CN115580510A - Frequency hopping pattern generation method based on deep neural network and intelligent communication system - Google Patents
Frequency hopping pattern generation method based on deep neural network and intelligent communication system Download PDFInfo
- Publication number
- CN115580510A CN115580510A CN202211573644.0A CN202211573644A CN115580510A CN 115580510 A CN115580510 A CN 115580510A CN 202211573644 A CN202211573644 A CN 202211573644A CN 115580510 A CN115580510 A CN 115580510A
- Authority
- CN
- China
- Prior art keywords
- frequency hopping
- hopping pattern
- transmission channel
- pattern generation
- generation network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
-
- 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
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
- H04B1/69—Spread spectrum techniques
- H04B1/713—Spread spectrum techniques using frequency hopping
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/024—Channel estimation channel estimation algorithms
- H04L25/0254—Channel estimation channel estimation algorithms using neural network algorithms
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Power Engineering (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The application relates to a frequency hopping pattern generation method based on a deep neural network and an intelligent communication system. The method comprises the following steps: the method comprises the steps of inputting various known transmission channels and respectively corresponding optimal frequency hopping patterns into a frequency hopping pattern generation network to train the frequency hopping pattern generation network, obtaining the trained frequency hopping pattern generation network, and inputting channel transmission characteristic parameters of an actual application environment into the trained frequency hopping pattern generation network to generate the frequency hopping patterns suitable for the actual application environment when the network trained network is actually applied. By adopting the method, the confidentiality of the communication process can be enhanced, and the communication safety can be guaranteed.
Description
Technical Field
The application relates to the technical field of electronic countermeasure, in particular to a frequency hopping pattern generation method based on a deep neural network and an intelligent communication system.
Background
The development of opposition of both enemies and my parties around the electromagnetic space has become a trend, and how to maintain stable and reliable communication in a complex channel environment is a great challenge in military activities. The frequency hopping spread spectrum communication is a common mode of military communication and has the characteristics of strong concealment, interference resistance, good confidentiality and the like. In the face of a changeable and complex electromagnetic environment, related parameters of a communication link are often required to be manually adjusted, and a large amount of professional knowledge and experience are required, so that the actual communication requirements cannot be met; meanwhile, the communication process has the possibility of being intercepted, and the confidentiality of communication needs to be enhanced urgently.
With the rapid development of big data and hardware computing power, artificial intelligence, especially deep learning technology, makes great progress and is increasingly closely related to production practice. The deep learning has great promotion effect on the solution of the problems of difficult accurate description and complex modeling. Physical layer communication faces a complex electromagnetic environment, statistical modeling analysis needs to be carried out by means of a large amount of measurement data, and the existing theory does not necessarily completely cover various unknown communication scenes; the combination of the confidentiality and reliability of the physical layer communication with the deep learning technique is inevitable.
Therefore, it is necessary to develop a frequency hopping spread spectrum communication system having an artificial intelligence characteristic. And the prototype verification of the communication system is carried out by depending on the software technology, so that the research and development period can be greatly saved without carrying out complicated tests on special equipment.
Disclosure of Invention
Therefore, it is necessary to provide a frequency hopping pattern generation method based on a deep neural network and an intelligent communication system, which can enhance the security of the communication process and ensure the communication security, in order to solve the above technical problems.
A method for frequency hopping pattern generation based on a deep neural network, the method comprising:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns;
inputting a training sample pair into a frequency hopping pattern generation network to train the training sample pair, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
In one embodiment, the sample set of hopping patterns includes three types, a linear hopping pattern, a pseudo-random hopping pattern, and an irregular hopping pattern.
In one embodiment, the frequency hopping pattern is composed of a series of frequency hopping frequency offset sequences composed according to a specific rule.
In one embodiment, the determining, in the hopping pattern sample set, the optimal hopping patterns corresponding to the known transmission channels in the transmission channel set respectively includes:
sequentially adopting each frequency hopping pattern in the frequency hopping pattern sample set in each known transmission channel by using a frequency hopping communication system to carry out frequency hopping communication, and calculating a corresponding error rate;
and when the frequency hopping communication system communicates in each known transmission channel, the frequency hopping pattern corresponding to the lowest error rate is used as the optimal frequency hopping pattern of the known transmission channel.
In one embodiment, the frequency hopping pattern generation network employs a multi-layer deep learning neural network.
In one embodiment, the training the frequency hopping pattern generation network according to the reconstructed frequency hopping pattern and the loss function constructed by the optimal frequency hopping pattern in the training sample pair includes:
and optimizing the parameters of the frequency hopping pattern generation network according to the loss function until the loss function is converged, and obtaining the trained frequency hopping pattern generation network.
In one embodiment, when the transmission channel of the practical application environment of the frequency hopping communication system is an unknown transmission channel:
extracting unknown characteristic parameters of the unknown transmission channel;
inputting the unknown characteristic parameters into the trained frequency hopping pattern generation network, and outputting a reconstructed frequency hopping pattern;
performing frequency hopping communication by using the reconstructed frequency hopping pattern in the unknown transmission channel by using a frequency hopping communication system, and calculating a corresponding error rate;
and optimizing parameters of the trained frequency hopping pattern generation network according to the error rate, generating a new reconstructed frequency hopping pattern according to the unknown characteristic parameters by using the optimized frequency hopping pattern generation network, and obtaining a corresponding error rate in the unknown transmission channel by a frequency hopping communication system according to the new reconstructed frequency hopping pattern until the error rate is converged, thereby obtaining the optimal frequency hopping pattern of the unknown transmission channel.
In one embodiment, a frequency hopping knowledge base is constructed according to each known transmission channel characteristic parameter and the corresponding optimal frequency hopping pattern.
In one embodiment, the characteristic parameters of the unknown transmission channel and the corresponding optimal frequency hopping pattern are used to update the frequency hopping knowledge base.
An intelligent communications system, comprising: the device comprises a transmitting unit, a receiving unit, a transmission channel for communication between the transmitting unit and the receiving unit, and a frequency hopping pattern generating unit respectively connected with the transmitting unit and the receiving unit;
the frequency hopping pattern generating unit generates frequency hopping patterns according to the frequency hopping pattern generating method based on the deep neural network, and sends the frequency hopping patterns to the transmitting unit and the receiving unit respectively;
the transmitting unit comprises a transmitting baseband modulation part and a transmitting radio frequency processing part, the transmitting baseband modulation part performs digital modulation on a data message to form a constellation symbol, the constellation symbol is mapped according to the frequency hopping pattern transmitted by the frequency hopping pattern generating unit to obtain a modulation message, and the modulation message is subjected to up-conversion processing by the transmitting radio frequency processing part and then transmitted to the receiving unit through the transmission channel;
the receiving unit comprises a receiving baseband modulation part and a receiving radio frequency processing part, the receiving radio frequency processing part carries out down-conversion processing on the modulation message received through the transmission channel, the receiving baseband modulation part synchronizes the modulation message after down-conversion processing according to the frequency hopping pattern sent by the frequency hopping pattern generating unit, and then constellation demodulation is carried out to obtain the data message so as to finish the communication between the transmitting unit and the receiving unit.
An apparatus for frequency hopping pattern generation based on a deep neural network, the apparatus comprising:
a sample set obtaining module, configured to obtain a frequency hopping pattern sample set and a transmission channel sample set, where the frequency hopping pattern sample set includes multiple frequency hopping patterns, and the transmission channel sample set includes characteristic parameters of multiple known transmission channels;
a training sample pair construction module, configured to respectively determine, in the frequency hopping pattern sample set, an optimal frequency hopping pattern corresponding to each known transmission channel in the transmission channel set, and construct a training sample pair from the characteristic parameters of each known transmission channel and the corresponding optimal frequency hopping pattern;
the frequency hopping pattern generation network training module is used for inputting a training sample pair into a frequency hopping pattern generation network to train the training sample pair and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and the frequency hopping pattern generating module is used for acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generating network to generate the frequency hopping pattern suitable for the actual application environment.
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns:
inputting a training sample pair into a frequency hopping pattern generation network to train the training sample pair, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns;
inputting a training sample pair into a frequency hopping pattern generation network to train the training sample pair, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
According to the frequency hopping pattern generation method based on the deep neural network and the intelligent communication system, the deep neural network is trained by adopting various known transmission channel characteristic parameters and the corresponding optimal frequency hopping patterns to obtain the frequency hopping pattern generation network capable of obtaining the optimal frequency hopping patterns corresponding to the channels according to the channel characteristic parameters, and the frequency hopping pattern generation network is applied to the frequency hopping communication system.
Drawings
FIG. 1 is a schematic flow chart of a method for generating a frequency hopping pattern based on a deep neural network according to an embodiment;
FIG. 2 is an exemplary diagram of a linear frequency hopping pattern in one embodiment;
FIG. 3 is an exemplary diagram of a pseudo-random frequency hopping pattern in one embodiment;
FIG. 4 is an exemplary diagram of an irregular frequency hopping pattern in one embodiment;
FIG. 5 is a schematic diagram of an intelligent communications system in one embodiment;
FIG. 6 is a block diagram of an embodiment of a method for obtaining channel characterization parameters;
FIG. 7 is a block diagram of an apparatus for generating a hopping pattern based on a deep neural network according to an embodiment;
FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
As shown in fig. 1, a method for generating a frequency hopping pattern based on a deep neural network is provided, which includes the following steps:
step S100, obtaining a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of various known transmission channels;
step S110, respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns;
step S120, inputting a training sample pair into a frequency hopping pattern generation network for training, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in a training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
step S130, obtaining the channel transmission characteristic parameter of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameter into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
The method aims at the problems that in the prior art, in the face of changeable and complex electromagnetic environments, relevant parameters of a communication link often need to be adjusted manually, a large amount of professional knowledge and experience are needed, actual communication requirements cannot be met, and meanwhile, safety during communication needs to be enhanced due to the fact that message data are possibly intercepted during communication.
In step S100, various transmission channels in the sample set of transmission channels are transmitted to simulate various types of electromagnetic environments such as mountains, hills, plains, etc., wherein different types of electromagnetic environments are simulated through different characteristic parameters of the transmission channels.
In this embodiment, the frequency hopping pattern sample set includes three types, i.e., a linear frequency hopping pattern, a pseudo random frequency hopping pattern, and an irregular frequency hopping pattern, and each type of frequency hopping pattern includes a plurality of frequency hopping patterns.
Specifically, as shown in fig. 2, which is an example of a linear frequency hopping pattern, it can be seen from the figure that the linear frequency hopping pattern is a frequency hopping offset value that changes in an increasing or decreasing manner with time, and each frequency point occupies equal frequency hopping time.
Specifically, as shown in fig. 3, a pseudo-random frequency hopping pattern is shown, and it can be seen from the figure that the pseudo-random frequency hopping pattern is a sequence of changes of frequency hopping frequency offset, where the change sequence is controlled by a pseudo-random sequence, and frequency points occupy equal frequency hopping time.
Specifically, as shown in fig. 4, the irregular frequency hopping pattern is a frequency hopping frequency offset sequence that changes completely randomly, and the frequency hopping time occupied by each frequency point may not be equal.
In fig. 2-4, the frequency hopping time refers to the frequency dwell time, and the digital frequency hopping operation ensures that the frequency hopping switching time is negligible.
In this embodiment, the characteristic parameters of the transmission channel include multipath delay, multipath average power, RMS delay spread, and coherence bandwidth.
In step S110, determining the optimal hopping pattern corresponding to each known transmission channel in the transmission channel set according to the hopping pattern sample set includes: and when the frequency hopping communication system communicates in each known transmission channel, the frequency hopping pattern corresponding to the lowest error rate is used as the optimal frequency hopping pattern of the known transmission channel.
Specifically, the method comprises the following steps:
S1.1A Transmission channel in a sample set of Transmission channels is first addressedThe transmitting unit in the frequency hopping communication system selects a certain sample in a first type of frequency hopping pattern in a frequency hopping pattern sample set to modulate original message data, the receiving unit also selects the frequency hopping pattern to synchronize and demodulate to obtain demodulated message data, and the error rate is calculated according to the original message data and the demodulated message data;
s1.2, sequentially selecting other samples in the first type of frequency hopping pattern, and repeating the step S1.1 to obtain a series of error code rate values;
s1.3, the transmitting unit selects a certain sample in the second type of frequency hopping pattern to modulate, repeats the step S1.2, and records a series of error code rate values;
s1.4, the transmitting unit selects a certain sample in the third type of frequency hopping patterns for modulation, repeats the step S1.2 and records a series of error code rate values;
s1.5, comparing the bit error rate, and acquiring a frequency hopping pattern corresponding to the minimum value as a channelOptimal frequency hopping pattern of;
S2 aiming at another transmission channel in the transmission channel sample setRepeating steps S1.1-S1.5 to determine the channelOptimal frequency hopping pattern of;
S3 for another transport channel in a sample set of transport channelsRepeating the steps S1.1-S1.5 to determine the channelLocal optimum frequency hopping pattern;
Then, the known channels are collected for the transmission channel sample in turnRespectively determining locally optimal hopping patterns。
Finally, the characteristic parameters of various known transmission channels and the corresponding optimal frequency hopping patterns are used for constructing training sample pairs。
In step S120, the frequency hopping pattern generation network employs a multilayer deep learning neural network.
In the present embodiment, in training the frequency hopping pattern generation networkThe set of hopping sequences is also input into the hopping pattern generation network along with the training sample pairs to train them. Wherein the frequency hopping sequence set is a series of frequency hopping frequency offset value sets relative to the central carrier. In addition, when training the network, the frequency hopping offset value in the frequency hopping sequence set may be fixed, and may also be adjusted according to the actual environment.
In this embodiment, training the hopping pattern generation network according to the reconstructed hopping pattern and the loss function constructed by the optimal hopping pattern in the training sample pair includes: and optimizing the parameters of the frequency hopping pattern generation network according to the loss function until the loss function is converged, thereby obtaining the trained frequency hopping pattern generation network.
In this embodiment, digital frequency hopping is adopted in the frequency hopping communication system, so that frequency hopping modulation can be completed in the baseband only by knowing the frequency offset value of each frequency hopping point relative to the central carrier, and therefore, the frequency hopping pattern can be simplified into a series of frequency hopping offset sequences composed according to a specific rule. And in fact the data output by the hopping pattern generation network is not image data but a series of hopping frequency offset sequences composed according to a specific rule.
In step S130, when the trained frequency hopping pattern generation network is applied, a corresponding channel transmission characteristic parameter is generated according to an actual transmission environment, and the characteristic parameter is input into the trained frequency hopping pattern generation network to output a frequency hopping pattern adapted to the environment.
In practical application, it is likely that the transmission channel of the practical application environment is an unknown transmission channel, that is, the transmission channel type corresponding to the practical application environment is not in the transmission channel sample set, and at this time: extracting unknown characteristic parameters of an unknown transmission channel, inputting the unknown characteristic parameters into a trained frequency hopping pattern generation network, outputting a reconstructed frequency hopping pattern, performing frequency hopping communication in the unknown transmission channel by using a frequency hopping communication system by using the reconstructed frequency hopping pattern, calculating a corresponding error rate, optimizing parameters of the trained frequency hopping pattern generation network according to the calculated error rate, generating a new reconstructed frequency hopping pattern according to the unknown characteristic parameters by using the optimized frequency hopping pattern generation network, and obtaining a corresponding error rate according to the new reconstructed frequency hopping pattern in the unknown transmission channel by using the frequency hopping communication system until the error rate is converged, thereby obtaining the optimal frequency hopping pattern of the unknown transmission channel.
Specifically, for an unknown transmission channel, a pseudorandom sequence is used as a sending sequence, and when the unknown transmission channel reaches a receiving unit through a channel after modulation, sliding correlation is performed with an original sending sequence, so that characteristic parameters such as a delay power spectrum, mean-square delay expansion and the like corresponding to the unknown application environment channel can be estimated. Similar iterative training is performed for more unknown signal transmission environments.
In this embodiment, a frequency hopping knowledge base is constructed according to each known transmission channel characteristic parameter and the corresponding optimal frequency hopping pattern.
And further, updating the frequency hopping knowledge base by using the characteristic parameters of the unknown transmission channel and the corresponding optimal frequency hopping pattern.
Specifically, the mapping set obtained in the generalization process of the neural network training is usedAnd accumulated as a knowledge base. In the face of complex and changeable channel environments, firstly, extracting characteristic parameters of the channel environments, retrieving and matching a knowledge base, and then, supplementing a new channel characteristic model and an optimal frequency hopping pattern to the knowledge base to complete updating. So that the situation of similar channels is met at a later period and rapid decision is made.
As shown in fig. 6, in this embodiment, an intelligent communication system is further provided, including: the device comprises a transmitting unit, a receiving unit, a transmission channel for communication between the transmitting unit and the receiving unit, and a frequency hopping pattern generating unit respectively connected with the transmitting unit and the receiving unit.
The frequency hopping pattern generating unit generates frequency hopping patterns according to the frequency hopping pattern generating method based on the deep neural network, and sends the frequency hopping patterns to the transmitting unit and the receiving unit respectively;
the transmitting unit comprises a transmitting baseband modulation part and a transmitting radio frequency processing part, the transmitting baseband modulation part digitally modulates the data message to form a constellation symbol, the constellation symbol is mapped according to the frequency hopping pattern sent by the frequency hopping pattern generating unit to obtain a modulation message, and the modulation message is subjected to up-conversion processing by the transmitting radio frequency processing part and then is sent to the receiving unit through a transmission channel;
the receiving unit comprises a receiving baseband modulation part and a receiving radio frequency processing part, the receiving radio frequency processing part carries out down-conversion processing on the modulation message received through the transmission channel, the receiving baseband modulation part synchronizes the modulation message after down-conversion processing according to the frequency hopping pattern sent by the frequency hopping pattern generating unit, and then constellation demodulation is carried out to obtain a data message so as to finish communication between the transmitting unit and the receiving unit.
In this embodiment, the baseband modulation and demodulation section is implemented by software programming according to the radio platform to process baseband IQ data, the radio frequency section is implemented by an FPGA, and the hopping pattern generation unit is implemented by a dedicated GPU.
In the frequency hopping pattern generation method based on the deep neural network, the method constructs a frequency hopping pattern suitable for channel transmission. Here, for a given channel, modeling is performed, and channel characteristic parameters are extracted as input for multi-layer neural network training to obtain an optimal frequency hopping pattern. The method solves the problem of professional and complex parameter adjustment of the physical layer of the communication system by utilizing a deep learning technology, enhances the confidentiality of the communication process and guarantees the communication safety by combining a basic theory and the frequency hopping design of the physical layer of artificial intelligence, and accelerates the verification development of an intelligent frequency hopping communication prototype platform by adopting a software reconfigurable technology and relying on a general architecture and a general platform.
It should be understood that, although the steps in the flowchart of fig. 1 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the sub-steps or stages of other steps.
In one embodiment, as shown in fig. 7, there is provided a frequency hopping pattern generating apparatus based on a deep neural network, including: a sample set obtaining module 200, a training sample pair constructing module 210, a frequency hopping pattern generation network training module 220, and a frequency hopping pattern generating module 230, wherein:
a sample set obtaining module 200, configured to obtain a frequency hopping pattern sample set and a transmission channel sample set, where the frequency hopping pattern sample set includes multiple frequency hopping patterns, and the transmission channel sample set includes characteristic parameters of multiple known transmission channels;
a training sample pair constructing module 210, configured to respectively determine, in the hopping pattern sample set, an optimal hopping pattern corresponding to each known transmission channel in the transmission channel set according to the known transmission channel, and construct a training sample pair from the characteristic parameters of each known transmission channel and the corresponding optimal hopping pattern;
a frequency hopping pattern generation network training module 220, configured to train the training sample pair to an input frequency hopping pattern generation network, and obtain a trained frequency hopping pattern generation network, where when the frequency hopping pattern generation network is trained, a characteristic parameter in the training sample pair is input into the frequency hopping pattern generation network, and a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by an optimal frequency hopping pattern in the training sample pair;
a frequency hopping pattern generating module 230, configured to obtain channel transmission characteristic parameters of an actual application environment of the frequency hopping communication system, and input the channel transmission characteristic parameters into the trained frequency hopping pattern generating network to generate a frequency hopping pattern suitable for the actual application environment.
For specific limitations of the deep neural network-based frequency hopping pattern generation apparatus, reference may be made to the above limitations on the deep neural network-based frequency hopping pattern generation method, and details are not repeated here. The various modules in the deep neural network-based frequency hopping pattern generating device can be wholly or partially implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent of a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method for frequency hopping pattern generation based on a deep neural network. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the configuration shown in fig. 8 is a block diagram of only a portion of the configuration associated with the present application, and is not intended to limit the computing device to which the present application may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining the corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns:
inputting a training sample to a frequency hopping pattern generation network to train the training sample, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input to the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns:
inputting a training sample pair into a frequency hopping pattern generation network to train the training sample pair, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct Rambus Dynamic RAM (DRDRAM), and Rambus Dynamic RAM (RDRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent application shall be subject to the appended claims.
Claims (10)
1. A frequency hopping pattern generation method based on a deep neural network is characterized by comprising the following steps:
acquiring a frequency hopping pattern sample set and a transmission channel sample set, wherein the frequency hopping pattern sample set comprises a plurality of frequency hopping patterns, and the transmission channel sample set comprises characteristic parameters of a plurality of known transmission channels;
respectively determining corresponding optimal frequency hopping patterns in the frequency hopping pattern sample set according to various known transmission channels in the transmission channel set, and constructing training sample pairs by using the characteristic parameters of the various known transmission channels and the corresponding optimal frequency hopping patterns;
inputting the training samples into a frequency hopping pattern generation network to train the training samples, and obtaining a trained frequency hopping pattern generation network, wherein when the frequency hopping pattern generation network is trained, characteristic parameters in the training sample pair are input into the frequency hopping pattern generation network, a reconstructed frequency hopping pattern is output, and the frequency hopping pattern generation network is trained according to the reconstructed frequency hopping pattern and a loss function constructed by the optimal frequency hopping pattern in the training sample pair;
and acquiring channel transmission characteristic parameters of the actual application environment of the frequency hopping communication system, and inputting the channel transmission characteristic parameters into the trained frequency hopping pattern generation network to generate the frequency hopping pattern suitable for the actual application environment.
2. The method of claim 1, wherein the set of hopping pattern samples comprises three types of linear hopping patterns, pseudo-random hopping patterns, and irregular hopping patterns.
3. The method of claim 1, wherein the hopping pattern is comprised of a sequence of hopping frequency offsets organized according to a specific rule.
4. The method of claim 1, wherein determining the optimal hopping pattern in the hopping pattern sample set according to the known transmission channels in the transmission channel set comprises:
sequentially adopting each frequency hopping pattern in the frequency hopping pattern sample set to carry out frequency hopping communication in each known transmission channel by using a frequency hopping communication system, and calculating a corresponding error rate;
and when the frequency hopping communication system communicates in each known transmission channel, the frequency hopping pattern corresponding to the lowest error rate is used as the optimal frequency hopping pattern of the known transmission channel.
5. The frequency hopping pattern generation method of claim 1, wherein the frequency hopping pattern generation network employs a multi-layer deep learning neural network.
6. The method of claim 1, wherein training the hopping pattern generation network according to the reconstructed hopping pattern and a loss function constructed from the optimal hopping pattern in the training sample pair comprises:
and optimizing the parameters of the frequency hopping pattern generation network according to the loss function until the loss function is converged, thereby obtaining the trained frequency hopping pattern generation network.
7. The frequency hopping pattern generating method as claimed in any one of claims 1 to 6, wherein when the transmission channel of the actual application environment of the frequency hopping communication system is an unknown transmission channel:
extracting unknown characteristic parameters of the unknown transmission channel;
inputting the unknown characteristic parameters into the trained frequency hopping pattern generation network, and outputting a reconstructed frequency hopping pattern;
performing frequency hopping communication by using the reconstructed frequency hopping pattern in the unknown transmission channel by using a frequency hopping communication system, and calculating a corresponding error rate;
and optimizing the parameters of the trained frequency hopping pattern generation network according to the error rate, generating a new reconstructed frequency hopping pattern by using the optimized frequency hopping pattern generation network according to the unknown characteristic parameters, and obtaining a corresponding error rate in the unknown transmission channel by a frequency hopping communication system according to the new reconstructed frequency hopping pattern until the error rate is converged, thereby obtaining the optimal frequency hopping pattern of the unknown transmission channel.
8. The method of claim 7, wherein a frequency hopping knowledge base is constructed according to each known transmission channel characteristic parameter and the corresponding optimal frequency hopping pattern.
9. The method of claim 8, wherein the knowledge base of frequency hopping is updated with the parameters of the unknown transmission channel and the corresponding optimal frequency hopping pattern.
10. An intelligent communication system, comprising: the device comprises a transmitting unit, a receiving unit, a transmission channel for communication between the transmitting unit and the receiving unit, and a frequency hopping pattern generating unit respectively connected with the transmitting unit and the receiving unit;
the frequency hopping pattern generating unit generates a frequency hopping pattern according to the frequency hopping pattern generating method based on the deep neural network of any one of claims 7 to 9, and transmits the frequency hopping pattern to the transmitting unit and the receiving unit, respectively;
the transmitting unit comprises a transmitting baseband modulation part and a transmitting radio frequency processing part, the transmitting baseband modulation part performs digital modulation on a data message to form a constellation symbol, the constellation symbol is mapped according to the frequency hopping pattern transmitted by the frequency hopping pattern generating unit to obtain a modulation message, and the modulation message is subjected to up-conversion processing by the transmitting radio frequency processing part and then transmitted to the receiving unit through the transmission channel;
the receiving unit comprises a receiving baseband modulation part and a receiving radio frequency processing part, the receiving radio frequency processing part carries out down-conversion processing on the modulation message received through the transmission channel, the receiving baseband modulation part synchronizes the modulation message after down-conversion processing according to the frequency hopping pattern sent by the frequency hopping pattern generating unit, and then constellation demodulation is carried out to obtain the data message so as to finish the communication between the transmitting unit and the receiving unit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211573644.0A CN115580510B (en) | 2022-12-08 | 2022-12-08 | Frequency hopping pattern generation method based on deep neural network and intelligent communication system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211573644.0A CN115580510B (en) | 2022-12-08 | 2022-12-08 | Frequency hopping pattern generation method based on deep neural network and intelligent communication system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115580510A true CN115580510A (en) | 2023-01-06 |
CN115580510B CN115580510B (en) | 2023-04-07 |
Family
ID=84590815
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211573644.0A Active CN115580510B (en) | 2022-12-08 | 2022-12-08 | Frequency hopping pattern generation method based on deep neural network and intelligent communication system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115580510B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102651655A (en) * | 2011-02-24 | 2012-08-29 | 北京化工大学 | Realization method of fast frequency hopping communication |
CN103944606A (en) * | 2014-02-28 | 2014-07-23 | 电子科技大学 | Self-adaptation frequency hopping pattern generation method |
US20160065249A1 (en) * | 2014-09-02 | 2016-03-03 | Texas Instruments Incorporated | Capture selection for digital pre-distortion adaptation and capture concatenation for frequency hopping pre-distortion adaptation |
CN107682043A (en) * | 2017-10-30 | 2018-02-09 | 四川九洲电器集团有限责任公司 | A kind of generation method of frequency hopping pattern |
CN113098565A (en) * | 2021-04-02 | 2021-07-09 | 甘肃工大舞台技术工程有限公司 | Stage carrier communication self-adaptive frequency hopping anti-interference technology based on deep network |
CN113784359A (en) * | 2021-09-08 | 2021-12-10 | 昆明理工大学 | Dynamic channel access method based on improved BP neural network algorithm |
CN115118306A (en) * | 2022-06-27 | 2022-09-27 | 深圳友讯达科技股份有限公司 | Power Internet of things frequency hopping sequence pattern generation method and power Internet of things equipment |
-
2022
- 2022-12-08 CN CN202211573644.0A patent/CN115580510B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102651655A (en) * | 2011-02-24 | 2012-08-29 | 北京化工大学 | Realization method of fast frequency hopping communication |
CN103944606A (en) * | 2014-02-28 | 2014-07-23 | 电子科技大学 | Self-adaptation frequency hopping pattern generation method |
US20160065249A1 (en) * | 2014-09-02 | 2016-03-03 | Texas Instruments Incorporated | Capture selection for digital pre-distortion adaptation and capture concatenation for frequency hopping pre-distortion adaptation |
CN107682043A (en) * | 2017-10-30 | 2018-02-09 | 四川九洲电器集团有限责任公司 | A kind of generation method of frequency hopping pattern |
CN113098565A (en) * | 2021-04-02 | 2021-07-09 | 甘肃工大舞台技术工程有限公司 | Stage carrier communication self-adaptive frequency hopping anti-interference technology based on deep network |
CN113784359A (en) * | 2021-09-08 | 2021-12-10 | 昆明理工大学 | Dynamic channel access method based on improved BP neural network algorithm |
CN115118306A (en) * | 2022-06-27 | 2022-09-27 | 深圳友讯达科技股份有限公司 | Power Internet of things frequency hopping sequence pattern generation method and power Internet of things equipment |
Non-Patent Citations (4)
Title |
---|
GUOAN HAN: "Two-dimensional anti-jamming communication based on deep reinforcement learning", 《2017 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)》 * |
何雨桐: "基于GRU神经网络的自适应跳频技术研究", 《基于GRU神经网络的自适应跳频技术研究》 * |
刘尧等: "一种基于向量回归的无人机通信信道选择方法", 《软件导刊》 * |
申鹏: "跳频通信系统中的干扰识别技术研究", 《中国优秀硕士学位论文全文数据库(电子期刊)》 * |
Also Published As
Publication number | Publication date |
---|---|
CN115580510B (en) | 2023-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Kim et al. | Blockchained on-device federated learning | |
Ding et al. | Spectrum inference in cognitive radio networks: Algorithms and applications | |
Abel et al. | Chaos communications-principles, schemes, and system analysis | |
Kumar | Approximate Outage Probability and Capacity for $\kappa $-$\mu $ Shadowed Fading | |
Haenggi et al. | Diversity polynomials for the analysis of temporal correlations in wireless networks | |
CN109325584A (en) | Federation's modeling method, equipment and readable storage medium storing program for executing neural network based | |
Quan et al. | Optimal linear fusion for distributed detection via semidefinite programming | |
CN110956262A (en) | Hyper network training method and device, electronic equipment and storage medium | |
Qin et al. | Malicious user detection based on low-rank matrix completion in wideband spectrum sensing | |
CN107231214A (en) | Optimum detectors method based on evolution chaos quantum neutral net | |
CN108256334A (en) | Loophole test method, device, computer equipment and storage medium | |
Zhang et al. | Novel Bayesian inference algorithms for multiuser detection in M2M communications | |
Cui et al. | Side-information aided compressed multi-user detection for up-link grant-free NOMA | |
CN115580510B (en) | Frequency hopping pattern generation method based on deep neural network and intelligent communication system | |
CN108683465A (en) | Signaling testing method, device, computer equipment and storage medium | |
CN112988606B (en) | Test case generation method, device, equipment and storage medium | |
Dakic et al. | HybNet: A hybrid deep learning-matched filter approach for IoT signal detection | |
CN108345453A (en) | Code generating method, code generator and readable storage medium storing program for executing | |
KR20010005860A (en) | A code synchronization unit and method | |
Srinivasan et al. | Smart Channel State Information Pre-Processing for Authentication and Symmetric Key Distillation | |
US20230344685A1 (en) | A communication unit for soft-decision demodulation and method therefor | |
Zhu et al. | Asynchronous massive connectivity with deep-learned approximate message passing | |
Wang et al. | Spectrum breathing: Protecting over-the-air federated learning against interference | |
Glinskiy et al. | ALPACA: An Asymmetric Loss Prediction Algorithm for Channel Adaptation Based on a Convolutional-Recurrent Neural Network in URLLC Systems | |
Hu et al. | Signal Classification in Real-time Based on SDR using Convolutional Neural 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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |