CN110996343B - Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network - Google Patents
Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network Download PDFInfo
- Publication number
- CN110996343B CN110996343B CN201911307930.0A CN201911307930A CN110996343B CN 110996343 B CN110996343 B CN 110996343B CN 201911307930 A CN201911307930 A CN 201911307930A CN 110996343 B CN110996343 B CN 110996343B
- Authority
- CN
- China
- Prior art keywords
- interference
- model
- neural network
- recognition
- receiver
- 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
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/06—Testing, supervising or monitoring using simulated traffic
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Image Analysis (AREA)
- Noise Elimination (AREA)
Abstract
An interference recognition model and an intelligent recognition algorithm based on a depth convolution neural network are characterized in that: the receiver for interference recognition collects data of interference signals sent by a single or a plurality of interference machines, the frequency spectrum waterfall of the receiver as a receiving end is used as a network input layer to carry out a plurality of exercises, after a training model with enough fitting degree is achieved, the training model and the frequency spectrum waterfall of the receiving end are saved according to the exercises, and the training model and the frequency spectrum waterfall of the receiving end are used as network inputs to carry out online recognition. The model is complete by combining other structures or methods, the physical meaning is clear, the design algorithm is reasonable and effective, and the interference recognition scene based on the deep convolutional neural network algorithm can be better depicted.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to an intelligent recognition system and a recognition method of an interference recognition model based on a deep convolutional neural network
Background
Communication countermeasure refers to electromagnetic fight by the opponent and the opponent in the field of radio communication by using common radio communication equipment and special communication countermeasure equipment. As frequency-consuming devices in a communication environment increase, communication has become a popular topic, and intensive research has become important in how communication parties can better avoid the influence of enemies and own parties on own frequency consumption. Along with the rapid development of machine learning, the intelligent level of various devices is continuously improved, and intelligent interference and intelligent anti-interference are also one of the research subjects in the communication field. Communication immunity includes two aspects: firstly, a certain means is utilized to disturb or destroy the investigation equipment of the enemy; secondly, electromagnetic reinforcement and interference suppression measures are adopted for own communication, so that anti-reconnaissance and anti-interference capabilities of own communication equipment are enhanced. The anti-interference capability mainly refers to the capability of inhibiting communication interference of an adversary. However, in the current research situation, most of research on the identification of interference signals is focused on the feature extraction of different interference signals of different communication systems, and less research on classification algorithms is performed, which also shows the importance of feature extraction in the method of interference identification based on feature extraction, and a large research space is provided for the general interference signal classification algorithms such as convolutional neural networks. The receiver combines the received spectrum information with the intelligent agent into a new method of interference pattern recognition.
Because of the complex dynamic characteristics of communication and interference in the environment, an intelligent interference recognition method becomes one of important solutions to the problem of interference recognition, a deep convolutional neural network becomes a popular classification tool (references: leCun Y, bottou L, bengio Y, et al., gradient-based learning applied to document recognition [ J ]. Proceedings of the IEEE,1998,86 (11): 2278-2324 ]), but how to better model to adapt to the problem to be solved is also an important point of research.
Disclosure of Invention
In order to solve the problems, the invention provides an intelligent recognition system and an intelligent recognition method for an interference recognition model based on a deep convolutional neural network, which well describe an interference recognition scene based on a deep learning algorithm.
In order to overcome the defects in the prior art, the invention provides an intelligent recognition system and a recognition method of an interference recognition model based on a deep convolutional neural network, which concretely comprises the following steps:
an intelligent recognition system based on an interference recognition model of a deep convolutional neural network, comprising:
the interference identification model based on the depth convolution neural network is characterized as follows: the receiver for interference recognition collects data of interference signals sent by a single or a plurality of interference machines, the frequency spectrum waterfall of the receiver as a receiving end is used as a network input layer to carry out a plurality of exercises, after a training model with enough fitting degree is achieved, the training model and the frequency spectrum waterfall of the receiving end are saved according to the exercises, and the training model and the frequency spectrum waterfall of the receiving end are used as network inputs to carry out online recognition.
Power spectral density s of the receiver as receiving end t (f) The expression of (2) is shown in formula (1):
in equation (1), the jth jammer may select its frequency f at the t-th sampling t j And a power spectral density function
When the receiver perceives the full frequency band, the interference signal and the noise are received simultaneously; g j Is the channel link gain from the jammer to the receiver, f is the frequency of the receiver, n (f) is the PSD function of the noise, J is a positive integer and is the number of jammers, t is a positive integer and represents the sequence number of receiver samples.
