工程大学 | VICTS antenna satellite-to-satellite tracking method based on deep neural network |
| CN112054869A (en) * | 2020-08-05 | 2020-12-08 | 西安电子科技大学 | Wavelength allocation method based on continuous Hopfield neural network |
| CN113777571A (en) * | 2021-08-04 | 2021-12-10 | 中山大学 | Unmanned aerial vehicle cluster dynamic directional diagram synthesis method based on deep learning |
| CN114398831A (en) * | 2022-01-14 | 2022-04-26 | 厦门大学 | Surface conformal array design method based on transformation optics and neural network algorithm |
| WO2023284571A1 (en) * | 2021-07-14 | 2023-01-19 | 中兴通讯股份有限公司 | Beamforming method and apparatus, base station, and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100317420A1 (en) * | 2003-02-05 | 2010-12-16 | Hoffberg Steven M | System and method |
| CN104485513A (en) * | 2014-12-17 | 2015-04-01 | 镇江市高等专科学校 | Real-time integrating method for wideband antenna array |
| CN104517035A (en) * | 2014-12-22 | 2015-04-15 | 中国舰船研究设计中心 | Planar array antenna active scattering directional diagram predication method |
| CN104569907A (en) * | 2014-09-04 | 2015-04-29 | 深圳市金溢科技股份有限公司 | Wireless positioning method and system based on neural network and road side unit |
-
2016
- 2016-09-12 CN CN201610817682.4A patent/CN106355245B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100317420A1 (en) * | 2003-02-05 | 2010-12-16 | Hoffberg Steven M | System and method |
| CN104569907A (en) * | 2014-09-04 | 2015-04-29 | 深圳市金溢科技股份有限公司 | Wireless positioning method and system based on neural network and road side unit |
| CN104485513A (en) * | 2014-12-17 | 2015-04-01 | 镇江市高等专科学校 | Real-time integrating method for wideband antenna array |
| CN104517035A (en) * | 2014-12-22 | 2015-04-15 | 中国舰船研究设计中心 | Planar array antenna active scattering directional diagram predication method |
Non-Patent Citations (2)
| Title |
|---|
| 刘建霞: "基于RBF神经网络的天线阵方向图建模", 《太原理工大学学报》 * |
| 姜正茂 等: "BP神经网络天线方向图获取方法研究", 《电子科技大学学报》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107729627A (en) * | 2017-09-26 | 2018-02-23 | 西安电子科技大学 | The unit fast selecting method of chance array antenna |
| CN107729627B (en) * | 2017-09-26 | 2020-12-08 | 西安电子科技大学 | Quick selection method for unit of opportunistic array antenna |
| CN107623691A (en) * | 2017-09-29 | 2018-01-23 | 长沙市智为信息技术有限公司 | A kind of ddos attack detecting system and method based on reverse transmittance nerve network algorithm |
| CN107729694B (en) * | 2017-11-17 | 2020-09-25 | 电子科技大学 | Multi-parameter electromagnetic field modeling simulation method based on neural network |
| CN107729694A (en) * | 2017-11-17 | 2018-02-23 | 电子科技大学 | A kind of multi-parameter electromagnetic field modeling and simulating method based on neutral net |
| CN108957406A (en) * | 2018-05-18 | 2018-12-07 | 电子科技大学 | A kind of radar main lobe interference suppression method neural network based |
| CN109271695A (en) * | 2018-09-07 | 2019-01-25 | 中南大学 | Multiple target antenna design method neural network based |
| CN109271695B (en) * | 2018-09-07 | 2022-12-23 | 中南大学 | Multi-target antenna design method based on neural network |
| CN109631997B (en) * | 2018-12-20 | 2021-07-09 | 树根互联股份有限公司 | Engineering equipment health detection method and device and electronic equipment |
| CN109631997A (en) * | 2018-12-20 | 2019-04-16 | 树根互联技术有限公司 | Engineering equipment health detecting method, device and electronic equipment |
| CN110427669A (en) * | 2019-07-20 | 2019-11-08 | 中国船舶重工集团公司第七二四研究所 | A kind of neural network model calculation method of phase-array scanning radiation beam |
| CN110427590B (en) * | 2019-08-01 | 2022-03-25 | 电子科技大学 | Large sparse array antenna efficient comprehensive method based on adaptive probability learning |
| CN110427590A (en) * | 2019-08-01 | 2019-11-08 | 电子科技大学 | The efficient integrated approach of Large Scale Sparse array antenna based on adaptive probability study |
| CN110554352A (en) * | 2019-09-11 | 2019-12-10 | 哈尔滨工业大学 | Method for estimating direction of arrival of interference source of aerospace measurement and control system based on VGG16 neural network |
| CN110596668A (en) * | 2019-09-20 | 2019-12-20 | 中国人民解放军国防科技大学 | Target external radiation source passive positioning method based on reciprocal deep neural network |
| CN111276822A (en) * | 2020-02-24 | 2020-06-12 | 西安电子科技大学 | Antenna pattern main lobe controllable antenna array design method |
| CN112013833A (en) * | 2020-07-23 | 2020-12-01 | 中国人民解放军战略支援部队信息工程大学 | VICTS antenna satellite-to-satellite tracking method based on deep neural network |
| CN112054869A (en) * | 2020-08-05 | 2020-12-08 | 西安电子科技大学 | Wavelength allocation method based on continuous Hopfield neural network |
| WO2023284571A1 (en) * | 2021-07-14 | 2023-01-19 | 中兴通讯股份有限公司 | Beamforming method and apparatus, base station, and storage medium |
| CN113777571A (en) * | 2021-08-04 | 2021-12-10 | 中山大学 | Unmanned aerial vehicle cluster dynamic directional diagram synthesis method based on deep learning |
| CN113777571B (en) * | 2021-08-04 | 2023-08-11 | 中山大学 | Unmanned aerial vehicle cluster dynamic pattern synthesis method based on deep learning |
| CN114398831A (en) * | 2022-01-14 | 2022-04-26 | 厦门大学 | Surface conformal array design method based on transformation optics and neural network algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106355245B (en) | 2018-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106355245A (en) | Method for integrating array antenna directional images on basis of neural network algorithms | |
| CN111294096A (en) | Channel capacity optimization method of intelligent reflector MISO wireless communication system | |
| CN106407549B (en) | A kind of Pattern Synthesis of Antenna Array optimization method based on genetic algorithm | |
| CN110535486B (en) | Radio frequency signal direct processing type wireless transceiver based on super surface neural network | |
| Zeng et al. | Intelligent omni-surfaces: Reflection-refraction circuit model, full-dimensional beamforming, and system implementation | |
| CN110535518B (en) | Optimization design method for wide beam transmission beam forming | |
| CN113093094B (en) | Intelligent incident wave direction detection method based on phase regulation and control super surface | |
| CN107104720A (en) | The relatively prime array adaptive beamforming method rebuild based on covariance matrix virtual Domain discretization | |
| Tang et al. | Transmissive RIS for B5G communications: Design, prototyping, and experimental demonstrations | |
| CN107135023A (en) | Three-dimensional training code book design method and beam alignment for millimeter-wave communication system | |
| Wu et al. | Distance-aware precoding for near-field capacity improvement in XL-MIMO | |
| CN104102775B (en) | A kind of electricity based on Sidelobe Suppression adjusts passive antenna beam optimization method | |
| JP2002118414A (en) | Controller and controlling method of array antenna | |
| CN114726687A (en) | Channel estimation method of intelligent reflector assisted millimeter wave large-scale MIMO system | |
| Chen et al. | Channel customization for joint Tx-RISs-Rx design in hybrid mmWave systems | |
| Cerri et al. | Feed forward neural networks for path loss prediction in urban environment | |
| CN105510872B (en) | A kind of assay method of two-dimentional direction of arrival suitable for extensive mimo system | |
| Kao et al. | AI-aided 3-D beamforming for millimeter wave communications | |
| Jiang et al. | Active sensing for two-sided beam alignment and reflection design using ping-pong pilots | |
| CN110518949A (en) | The design method and training method of extensive antenna training code book under constant power constraint | |
| CN106301632A (en) | A kind of method of rapid evaluation multiaerial system channel capacity stability | |
| Zeng et al. | Transmit and receive hybrid beamforming design for OFDM dual-function radar-communication systems | |
| Park et al. | Angle tracking automatic beamforming for microwave power transfer systems | |
| Patnaik et al. | A frequency reconfigurable antenna design using neural networks | |
| Chen et al. | Razor-sharp Narrowbeam Communications and Management for Terahertz Wireless Networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |