CN110048978A - A kind of signal modulate method - Google Patents
A kind of signal modulate method Download PDFInfo
- Publication number
- CN110048978A CN110048978A CN201910280889.6A CN201910280889A CN110048978A CN 110048978 A CN110048978 A CN 110048978A CN 201910280889 A CN201910280889 A CN 201910280889A CN 110048978 A CN110048978 A CN 110048978A
- Authority
- CN
- China
- Prior art keywords
- signal
- recognition
- identification
- modulation
- signal modulate
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/0012—Modulated-carrier systems arrangements for identifying the type of modulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of signal modulate methods, including the recognition methods based on decision theory and are based on statistical-simulation spectrometry;The modulation system of the signal of communication received is identified in the presence of noise and interference, to provide foundation to be further processed and analyzing signal of communication, it is described that multiple hypothesis test problem as is regarded identification problem based on the recognition methods of decision theory, theory deduction is carried out to the test statistics of modulated signal.In the present invention, by building convolutional neural networks model, use training set training pattern, the changing rule of internal characteristics or signal sequence data that model is learnt to known modulation type signal, trained model is tested on test set finally, can effectively identify the modulation type of signal.This mode not only solves the artificial problem for extracting feature difficulty, and can obtain preferable recognition effect, and model proposed in this paper, the accuracy rate identified on test set can achieve 95% or more.
Description
Technical field
The present invention relates to blipology field more particularly to a kind of signal modulate methods.
Background technique
Automatic Modulation Recognition also known as automatic Modulation classification, it is to realize the automatic Modulation category classification for receiving signal.
Currently, Automatic Modulation Recognition is the important technology of many dual-use scenes.Automatic Modulation Recognition technology is initially applied to
The military fields such as electronic countermeasure and information investigation, with the continuous development of communication system, Automatic Modulation Recognition is in civilian cognition
There has also been very big development in the radio of radio and software definition.By the help of Automatic Modulation Recognition, user can
The correctly parameter of setting communication transmission process, to ensure the quality communicated.
In military domain, how Modulation Identification for the interception of information and selects optimal interference to provide in electronic warfare system
Important foundation, is mainly manifested in enemy radar type identification, in enemy intelligence interception and enemy's Radar recognition, so adjusting
System identification has very important status at military aspect.In civil field, Modulation Identification is mainly used in the prison of radio station
It surveys and uses, the service condition of monitoring radio platform, the information that can also be carried to the signal of communication of non-partner obtains,
The effectively frequency spectrum of supervision communication.
In recent years, deep-neural-network is flourished in fields such as biology, computer and electronics, attracts crowd
More scientists fall over each other to inquire into.After being suggested from it, it just becomes a popular research direction of artificial intelligence field, breaches
Limitation of the traditional neural network to the number of plies, the multilayered structure model by establishing similar human brain solve challenge.It is main
Heterosis exists: deep layer network itself belongs to a kind of feedback network with store function, can use the high speed fortune of computer
Calculation ability solves the optimal solution of challenge, improves online problem-solving ability.Most outstanding be advantage is large-scale
Sample data can successively extract internal characteristics information by layered structure, have a wide range of applications scene.Deep layer network obtains
Characteristic information it is more representative than the expert features manually extracted, it is easier to realize classification and prediction to initial data.Base
In the above advantage, the research of deep learning has had changed into a cross-cutting, multi-disciplinary research topic, and widely successfully answers
Speech recognition and field of image recognition are used, the application of other field still needs to be explored.
The development of wireless communication technique is that Modulation identification technology brings many new problems and challenge.In face of nothing of new generation
The revolution of line communication, conventional modulated recognition methods have shown many shortcoming and deficiency.In military domain, with International Politics shape
Gesture is increasingly complicated, in view of the future may appear electronic warfare, Modulation identification technology had received the height of various countries national defence expert
Degree concern and attention;In civil field, as the use of radio platforms is more and more lack of standardization, major wireless provider is for adjusting
Identification technology processed focuses more on.
Summary of the invention
It is an object of the invention to: to solve the above-mentioned problems, and a kind of signal modulate method proposed.
To achieve the goals above, present invention employs following technical solutions:
The basic task of signal modulate is to identify the signal of communication received in the presence of noise and interference
Modulation system, to provide foundation to be further processed and analyzing signal of communication.Currently, Modulation identification technology mainly have it is two big
Class method: recognition methods based on decision theory and it is based on statistical pattern recognition method.
Regard identification problem as multiple hypothesis test problem based on the recognition methods of decision theory, the inspection to modulated signal
Statistic carries out theory deduction, finds suitable threshold value using minimum likelihood method and differentiates to modulation.This method
Maximum advantage is that have more complete theoretical basis, but the derivation of likelihood function is extremely complex in the analysis process, method
It is poor for applicability, and a large amount of priori knowledge is needed, therefore this method is not very practical, is rarely employed in practice.
Modulation Identification method based on pattern recognition theory is the method for current relatively mainstream, and the method asks Modulation Identification
Topic regards a kind of typical pattern recognition problem as, and pattern-recognition refers to: disaggregatedly being identified things by software and hardware algorithm
Out.With the development of artificial intelligence in recent years, mode identification technology is also more and more to obtain the wide of researchers at home and abroad
General attention and research, and be more and more applied in daily life, for example, fingerprint recognition, optical character identification,
Recognition of face, speech recognition, Car license recognition and many smart fields such as unmanned.
Mode identification technology can be divided into two classes: the first is the recognition methods for having supervision, and this method is firstly the need of with big
Amount has the sample of calibration to be trained recognizer, then trained model could be used to identify unknown sample;
For second unsupervised recognition methods, this method is mainly some clustering algorithms, this method according to sample in space
Distribution clusters different types of sample.Usually, mode identification method mainly includes three parts: signal is located in advance
Reason, feature extraction and Classification and Identification.
Signal Pretreatment part main function is to provide the data for being easily handled and analyzing for subsequent characteristic extraction part.
For the Modulation Identification of wireless communication signals, the pretreatment of signal generally has following some methods: the frequency of radio-frequency front-end
Down coversion, the filtering of signal and amplification, estimating carrier frequencies, noise estimation etc., in certain circumstances, can be according to not yet
Same identification problem selects different Signal Pre-Processing Methods.
For general mode identification method, the selection and extraction of feature are very crucial parts.It is ideal special
Sign should have different modulated signals good differentiation to act on, in all cases all simultaneously for identical modulated signal
It should keep good consistency.For wireless communication signals, the modulation intelligence of signal is contained mainly in modulated signal
In amplitude, frequency and phase, thus theoretically for, utilize signal these features can identification signal modulation system.But
It is in the actual environment, since there are much noises and interference for wireless channel, directly to be generally difficult correctly to know using these features
The modulation system of level signal.Therefore, researchers at home and abroad, which have been working hard, finds other better characteristic parameters, enables
Inhibit noise and interference present in signal, to simplify the process of identification and improve recognition correct rate.
In Modulation identification technology, the recognition correct rate and computation complexity of recognizer have final recognition effect
Vital influence.The recognizer of mainstream nearly all identifies modulated signal using the method for machine learning at present, these
Recognizer can substantially be divided into these four types of methods: traditional decision-tree, clustering method, support vector machines side according to its recognition principle
Method and neural network method.
On the whole, the Modulation identification technology based on statistical pattern recognition method is the method for current mainstream, and at present
The method of domestic and international expert's extensive concern, major advantage are that the preprocessing part of signal is simple, and applicable modulation type is more, right
The dependence of modulated signal prior information is small, also there is good recognition effect when noise is relatively low, has preferable practical value
Deng;Major defect is the extraction of recognition effect excessively dependence characteristics, and the generally recognized algorithm is the learning algorithm for having supervision, is needed
Largely there is the training sample of label to be trained, and the training process of algorithm is more complicated.
In conclusion by adopting the above-described technical solution, the beneficial effects of the present invention are:
1, in the present invention, relative to traditional mode identification method, deep learning method is independent of manually for feature
Extraction, can also obtain preferable recognition effect.Therefore, more scientific research personnel begin to use the method conduct of deep learning
Modulation Recognition obtains preferably identification effect using the powerful feature learning ability of deep learning and nonlinear fitting ability
Fruit.
2, in the present invention, by building convolutional neural networks model, using training set training pattern, enable model
The internal characteristics of known modulation type signal or the changing rule of signal sequence data are practised, finally trained model exists
It is tested on test set, can effectively identify the modulation type of signal.It is special that this mode not only solves artificial extraction
Difficult problem is levied, and preferable recognition effect can be obtained, model proposed in this paper, the accuracy rate identified on test set
It can achieve 95% or more.
Detailed description of the invention
Fig. 1 is a kind of basic flow chart of signal modulate method proposed by the present invention;
Fig. 2 is a kind of convolutional neural networks structural schematic diagram of signal modulate method proposed by the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts all other
Embodiment shall fall within the protection scope of the present invention.
Convolution algorithm is one of functional analysis important operation, and the definition of convolution refers to that two can accumulate between real function
A kind of mathematical operation, the calculating process can indicate are as follows:
In above formula, the convolution algorithm of f*g representative function f and function g, wherein function f and function g are defined in real number field
Accumulate real function, can prove to belong to all t (- infinite ,+infinite) above-mentioned integral is existing, and convolution results
It is still the function that can be accumulated.Its physical significance is it is to be understood that the output at system a certain moment is by multiple input collective effects
Result.When we use computer processing data, usually discrete data is handled and is analyzed, correspondingly, we
The convolution of discrete form can be defined:
In above formula, function f and g indicate that discrete function, f*g are convolution results.
In convolutional neural networks, function f can be understood as original image vegetarian refreshments, and all original image vegetarian refreshments, which add up, is exactly
Original graph, function g are properly termed as position, we are known as convolution kernel to all positions altogether.
Such as in two dimensional image, we can indicate the two-dimensional convolution operation of image with following formula:
In above formula, Xi, j indicate the element of the i-th row jth in image;Wm, n indicate the element that m row n-th arranges in convolution kernel.
We introduce the turning operation of convolution kernel in convolution algorithm, although this turning operation have in many other fields it is very outstanding
Property, but there is no what actual effect for convolutional neural networks.Therefore, in actual convolutional neural networks
In, we generally use computing cross-correlation to replace convolution, and this computing cross-correlation and convolution algorithm are closely similar, only in operation
When to convolution kernel without overturning, calculation indicates as the following formula:
The structure of convolutional neural networks
What input layer typically entered is matrix data, such as in image domains, and the input of convolutional neural networks is usually
The picture element matrix of image, the latter linked of input layer is convolutional layer.
Convolutional layer is the core of entire convolutional neural networks, brings many excellent spies to convolutional neural networks
Property.In general, pass through different volumes comprising multiple convolution kernels in each convolutional layer for extracting different characteristics of image
The product available different characteristic layer of core, calculating process such as following formula:
In above formula, i, jy indicate that input passes through calculated characteristic layer result after convolutional layer;B is amount of bias, each convolution
Core all corresponds to the amount of bias of oneself;F () indicates activation primitive, selects activation letter of the Re LU function as neural network herein
Number.Re LU activation primitive was used successfully in Alex Net depth convolutional neural networks model in 2012, and verify its
Effect is very good in deeper convolutional neural networks, and Re LU activation primitive is in various convolutional neural networks models after this
It is widely used [40].Re LU function is very simple, and expression formula is as follows:
F (x)=max (0, x)
Re LU activation primitive is that convolutional neural networks model joined non-linear factor, can largely increase model
Nonlinear fitting ability.Meanwhile Re LU activation primitive calculating speed very block, when input is greater than zero, which is led
Permanent number is 1, is very suitable to the study based on gradient descent algorithm;When input be negative when, Re LU activation primitive output and
Gradient is 0, to make neural network weight that can not update in the training process, this can effectively increase the sparsity of model.
It, can connection pool layer after one or more convolution algorithm usually in convolutional neural networks.Pond function makes
The output of the position is calculated with the general evaluation system feature of a certain position of characteristic layer and its adjacent output, it can be using similar one
The structure of a filter completes the process of the forward calculation of pond layer, and common pond function has maximum pond function, average pond
Change function etc..
The input of convolutional neural networks by the alternating of multiple convolutional layers and pond layer propagation after, for identify class problem
For, it generally will use one layer or multilayer fully-connected network for characteristic layer and be converted into classification results.One can consider that convolution
Layer and pond layer complete automatically extracting for feature, and the information in image has been conceptualized as high-level information, last only to need
To utilize full Connection Neural Network from global comprehensive characteristics, it will be able to complete identification mission
The foregoing is only a preferred embodiment of the present invention, but scope of protection of the present invention is not limited thereto,
Anyone skilled in the art in the technical scope disclosed by the present invention, according to the technique and scheme of the present invention and its
Inventive concept is subject to equivalent substitution or change, should be covered by the protection scope of the present invention.
Claims (6)
1. a kind of signal modulate method, which is characterized in that including the recognition methods based on decision theory and based on statistics mould
Formula identification;The modulation system of the signal of communication received is identified in the presence of noise and interference, thus for further place
Reason and analysis signal of communication provide foundation.
2. a kind of signal modulate method according to claim 1, which is characterized in that the knowledge based on decision theory
Other method regards identification problem as multiple hypothesis test problem, carries out theory deduction to the test statistics of modulated signal, uses
Minimum likelihood method is found suitable threshold value and is differentiated to modulation.
3. a kind of signal modulate method according to claim 1, which is characterized in that described to be based on pattern recognition theory
Modulation Identification method be current relatively mainstream method, the method regards Modulation Identification problem as a kind of typical pattern-recognition
Problem.
4. a kind of typical pattern recognition problem of signal modulate method according to claim 3, which is characterized in that
The pattern-recognition refers to: disaggregatedly being identified things by software and hardware algorithm.
5. a kind of signal modulate method according to claim 4, which is characterized in that the mode identification technology can be with
Be divided into two classes: the first is the recognition methods for having supervision, and this method is firstly the need of with largely having the sample of calibration to recognizer
It is trained, then trained model could be used to identify unknown sample;Second is unsupervised recognition methods,
This method is mainly some clustering algorithms, and this method gathers different types of sample according to the distribution of sample in space
Class.
6. a kind of signal modulate method according to claim 4, which is characterized in that the mode identification method is main
Including three parts: Signal Pretreatment, feature extraction and Classification and Identification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910280889.6A CN110048978A (en) | 2019-04-09 | 2019-04-09 | A kind of signal modulate method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910280889.6A CN110048978A (en) | 2019-04-09 | 2019-04-09 | A kind of signal modulate method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110048978A true CN110048978A (en) | 2019-07-23 |
Family
ID=67276635
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910280889.6A Pending CN110048978A (en) | 2019-04-09 | 2019-04-09 | A kind of signal modulate method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110048978A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111585922A (en) * | 2020-03-23 | 2020-08-25 | 成都奥特为科技有限公司 | Modulation mode identification method based on convolutional neural network |
CN111709496A (en) * | 2020-08-18 | 2020-09-25 | 北京邮电大学 | Modulation mode recognition and model training method and device based on neural network |
CN112202527A (en) * | 2020-10-01 | 2021-01-08 | 西北工业大学 | Intelligent electromagnetic signal identification system interference method based on momentum gradient disturbance |
CN113518050A (en) * | 2021-06-24 | 2021-10-19 | 华东交通大学 | Modulation identification method, system, readable storage medium and device |
CN114065823A (en) * | 2021-12-02 | 2022-02-18 | 中国人民解放军国防科技大学 | Modulation signal identification method and system based on sparse deep neural network |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101667252A (en) * | 2009-10-15 | 2010-03-10 | 哈尔滨工业大学 | Classification and identification method for communication signal modulating mode based on ART2A-DWNN |
CN102497343A (en) * | 2011-11-25 | 2012-06-13 | 南京邮电大学 | Combined modulation recognition method based on clustering and support vector machine |
CN107547460A (en) * | 2017-08-21 | 2018-01-05 | 西安电子科技大学 | Radio communication Modulation Signals Recognition method based on deep learning |
CN109523023A (en) * | 2018-11-16 | 2019-03-26 | 泰山学院 | A kind of deep learning network and system for subsurface communication Modulation Identification |
-
2019
- 2019-04-09 CN CN201910280889.6A patent/CN110048978A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101667252A (en) * | 2009-10-15 | 2010-03-10 | 哈尔滨工业大学 | Classification and identification method for communication signal modulating mode based on ART2A-DWNN |
CN102497343A (en) * | 2011-11-25 | 2012-06-13 | 南京邮电大学 | Combined modulation recognition method based on clustering and support vector machine |
CN107547460A (en) * | 2017-08-21 | 2018-01-05 | 西安电子科技大学 | Radio communication Modulation Signals Recognition method based on deep learning |
CN109523023A (en) * | 2018-11-16 | 2019-03-26 | 泰山学院 | A kind of deep learning network and system for subsurface communication Modulation Identification |
Non-Patent Citations (2)
Title |
---|
李佳宸: "基于深度学习的数字调制信号识别方法研究", 《中国优秀硕士学位论文全文数据库_信息科技辑》 * |
肖嵩: "《计算机图形学原理及应用》", 30 June 2014 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111585922A (en) * | 2020-03-23 | 2020-08-25 | 成都奥特为科技有限公司 | Modulation mode identification method based on convolutional neural network |
CN111709496A (en) * | 2020-08-18 | 2020-09-25 | 北京邮电大学 | Modulation mode recognition and model training method and device based on neural network |
CN112202527A (en) * | 2020-10-01 | 2021-01-08 | 西北工业大学 | Intelligent electromagnetic signal identification system interference method based on momentum gradient disturbance |
CN112202527B (en) * | 2020-10-01 | 2022-09-13 | 西北工业大学 | Intelligent electromagnetic signal identification system interference method based on momentum gradient disturbance |
CN113518050A (en) * | 2021-06-24 | 2021-10-19 | 华东交通大学 | Modulation identification method, system, readable storage medium and device |
CN113518050B (en) * | 2021-06-24 | 2021-12-24 | 华东交通大学 | Modulation identification method, system, readable storage medium and device |
CN114065823A (en) * | 2021-12-02 | 2022-02-18 | 中国人民解放军国防科技大学 | Modulation signal identification method and system based on sparse deep neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110048978A (en) | A kind of signal modulate method | |
CN107220600B (en) | A kind of Picture Generation Method and generation confrontation network based on deep learning | |
Huynh-The et al. | Accurate LPI radar waveform recognition with CWD-TFA for deep convolutional network | |
CN111767848B (en) | Radiation source individual identification method based on multi-domain feature fusion | |
CN105512676A (en) | Food recognition method at intelligent terminal | |
CN114564982B (en) | Automatic identification method for radar signal modulation type | |
CN102096824B (en) | Multi-spectral image ship detection method based on selective visual attention mechanism | |
CN111832417A (en) | Signal modulation pattern recognition method based on CNN-LSTM model and transfer learning | |
CN111461037B (en) | End-to-end gesture recognition method based on FMCW radar | |
CN106529570B (en) | Image classification method based on depth ridge ripple neural network | |
CN102136024A (en) | Biometric feature identification performance assessment and diagnosis optimizing system | |
CN104751186A (en) | Iris image quality classification method based on BP (back propagation) network and wavelet transformation | |
Guo et al. | Radar target recognition based on feature Pyramid fusion lightweight CNN | |
CN109685071A (en) | Brain electricity classification method based on the study of common space pattern feature width | |
CN110503082A (en) | A kind of model training method and relevant apparatus based on deep learning | |
Gao et al. | Vacl: Variance-aware cross-layer regularization for pruning deep residual networks | |
CN112084842B (en) | Hydrological remote sensing image target recognition method based on depth semantic model | |
CN113822201A (en) | Deep learning method for underwater object shape recognition based on flow field velocity component time course | |
CN114980122A (en) | Small sample radio frequency fingerprint intelligent identification system and method | |
Feng et al. | FCGCN: Feature Correlation Graph Convolution Network for Few-Shot Individual Identification | |
CN109145810A (en) | Details in fingerprint point detecting method, device, equipment, system and storage medium | |
CN116010783B (en) | RSVP weak hidden target induced electroencephalogram identification method, device and storage medium | |
Belan | Specialized cellular structures for image contour analysis | |
CN101964055A (en) | Visual perception mechansim simulation natural scene type identification method | |
Gai et al. | Spectrum sensing method based on residual cellular network |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190723 |