Discrete value sampling spectrum s in interference identification model based on depth convolution neural network t,i Defined as formula (2):
where Δf is the spectral resolution, f L Initial frequency s representing interference bandwidth of jammer t,i A sampling spectrum representing the ith sample received by the receiver at the nth sample, the frequency vector perceived by the receiver being s t ＝s t,1 ,s t,2 ,…s t,N N is the number of samples sampled at the t-th time and i is a positive integer.
In a dynamic unknown environment, taking a frequency spectrum waterfall diagram of the receiver as a receiving end as an input layer of a neural network, and finally obtaining an interference recognition model after convolution extracting features and monitoring and training for a plurality of times, wherein the method comprises the following steps of:
in a dynamically unknown communication environment, the interference recognition problem in the model is modeled as a convolution layer extraction process, and the complex interference pattern existing in the environment has close relation with the historical information, so that the environment state S t Defined as S t ＝{s t ,s t-1 ,…,s t-T+1 Where T represents the number of historical states tracked, then S t Is a two-dimensional matrix T x N thermodynamic block diagram containing time and frequency domain information, thereby constructing a spectral waterfall diagram.
In the historical environmental state information, the data is used as an input layer for designing a CNN model to carry out iterative training, x is defined as a data vector from input data to the input layer, and x is defined as a data vector from the input layer to the input layer i For the ith element in x and representing the data input to the input layer the ith time, y is the label vector corresponding to x, optimize the objective
Is shown in formula (3):
An intelligent recognition method based on a deep convolutional neural network interference recognition model comprises the following steps:
step 1, initializing, namely setting the weight and the bias value of each network node in the deep convolutional neural network to be random and meet the value distributed according to the probability, and sensing an initial environment;
step 2, inputting the data collected by the receiver into a deep convolutional neural network which is trained as a supervised neural network, wherein the input data and the labels have a certain loss function
Thereby providing a structural risk function R srm (f) Is shown in formula (5):
λ is a set coefficient greater than 0, J (f) is the complexity of the model;
wherein eta is learning rate omega t Weighting the empirical risk function for the t-th sample, ω being such that f when the learning rate is sufficiently small ω (x i ) The optimal solution is reached, and the algorithm is linearly converged;
and 4, carrying out online identification on the trained interference identification model based on the deep convolutional neural network by taking in the reserved test data set sample.
The beneficial effects of the invention are as follows:
(1) The traditional feature recognition is constructed into a convolutional neural network, and the calculation complexity is reduced under the condition that the performance is not affected;
(2) The model is perfect, the physical meaning is clear, and the intelligent recognition algorithm based on the deep convolutional neural network is provided to realize the considerable recognition rate of the proposed model;
(3) The method can effectively identify the interference mode and well delineate the interference identification scene based on the deep convolutional neural network algorithm.
Drawings
Fig. 1 is a model diagram of a deep convolutional neural network interference recognition system of the present invention.
Fig. 2 is a schematic diagram of the structure of the convolutional neural network for interference recognition of the present invention.
Fig. 3 is a flowchart of a training update process for convolutional neural network to interference recognition in the present invention.
Fig. 4 is a graph showing the identification and accuracy of the constant frequency interference pattern in embodiment 1 of the present invention.
Fig. 5 is a graph of the identification and accuracy of the interference pattern of the sweep in embodiment 2 of the present invention.
Detailed Description
The invention provides a deep convolutional neural network interference recognition system and an intelligent recognition algorithm, and aims to provide a scheme for solving the intelligent recognition problem. The method is based on a machine learning algorithm, takes a frequency spectrum waterfall diagram of a receiving end as an input layer of a neural network, adopts a deep convolution neural network to perform training fitting on acquired data, and performs supervised training on the acquired data; then, the network model is trained and stored for a plurality of times, so that the interference mode can be identified in real time.
Fig. 1 is a diagram of an interference recognition system model. In the model, a receiver collects data of a single or a plurality of jammers, a frequency spectrum waterfall diagram of a receiving end is used as a network input layer to carry out a plurality of exercises, and the model can be stored after the model is trained to reach enough fitting degree.
Fig. 2 is a model of an interference-recognition convolutional neural network. The interference recognition network architecture adopts a convolutional neural network framework, is input into a frequency spectrum waterfall diagram, obtains recognition data with labels through one-layer convolution one-layer pooling and two-layer full connection, and then achieves the best fitting degree through a plurality of times of training. The training model is saved and real-time interference pattern recognition can be performed using the input data and the model.
The model was characterized as follows: the single receiver collects data for single or multiple jammers, the frequency spectrum waterfall of the receiving end is used as a network input layer to carry out a plurality of exercises, after a training model with enough fitting degree is achieved, the frequency spectrum waterfall of the receiving end is used as input in a dynamic position environment, and therefore online mode identification is achieved.
Fig. 3 is a modeling and updating process of an interference recognition convolutional neural network, when a proxy receives interference data of a receiving end, the interference data is brought into the designed convolutional neural network, the fitting degree of the data is observed through a plurality of iterations, a trained model is obtained after the network structure is optimized and the weight is trained, and finally, the collected real data is input as the model, so that a recognition result can be obtained.
Based on the Le-Net5 model and combined with the intelligent interference recognition problem, the invention considers factors such as huge interference data, difficult data preprocessing, time consumption of learning and calculation and the like, adopts a deep convolutional neural network to process and classify the interference data, and continuously updates the network weight in the process to obtain an optimal model and is applied to actual data.
The invention will be further described with reference to the accompanying drawings and examples:
as shown in fig. 1-5, the deep convolutional neural network interference recognition model comprises:
the interference identification model based on the depth convolution neural network is characterized as follows: the receiver for interference recognition collects data of interference signals sent by a single or a plurality of interference machines, the frequency spectrum waterfall of the receiver as a receiving end is used as a network input layer to carry out a plurality of exercises, after a training model with enough fitting degree is achieved, the training model and the frequency spectrum waterfall of the receiving end are saved according to the exercises, and the training model and the frequency spectrum waterfall of the receiving end are used as network inputs to carry out online recognition.
Power spectral density s of the receiver as receiving end t (f) The expression of (2) is shown in formula (1):
in equation (1), the jth jammer may select its frequency f at the t-th sampling t j And a power spectral density function
When the receiver perceives the full frequency band, the interference signal and the noise are received simultaneously; g j Is the channel link gain from the jammer to the receiver, f is the frequency of the receiver, n (f) is the PSD function of the noise, J is a positive integer and is the number of jammers, t is a positive integer and represents the sequence number of receiver samples.
Discrete value sampling spectrum s in interference identification model based on depth convolution neural network t,i Defined as formula (2):
where Δf is the spectral resolution, f L Initial frequency s representing interference bandwidth of jammer t,i A sampling spectrum representing the ith sample received by the receiver at the nth sample, the frequency vector perceived by the receiver being s t ＝s t,1 ,s t,2 ,…s t,N N is the number of samples sampled at the t-th time and i is a positive integer.
In a dynamic unknown environment, taking a frequency spectrum waterfall diagram of the receiver as a receiving end as an input layer of a neural network, and finally obtaining an interference recognition model after convolution extracting features and monitoring and training for a plurality of times, wherein the method comprises the following steps of:
in a dynamically unknown communication environment, the interference recognition problem in the model is modeled as a convolution layer extraction process, and the complex interference pattern existing in the environment has close relation with the historical information, so that the environment state S t Defined as S t ＝{s t ,s t-1 ,…,s t-T+1 Where T represents the number of historical states tracked, thenS t Is a two-dimensional matrix T x N thermodynamic block diagram containing time and frequency domain information, thereby constructing a spectral waterfall diagram.
In the historical environmental state information, the data is used as an input layer for designing a CNN model to carry out iterative training, x is defined as a data vector from input data to the input layer, and x is defined as a data vector from the input layer to the input layer i For the ith element in x and representing the data input to the input layer the ith time, y is the label vector corresponding to x, optimize the objective
Is shown in formula (3):
The intelligent recognition method based on the deep convolutional neural network interference recognition model comprises the following steps:
step 1, initializing, namely setting the weight and the bias value of each network node in the deep convolutional neural network to be random and meet the value distributed according to the probability, and sensing an initial environment;
step 2, inputting the data acquired by the receiver into a deep convolutional neural network which is trained as a supervised neural network, and inputting the data into the deep convolutional neural networkHas a certain loss function on the data and the label
Thereby providing a structural risk function R srm (f) Is shown in formula (5):
λ is a set coefficient greater than 0, J (f) is the complexity of the model;
wherein eta is learning rate omega t Weighting the empirical risk function for the t-th sample, ω being such that f when the learning rate is sufficiently small ω (x i ) The optimal solution is reached, and the algorithm is linearly converged;
and 4, carrying out online identification on the trained interference identification model based on the deep convolutional neural network by taking in the reserved test data set sample.
Because the calculation classification is less, the complex convolutional neural network model can be simplified to achieve the effect of reducing the calculation amount.
The invention is further illustrated by the following examples:
example 1
The first embodiment of the invention is specifically described below, the system simulation adopts python language, and the parameter setting does not affect the generality based on the tensorflow deep learning framework. This embodiment verifies the validity of the proposed model and method and fig. 4 verifies the validity of the fixed frequency interference pattern. The parameters are set to be that the frequency band of interference is 20MHz and the frequency spectrumThe perceived frequency resolution is 100kHz, the receiver performs full-band sensing once for no 1ms and keeps the perceived spectrum data for 200ms, thus S t The matrix size is 200×200, the bandwidth of the interference signal is 4MHz, the signal waveform is a raised cosine wave, and the roll-off coefficient is α=0.5. The interference power is 30dBm. In embodiment 1, we consider 2 fixed frequency interference modes:
1. the single tone interference, the interference frequency is 2MHz.
2. Multitone interference, the fixed interference frequency is 2MHz,10MHz,18MHz respectively.
Fig. 4 is a waterfall diagram of a frequency spectrum of a constant-frequency interference pattern in embodiment 1 of the present invention, and it can be seen from the figure that the interference appears as a plurality of vertical stripes, and the right side is a training recognition rate graph in the interference pattern, and the accuracy fluctuates above 0.9.
Example 2
The second embodiment of the invention is specifically described below, the system simulation adopts python language, and the parameter setting does not affect the generality based on the tensorflow deep learning framework. This example verifies the validity of the proposed model and method, fig. 4 verifies the validity of the fixed frequency interference pattern, and fig. 5 verifies the validity of the swept frequency interference identification. The parameters are set to 20MHz of the interfering frequency band, 100kHz of the frequency resolution of the spectrum sensing, the receiver performs full-band sensing every 1ms, and the sensed spectrum data is kept for 200ms, thus S t The matrix size is 200×200, the bandwidth of the interference signal is 4MHz, the signal waveform is a raised cosine wave, and the roll-off coefficient is α=0.5. The interference power is 30dBm. In example 2, we consider the swept interference mode: the frequency sweep interference and the frequency sweep speed are 0.2GHz/s and 0.5GHz/s.
Fig. 5 is a waterfall diagram of the frequency spectrum of the swept interference pattern in example 2 of the present invention, from which it can be seen that the interference appears as a diagonal plot due to the linear frequency variation. The right side is the training accuracy graph in this interference mode. Therefore, the model has higher recognition degree, and the precision fluctuates above 0.9.
In summary, the deep convolutional neural network interference recognition model provided by the invention has the advantages that the problems of complex data preprocessing, complex network structure and huge calculation amount in the interference recognition are fully considered, and the deep convolutional neural network interference recognition model has more practical significance than the traditional model; the intelligent recognition algorithm based on the depth convolution neural network interference recognition model can realize the effectiveness solution of the proposed model, has higher recognition rate from the accuracy curve, and effectively handles the interference pattern recognition.
While the invention has been described by way of examples, it will be understood by those skilled in the art that the present disclosure is not limited to the examples described above, and that various changes, modifications and substitutions may be made without departing from the scope of the invention.
Claims (6)
1. An intelligent recognition method based on a deep convolutional neural network interference recognition model is characterized by comprising the following steps:
step 1, initializing, namely setting the weight and the bias value of each network node in the deep convolutional neural network to be random and meet the value distributed according to the probability, and sensing an initial environment;
step 2, inputting the data collected by the receiver into a deep convolutional neural network which is trained as a supervised neural network, wherein the input data and the labels have a certain loss function
Thereby providing a structural risk function R srm (f) Is shown in formula (5):
λ is a set coefficient greater than 0, J (f) is the complexity of the model;
step 3, updating the deep convolutional neural network by using a formula (6) after the set training times, repeating the step 2, and correspondingly optimizing the model to minimize the structural risk function and the empirical risk function weight omega t The update formula is:
wherein eta is learning rate omega t Weighting the empirical risk function for the t-th sample, ω being such that f when the learning rate is sufficiently small ω (x i ) The optimal solution is reached, and the algorithm is linearly converged;
and 4, carrying out online identification on the trained interference identification model based on the deep convolutional neural network by taking in the reserved test data set sample.
2. An intelligent recognition system based on the intelligent recognition method of claim 1, comprising: the interference identification model based on the deep convolutional neural network is characterized as follows: the receiver for interference recognition collects data of interference signals sent by a single or a plurality of interference machines, the frequency spectrum waterfall of the receiver as a receiving end is used as a network input layer to carry out a plurality of exercises, after a training model with enough fitting degree is achieved, the training model and the frequency spectrum waterfall of the receiving end are saved according to the exercises, and the training model and the frequency spectrum waterfall of the receiving end are used as network inputs to carry out online recognition.
3. The intelligent recognition system according to claim 2, wherein the power spectral density s of the receiver as a receiving end t (f) The expression of (2) is shown in formula (1):
in equation (1), the jth jammer may select its frequency at the t-th sampling
And Power spectral Density function->
When the receiver perceives the full frequency band, the interference signal and the noise are received simultaneously; g j Is the channel link gain from the jammer to the receiver, f is the frequency of the receiver, n (f) is the PSD function of the noise, J is a positive integer and is the number of jammers, t is a positive integer and represents the sequence number of receiver samples.
4. The intelligent recognition system according to claim 2, wherein the discrete value sampling spectrum s in the depth convolutional neural network-based interference recognition model t,i Defined as formula (2):
where Δf is the spectral resolution, f L Initial frequency s representing interference bandwidth of jammer t,i A sampling spectrum representing the ith sample received by the receiver at the nth sample, the frequency vector perceived by the receiver being s t ＝s t,1 ,s t,2 ,…s t,N N is the number of samples sampled at the t-th time and i is a positive integer.
5. The intelligent recognition system according to claim 2, wherein in a dynamic unknown environment, a spectrum waterfall diagram of the receiver as a receiving end is used as an input layer of a neural network, and the interference recognition model is finally obtained after convolution extracting features and a plurality of supervision training steps, wherein the method comprises the following steps:
in a dynamically unknown communication environment, the interference recognition problem in the model is modeled as a convolution layer extraction process, and the complex interference pattern existing in the environment has close relation with the historical information, so that the environment state S t Defined as S t ＝{s t ,s t-1 ,…,s t-T+1 Where T represents the number of historical states tracked, then S t Is a two-dimensional matrix T x N thermodynamic block diagram containing time and frequency domain information, thereby constructing a spectral waterfall diagram.
6. The intelligent recognition system of claim 2, wherein in the historical environmental state information, the data is iteratively trained as an input layer for which the CNN model is designed, defining x as a data vector for inputting the data to the input layer, and x i For the ith element in x and representing the data input to the input layer the ith time, y is the label vector corresponding to x, optimize the objective
Is shown in formula (3):
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911307930.0A CN110996343B (en) | 2019-12-18 | 2019-12-18 | Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911307930.0A CN110996343B (en) | 2019-12-18 | 2019-12-18 | Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110996343A CN110996343A (en) | 2020-04-10 |
| CN110996343B true CN110996343B (en) | 2023-06-30 |
Family
ID=70095337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911307930.0A Active CN110996343B (en) | 2019-12-18 | 2019-12-18 | Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110996343B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111562597A (en) * | 2020-06-02 | 2020-08-21 | 南京敏智达科技有限公司 | Beidou satellite navigation interference source identification method based on BP neural network |
| CN111669248A (en) * | 2020-06-04 | 2020-09-15 | 上海特金无线技术有限公司 | Unmanned aerial vehicle signal suppression method and device, electronic equipment and storage medium |
| CN111786738B (en) * | 2020-07-01 | 2022-06-03 | 中国人民解放军陆军工程大学 | Anti-interference learning network structure based on long-term and short-term memory and learning method |
| CN112435356B (en) * | 2020-11-04 | 2021-11-26 | 南京航天工业科技有限公司 | ETC interference signal identification method and detection system |
| CN112491442B (en) * | 2020-11-17 | 2021-12-28 | 中山大学 | Self-interference elimination method and device |
| CN113067653B (en) * | 2021-03-17 | 2022-01-14 | 北京邮电大学 | Spectrum sensing method and device, electronic equipment and medium |
| CN113138201B (en) * | 2021-03-24 | 2022-05-20 | 北京大学 | Metamaterial Internet of things system and method for wireless passive environment state detection |
| CN114358064B (en) * | 2021-12-23 | 2022-06-21 | 中国人民解放军海军工程大学 | Interference detection device and method based on deep support vector data description |
| CN114598403B (en) * | 2022-03-31 | 2024-03-12 | 中国人民解放军陆军工程大学 | Data link broadband noise electromagnetic signal interference prediction method and system |
| CN115243288B (en) * | 2022-07-11 | 2024-05-03 | 中国人民解放军国防科技大学 | Interference identification method and device based on multi-node cooperative sensing |
| CN116271539B (en) * | 2023-05-15 | 2023-09-08 | 苏州维伟思医疗科技有限公司 | Method for identifying shockable rhythm, wearable cardioverter-defibrillator and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764013A (en) * | 2018-03-28 | 2018-11-06 | 中国科学院软件研究所 | A kind of automatic Communication Signals Recognition based on end-to-end convolutional neural networks |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108509911B (en) * | 2018-04-03 | 2020-06-12 | 电子科技大学 | Interference signal identification method based on convolutional neural network |
| CN108777872B (en) * | 2018-05-22 | 2020-01-24 | 中国人民解放军陆军工程大学 | Intelligent anti-interference method and intelligent anti-interference system based on deep Q neural network anti-interference model |
| CN109961017A (en) * | 2019-02-26 | 2019-07-02 | 杭州电子科技大学 | A kind of cardiechema signals classification method based on convolution loop neural network |
-
2019
- 2019-12-18 CN CN201911307930.0A patent/CN110996343B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764013A (en) * | 2018-03-28 | 2018-11-06 | 中国科学院软件研究所 | A kind of automatic Communication Signals Recognition based on end-to-end convolutional neural networks |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110996343A (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110996343B (en) | Intelligent recognition system and recognition method of interference recognition model based on deep convolutional neural network | |
| CN108777872B (en) | Intelligent anti-interference method and intelligent anti-interference system based on deep Q neural network anti-interference model | |
| CN111901862B (en) | User clustering and power distribution method, device and medium based on deep Q network | |
| CN111970072B (en) | Broadband anti-interference system and method based on deep reinforcement learning | |
| CN108508411B (en) | Passive radar external radiation source signal identification method based on transfer learning | |
| CN113435247B (en) | Intelligent recognition method, system and terminal for communication interference | |
| CN110166154B (en) | Software radio frequency spectrum monitoring and identifying method based on neural network | |
| CN112910811B (en) | Blind modulation identification method and device under unknown noise level condition based on joint learning | |
| CN113406579B (en) | Camouflage interference waveform generation method based on deep reinforcement learning | |
| Shen et al. | Node identification in wireless network based on convolutional neural network | |
| CN112115830B (en) | Target distributed fusion recognition method based on bit domain feature extraction | |
| CN111786738B (en) | Anti-interference learning network structure based on long-term and short-term memory and learning method | |
| CN111222640A (en) | Signal recognition convolutional neural network convolutional kernel partition pruning method | |
| CN114764577A (en) | Lightweight modulation recognition model based on deep neural network and method thereof | |
| Cai et al. | Jamming pattern recognition using spectrum waterfall: A deep learning method | |
| Cai et al. | An integrated incremental self-organizing map and hierarchical neural network approach for cognitive radio learning | |
| Feng et al. | FCGCN: Feature Correlation Graph Convolution Network for Few-Shot Individual Identification | |
| Zhang et al. | Automatic classification and recognition method based on partially-connected differentiable architecture search for ISAC systems | |
| CN116866895A (en) | Intelligent countering method based on neural virtual self-game | |
| CN116866048A (en) | Anti-interference zero-and Markov game model and maximum and minimum depth Q learning method | |
| CN113095215B (en) | Solar radio filtering method and system based on improved LSTM network | |
| CN115276858A (en) | Dynamic spectrum multi-domain anti-interference method and system based on cognitive anti-interference model | |
| Yin et al. | Few-Shot Domain Adaption-Based Specific Emitter Identification Under Varying Modulation | |
| CN114298166A (en) | Spectrum availability prediction method and system based on wireless communication network | |
| Olausson et al. | Disturbance detection and classification in large microwave networks |
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 |