CN106961684A - The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply - Google Patents

The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply Download PDF

Info

Publication number
CN106961684A
CN106961684A CN201710181997.9A CN201710181997A CN106961684A CN 106961684 A CN106961684 A CN 106961684A CN 201710181997 A CN201710181997 A CN 201710181997A CN 106961684 A CN106961684 A CN 106961684A
Authority
CN
China
Prior art keywords
cognitive radio
deeply
convolutional neural
neural networks
secondary user
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
Application number
CN201710181997.9A
Other languages
Chinese (zh)
Inventor
肖亮
韩国安
李炎达
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xiamen University
Original Assignee
Xiamen University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xiamen University filed Critical Xiamen University
Priority to CN201710181997.9A priority Critical patent/CN106961684A/en
Publication of CN106961684A publication Critical patent/CN106961684A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/14—Spectrum sharing arrangements between different networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply, cognitive radio time user is in the state of unknown disturbances machine attack mode and wireless channel environment, access state, the signal interference ratio of wireless signal of cognitive radio primary user is observed, the one suitable frequency sending signal in disturbed region or selection where leaving is decided whether using deeply study mechanism.Learn with reference to depth convolutional neural networks and Q, using Q study in the wireless optimal Anti-interference Strategy of dynamic game learning, observer state and acquisition benefit are inputted into depth convolutional neural networks and accelerate pace of learning as training set.The mechanism learnt using deeply, improves the communication efficiency that cognitive radio resists hostility jammer under the wireless network environment scene of dynamic change.Artificial neural network, which can be overcome, to be needed to need in the training process first to classify and Q learning algorithms pace of learning meeting rapid decrease problem in the case where state set and behavior aggregate dimension are big to data.

Description

The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply
Technical field
The present invention relates to wireless network secure, more particularly, to the cognitive radio null tone two dimension learnt based on deeply Meaning interference method against the enemy.
Background technology
With the fast development of radio communication, increasingly serious, the cognition wireless the problems such as shortage of frequency spectrum resource and utilization rate The proposition of electric (Cognitive Radio, CR) technology can effectively improve the utilization rate of frequency spectrum.Opened because cognitive radio is used The frequency spectrum and dynamic access mode of formula are put, the attack easily disturbed by hostility, its safety problem is urgently to be resolved hurrily.
Hostility jammer prevents validated user from carrying out normal data forwarding by taking network node communication channel, after And start Denial of Service attack (DoS).Spread spectrum communication can effectively resist interference as traditional Anti-Jamming Technique, and frequency hopping, DSSS and main 3 kinds of modes (Viterbi A J.Spread spectrum that linear frequency modulation spread spectrum is spread spectrum communications:myths and realities[J].Communications Magazine,IEEE,2002,40 (5):34-41)。
However, with the development of software radio, the attack mode variation all the more of jammer and intelligent, tradition Anti-Jamming Technique to resist this kind of attack performance not good.Therefore, artificial neural network, intensified learning are used for attacking for learning interference machine Blow mode simultaneously implements the research of Anti-interference Strategy and obtains the extensive concern of domestic and foreign scholars.(Y.Wu,B.Wang,K.J.R.Liu, and T.C.Clancy,“Anti-jamming games in multi-channel cognitive radio Networks, " IEEE J.Sel.Areas Commun., vol.30, no.1, pp.4-15,2012) one kind is proposed based on Q The channel access strategy of habit.(K.Dabcevic,A.Betancourt,L.Marcenaro,and C.S.Regazzoni,“A fictitious play-based game-theoretical approach to alleviating jamming attacks for cognitive radios,”IEEE Int’l Conf.Acoustich,Speech and Signal Processing (ICASSP), pp.8208-8212,2014) propose a kind of anti-interference power allocation strategy of optimization.
However, artificial neural network needs first to classify to data in the training process.Meanwhile, nitrification enhancement example Such as Q learning algorithms are in the case where state set and behavior aggregate dimension are big, and its pace of learning can rapid decrease.These problems are restricted The application of artificial neural network and nitrification enhancement.
The content of the invention
Need to need in the training process first to carry out data it is an object of the invention to provide artificial neural network can be overcome Classification and Q learning algorithms in the case where state set and behavior aggregate dimension are big pace of learning can rapid decrease problem based on depth Spend the cognitive radio null tone two dimension meaning interference method against the enemy of intensified learning.
The present invention comprises the following steps：
1) action of cognitive radio time user is designated as x ∈ [0, N], wherein, x=0 represents that time user leaves the region, x ∈ [1, N] represents that time user's selection channel number is communicated for x channel, and N represents available channel quantity；
2) construction depth convolutional neural networks；
In step 2) in, the specific method of the construction depth convolutional neural networks can be：
(1) the conventional state action that the weight parameter θ of initialization depth convolutional neural networks, list entries are included is to number The number of operations B that W and network update；
(2) the Q values of depth convolutional neural networks are initialized, a corresponding output Q is distributed to the everything of secondary user Value；
(3) initialization discount factor γ, greedy factor ε.
In step 2) in, the construction depth convolutional neural networks include H layer of convolutional layer and full articulamentum M layers, H (H >= 1) in layer convolutional layer, the convolutional layer input size of first layer is 3 × B+2, with h1Individual wave filter；In M (M >=1), layer is connected entirely The full articulamentum output size of last layer and the optional actuating range of time user are in the same size in layer, i.e. N+1.
3) on the k moment, secondary user record a moment cognitive radio primary user access state (λk-1) and wireless communication Number signal interference ratio (SINRk-1), and constitute current state sk=[λk-1,SINRk-1]；
4) at the k moment, as k≤W, secondary user randomly selects an action xk∈[0,N]；As k ＞ W, construction depth The list entries of convolutional neural networksBy status switchDepth convolutional neural networks are input to, the Q of each action is calculated Value；Secondary user acts x according to greedy algorithm selectionk, the action with maximum Q values is chosen with 1- ε probability, it is random with ε probability Choose other actions；
In step 4) in, at kth moment, the input of depth convolutional neural networksIt is by current state and conventional record W state action to composition, i.e.,
5) x is worked askWhen=0, secondary user leaves the region；Work as xkDuring ∈ [1, N], secondary user's selection channel number is xkLetter Road is communicated；
6) the current access state (λ of secondary user's observation cognitive radio primary userk), work as λkWhen=1, cognition wireless is represented Electric primary user uses intended communication channel, conversely, λk=0；Signal interference ratio (the SINR of wireless signal is observed simultaneouslyk)；
In step 6) in, the signal interference ratio (SINR of the wireless signalk) it is the signal averaging measured in single call duration time Signal interference ratio.
7) benefit u is calculated according to formula (1)k：
uk=λkSINRk-Cmf(xk=0) (1)
Wherein, CmThe mobile loss in current region is left for secondary user；F (ξ) is an indicator function, when ξ is true its It is worth for 1, otherwise is 0；
8) secondary user obtains subsequent time state sk+1=[λk,SINRk], subsequent time list entries
9) k moment experiences are recordedInto experience pond D；
10) to the weight parameter θ of kth moment networkkCarry out B time update operate, in renewal process each time, at random from An experience is chosen in the D of experience pond
According to formula (2) counting loss value L (θk)：
Wherein, x ' is that list entries isUnder make the maximum action of Q values,Expression pairWith Machine variable does statistical expection calculating.
According to the Grad of formula (3) counting loss value
Using stochastic gradient descent method, the weight parameter of depth convolutional neural networks is updated using neutral net reverse feedback θkSo that GradMinimize；
11) according to environmental change, secondary user's repeat step 3)~10), until secondary user learning is selected to stable action Strategy.
Cognitive radio of the present invention time user is in the state of unknown disturbances machine attack mode and wireless channel environment, observation The access state of cognitive radio primary user, the signal interference ratio of wireless signal, decide whether to leave using deeply study mechanism Place is disturbed the one suitable frequency sending signal in region or selection.Learn with reference to depth convolutional neural networks and Q, utilize Q study is in the wireless optimal Anti-interference Strategy of dynamic game learning, by observer state and acquisition benefit input depth convolutional Neural Network accelerates pace of learning as training set.This method utilizes the mechanism that deeply learns, and improves cognitive radio dynamic The communication efficiency of hostility jammer is resisted under the wireless network environment scene of state change.
Embodiment
Technical scheme is further described with reference to embodiment.
A kind of cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply is comprised the following steps：
Step 1：A depth convolutional neural networks are constructed, 2 convolutional layers and 2 full articulamentums are included.First layer is volume Lamination, input size is 36, includes the convolution kernel of 20 3 × 3, stepping is 1, and output size is 20 × 4 × 4；The second layer is Convolutional layer, input size is 20 × 4 × 4, includes the convolution kernel of 40 2 × 2, stepping is 1, and output size is 40 × 3 × 3； Third layer is full articulamentum, and input size is 360, and output size is 180；Last layer is full articulamentum, and input size is 180, output size is 129.4 layers are all used as activation primitive using ReLU functions.
Step 2：The state action that the weight parameter θ of initialization depth convolutional neural networks, list entries are included is to number W =the 11 and number of operations B=16 of network renewal；The Q values of initialization time user's everything；Initialization discount factor γ= 0.5, greedy factor ε=0.1, available channel quantity N=128.
Step 3：Access state (the λ of a moment cognitive radio primary user on the k moment, secondary user recordk-1) and nothing Signal interference ratio (the SINR of line signalk-1), and constitute current state sk=[λk-1,SINRk-1]。
Step 4：At the k moment, as k≤W, secondary user randomly selects an action xk∈[0,128]；As k ＞ W, structure Make the list entries of depth convolutional NeuralAnd it is changed into 6 × 6 matrix.By state SequenceDepth convolutional neural networks are input to, the Q values of each action are calculated.Secondary user is chosen with 0.9 probability to be had most The action of big Q values, other actions are randomly selected with 0.1 probability.
Step 5：Work as xkWhen=0, secondary user leaves the region, works as xkDuring ∈ [1,128], secondary user's selection channel number is xkChannel communicated.
Step 6：Secondary user observes the access state (λ of present cognitive radio primary userk), the signal interference ratio of wireless signal (SINRk)。
Step 7：Benefit u is calculated according to formula (1)k：
uk=λkSINRk-Cmf(xk=0) (1)
Step 8：Secondary user obtains subsequent time state sk+1=[λk,SINRk], subsequent time list entries
Step 9：Record k moment experiencesInto experience pond D.
Step 10：Carry out 16 weight parameter θkUpdate operation.In renewal process each time, at random from the D of experience pond Choose an experience
According to formula (2) counting loss value L (θk)：
According to the Grad of formula (3) counting loss value
Using stochastic gradient descent method, the weight parameter of depth convolutional neural networks is updated using neutral net reverse feedback θkSo that GradMinimize.
Step 11：According to environmental change, secondary user's repeat step 3)~10), until secondary user learning to stable action Selection strategy.

Claims (5)

1. the cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply, it is characterised in that including following step Suddenly：
1) action of cognitive radio time user is designated as x ∈ [0, N], wherein, x=0 represents that time user leaves the region, x ∈ [1, N] represent that time user's selection channel number is communicated for x channel, N represents available channel quantity；
2) construction depth convolutional neural networks；
3) on the k moment, secondary user record a moment cognitive radio primary user access state (λk-1) and wireless signal Signal interference ratio (SINRk-1), and constitute current state sk=[λk-1,SINRk-1]；
4) at the k moment, as k≤W, secondary user randomly selects an action xk∈[0,N]；As k ＞ W, construction depth convolution god List entries through networkBy status switchDepth convolutional neural networks are input to, the Q values of each action are calculated；It is secondary User acts x according to greedy algorithm selectionk, the action with maximum Q values is chosen with 1- ε probability, it is randomly selected with ε probability He acts；
5) x is worked askWhen=0, secondary user leaves the region；Work as xkDuring ∈ [1, N], secondary user's selection channel number is xkChannel enter Row communication；
6) the current access state (λ of secondary user's observation cognitive radio primary userk), work as λkWhen=1, cognitive radio master is represented User uses intended communication channel, conversely, λk=0；Signal interference ratio (the SINR of wireless signal is observed simultaneouslyk)；
7) benefit u is calculated according to formula (1)k：
uk=λkSINRk-Cmf(xk=0) (1)
Wherein, CmThe mobile loss in current region is left for secondary user；F (ξ) is an indicator function, and when ξ is true, its value is 1, Otherwise for 0；
8) secondary user obtains subsequent time state sk+1=[λk,SINRk], subsequent time list entries
9) k moment experiences are recordedInto experience pond D；
10) to the weight parameter θ of kth moment networkkCarry out B times and update operation, in renewal process each time, at random from experience An experience is chosen in the D of pond
According to formula (2) counting loss value L (θk)：
Wherein, x ' is that list entries isUnder make the maximum action of Q values,Expression pairxd,ud,It is random to become Amount does statistical expection calculating；
According to the Grad of formula (3) counting loss value
Using stochastic gradient descent method, the weight parameter θ of depth convolutional neural networks is updated using neutral net reverse feedbackk, make Obtain GradMinimize；
11) according to environmental change, secondary user's repeat step 3)~10), until secondary user learning to stable action selection strategy.
2. the cognitive radio null tone two dimension meaning interference method against the enemy learnt as claimed in claim 1 based on deeply, it is special Levy is in step 2) in, the specific method of the construction depth convolutional neural networks is：
(1) the conventional state action that the initialization weight parameter θ of depth convolutional neural networks, list entries are included to number W with And the number of operations B that network updates；
(2) the Q values of depth convolutional neural networks are initialized, a corresponding output Q value is distributed to the everything of secondary user；
(3) initialization discount factor γ, greedy factor ε.
3. the cognitive radio null tone two dimension meaning interference method against the enemy learnt as claimed in claim 1 based on deeply, it is special Levy is in step 2) in, the construction depth convolutional neural networks include H layers and full articulamentum M layers of convolutional layer, in H layers of convolution In layer, the convolutional layer input size of first layer is 3 × B+2, with h1Individual wave filter；Last layer in M layers of full articulamentum Full articulamentum output size and the optional actuating range of time user are in the same size, i.e. N+1；Wherein, H >=1, M >=1.
4. the cognitive radio null tone two dimension meaning interference method against the enemy learnt as claimed in claim 1 based on deeply, it is special Levy is in step 4) in, at kth moment, the input of depth convolutional neural networksRecorded by current state and in the past W state action is to constituting, i.e.,
5. the cognitive radio null tone two dimension meaning interference method against the enemy learnt as claimed in claim 1 based on deeply, it is special Levy is in step 6) in, the signal interference ratio (SINR of the wireless signalk) it is the signal averaging letter measured in single call duration time Dry ratio.
CN201710181997.9A 2017-03-24 2017-03-24 The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply Pending CN106961684A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710181997.9A CN106961684A (en) 2017-03-24 2017-03-24 The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710181997.9A CN106961684A (en) 2017-03-24 2017-03-24 The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply

Publications (1)

Publication Number Publication Date
CN106961684A true CN106961684A (en) 2017-07-18

Family

ID=59470431

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710181997.9A Pending CN106961684A (en) 2017-03-24 2017-03-24 The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply

Country Status (1)

Country Link
CN (1) CN106961684A (en)

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107277065A (en) * 2017-08-11 2017-10-20 厦门大学 The resource regulating method of the senior constant threat of detection based on intensified learning
CN108563971A (en) * 2018-04-26 2018-09-21 广西大学 The more reader anti-collision algorithms of RFID based on depth Q networks
CN108599809A (en) * 2018-03-14 2018-09-28 中国信息通信研究院 Full duplex self-interference signal number removing method and device
CN109239661A (en) * 2018-09-18 2019-01-18 广西大学 A kind of RFID indoor locating system and algorithm based on depth Q network
CN109274456A (en) * 2018-09-10 2019-01-25 电子科技大学 A kind of imperfect information intelligence anti-interference method based on intensified learning
CN109861720A (en) * 2019-03-15 2019-06-07 中国科学院上海高等研究院 WSN anti-interference method, device, equipment and medium based on intensified learning
CN110190918A (en) * 2019-04-25 2019-08-30 广西大学 Cognition wireless sensor network frequency spectrum access method based on depth Q study
CN110213025A (en) * 2019-05-22 2019-09-06 浙江大学 Dedicated ad hoc network anti-interference method based on deeply study
CN110380776A (en) * 2019-08-22 2019-10-25 电子科技大学 A kind of Internet of things system method of data capture based on unmanned plane
CN110516566A (en) * 2019-08-08 2019-11-29 北京邮电大学 A kind of filtering method and device based on convolutional layer
CN110836518A (en) * 2019-11-12 2020-02-25 上海建科建筑节能技术股份有限公司 System basic knowledge based global optimization control method for self-learning air conditioning system
CN110856268A (en) * 2019-10-30 2020-02-28 西安交通大学 Dynamic multichannel access method for wireless network
CN112291495A (en) * 2020-10-16 2021-01-29 厦门大学 Wireless video low-delay anti-interference transmission method based on reinforcement learning
CN112422208A (en) * 2020-11-06 2021-02-26 西安交通大学 Signal detection method based on antagonistic learning under unknown channel model
CN112821971A (en) * 2021-02-01 2021-05-18 之江实验室 Time-varying channel signal detection method based on countermeasure learning
CN113227973A (en) * 2019-02-26 2021-08-06 谷歌有限责任公司 Reinforcement learning techniques for selecting software policy networks and autonomously controlling corresponding software clients based on the selected policy networks
CN113824469A (en) * 2021-08-05 2021-12-21 杭州电子科技大学 MFDRL-CTDE-based intelligent anti-interference decision method for frequency hopping networking
CN115276858A (en) * 2022-07-11 2022-11-01 中国人民解放军国防科技大学 Dynamic spectrum multi-domain anti-interference method and system based on cognitive anti-interference model

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103987018A (en) * 2014-06-10 2014-08-13 厦门大学 Cognitive radio network anti-hostile interference cooperative broadcasting method based on non-coordinated frequency hopping
CN104581738A (en) * 2015-01-30 2015-04-29 厦门大学 Cognitive radio hostile interference resisting method based on Q learning
CN104994569A (en) * 2015-06-25 2015-10-21 厦门大学 Multi-user reinforcement learning-based cognitive wireless network anti-hostile interference method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103987018A (en) * 2014-06-10 2014-08-13 厦门大学 Cognitive radio network anti-hostile interference cooperative broadcasting method based on non-coordinated frequency hopping
CN104581738A (en) * 2015-01-30 2015-04-29 厦门大学 Cognitive radio hostile interference resisting method based on Q learning
CN104994569A (en) * 2015-06-25 2015-10-21 厦门大学 Multi-user reinforcement learning-based cognitive wireless network anti-hostile interference method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
GUOAN HAN等: "Two-dimensional anti-jamming communication based on deep reinforcement learning", 《 2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)》 *

Cited By (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107277065A (en) * 2017-08-11 2017-10-20 厦门大学 The resource regulating method of the senior constant threat of detection based on intensified learning
CN107277065B (en) * 2017-08-11 2019-12-17 厦门大学 Resource scheduling method for detecting advanced persistent threat based on reinforcement learning
CN108599809A (en) * 2018-03-14 2018-09-28 中国信息通信研究院 Full duplex self-interference signal number removing method and device
CN108563971A (en) * 2018-04-26 2018-09-21 广西大学 The more reader anti-collision algorithms of RFID based on depth Q networks
CN109274456A (en) * 2018-09-10 2019-01-25 电子科技大学 A kind of imperfect information intelligence anti-interference method based on intensified learning
CN109239661A (en) * 2018-09-18 2019-01-18 广西大学 A kind of RFID indoor locating system and algorithm based on depth Q network
CN113227973A (en) * 2019-02-26 2021-08-06 谷歌有限责任公司 Reinforcement learning techniques for selecting software policy networks and autonomously controlling corresponding software clients based on the selected policy networks
CN109861720A (en) * 2019-03-15 2019-06-07 中国科学院上海高等研究院 WSN anti-interference method, device, equipment and medium based on intensified learning
CN110190918B (en) * 2019-04-25 2021-04-30 广西大学 Cognitive wireless sensor network spectrum access method based on deep Q learning
CN110190918A (en) * 2019-04-25 2019-08-30 广西大学 Cognition wireless sensor network frequency spectrum access method based on depth Q study
CN110213025A (en) * 2019-05-22 2019-09-06 浙江大学 Dedicated ad hoc network anti-interference method based on deeply study
CN110516566B (en) * 2019-08-08 2021-07-27 北京邮电大学 Filtering method and device based on convolutional layer
CN110516566A (en) * 2019-08-08 2019-11-29 北京邮电大学 A kind of filtering method and device based on convolutional layer
CN110380776B (en) * 2019-08-22 2021-05-14 电子科技大学 Internet of things system data collection method based on unmanned aerial vehicle
CN110380776A (en) * 2019-08-22 2019-10-25 电子科技大学 A kind of Internet of things system method of data capture based on unmanned plane
CN110856268A (en) * 2019-10-30 2020-02-28 西安交通大学 Dynamic multichannel access method for wireless network
CN110856268B (en) * 2019-10-30 2021-09-07 西安交通大学 Dynamic multichannel access method for wireless network
CN110836518A (en) * 2019-11-12 2020-02-25 上海建科建筑节能技术股份有限公司 System basic knowledge based global optimization control method for self-learning air conditioning system
CN112291495A (en) * 2020-10-16 2021-01-29 厦门大学 Wireless video low-delay anti-interference transmission method based on reinforcement learning
CN112422208A (en) * 2020-11-06 2021-02-26 西安交通大学 Signal detection method based on antagonistic learning under unknown channel model
CN112821971A (en) * 2021-02-01 2021-05-18 之江实验室 Time-varying channel signal detection method based on countermeasure learning
CN113824469A (en) * 2021-08-05 2021-12-21 杭州电子科技大学 MFDRL-CTDE-based intelligent anti-interference decision method for frequency hopping networking
CN115276858A (en) * 2022-07-11 2022-11-01 中国人民解放军国防科技大学 Dynamic spectrum multi-domain anti-interference method and system based on cognitive anti-interference model
CN115276858B (en) * 2022-07-11 2024-01-23 中国人民解放军国防科技大学 Dynamic spectrum multi-domain anti-interference method and system based on cognitive anti-interference model

Similar Documents

Publication Publication Date Title
CN106961684A (en) The cognitive radio null tone two dimension meaning interference method against the enemy learnt based on deeply
CN110620611B (en) Cooperative spectrum sensing method based on GEO and LEO double-layer satellite network
CN106358308A (en) Resource allocation method for reinforcement learning in ultra-dense network
CN108712748B (en) Cognitive radio anti-interference intelligent decision-making method based on reinforcement learning
Garnaev et al. A Bayesian jamming game in an OFDM wireless network
CN113382381B (en) Unmanned aerial vehicle cluster network intelligent frequency hopping method based on Bayesian Q learning
CN104994569B (en) Multi-user reinforcement learning-based method for resisting hostile interference of cognitive wireless network
Chen et al. Anti-jamming transmissions with learning in heterogenous cognitive radio networks
CN104467999B (en) Spectrum sensing algorithm based on quantum leapfrog
Yang et al. Partially observable multi-agent deep reinforcement learning for cognitive resource management
Han et al. Joint resource allocation in underwater acoustic communication networks: A game-based hierarchical adversarial multiplayer multiarmed bandit algorithm
CN116744311B (en) User group spectrum access method based on PER-DDQN
CN107371169A (en) Model selection based on evolutionary Game and frequency spectrum distribution mechanism in isomery full duplex D2D cellular networks
He et al. Dynamic channel assignment using ant colony optimization for cognitive radio networks
Li et al. Know Thy Enemy: An Opponent Modeling-Based Anti-Intelligent Jamming Strategy Beyond Equilibrium Solutions
CN105391490A (en) Satellite communication network selection algorithm based on cognition
CN114126021A (en) Green cognitive radio power distribution method based on deep reinforcement learning
CN105451268B (en) A kind of isomery cell accessing method of high energy efficiency
Sagduyu et al. Power control and rate adaptation as stochastic games for random access
Bojan et al. Genetic algorithm as energy optimization method in WSN
CN115378487B (en) Satellite communication dynamic spectrum access method based on electromagnetic barrier
Han et al. Joint Resource Allocation for Time-Varying Underwater Acoustic Communication System: A Self-Reflection Adversarial Bandit Approach
Altman et al. Evolutionary power control games in wireless networks
CN116318520A (en) Path control method, device and system of jammer and storage medium
Tan et al. Cell selection in two-tier femtocell networks using Q-learning algorithm

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

Application publication date: 20170718

RJ01 Rejection of invention patent application after publication