CN112383496A - Mine communication method, system, computer equipment and medium based on depth receiver - Google Patents

Mine communication method, system, computer equipment and medium based on depth receiver Download PDF

Info

Publication number
CN112383496A
CN112383496A CN202011241415.XA CN202011241415A CN112383496A CN 112383496 A CN112383496 A CN 112383496A CN 202011241415 A CN202011241415 A CN 202011241415A CN 112383496 A CN112383496 A CN 112383496A
Authority
CN
China
Prior art keywords
mine
mimo
ofdm
receiver
channel
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
CN202011241415.XA
Other languages
Chinese (zh)
Inventor
李旭虹
周孝铭
王煜仪
高钰凯
张衡
陈龙
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.)
Xian University of Science and Technology
Original Assignee
Xian University of Science and Technology
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 Xian University of Science and Technology filed Critical Xian University of Science and Technology
Priority to CN202011241415.XA priority Critical patent/CN112383496A/en
Publication of CN112383496A publication Critical patent/CN112383496A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/26—Systems using multi-frequency codes
    • H04L27/2601—Multicarrier modulation systems
    • H04L27/2626—Arrangements specific to the transmitter only
    • H04L27/2627—Modulators
    • H04L27/2628—Inverse Fourier transform modulators, e.g. inverse fast Fourier transform [IFFT] or inverse discrete Fourier transform [IDFT] modulators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0045—Arrangements at the receiver end
    • H04L1/0054—Maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0056—Systems characterized by the type of code used
    • H04L1/0057—Block codes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0056—Systems characterized by the type of code used
    • H04L1/0071—Use of interleaving
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/0014—Carrier regulation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/26—Systems using multi-frequency codes
    • H04L27/2601—Multicarrier modulation systems
    • H04L27/2647—Arrangements specific to the receiver only
    • H04L27/2649—Demodulators
    • H04L27/265—Fourier transform demodulators, e.g. fast Fourier transform [FFT] or discrete Fourier transform [DFT] demodulators
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/0014—Carrier regulation
    • H04L2027/0024—Carrier regulation at the receiver end
    • H04L2027/0026—Correction of carrier offset

Abstract

The invention belongs to the technical field of mine communication, and discloses a mine communication method, a mine communication system, computer equipment and a medium based on a depth receiver, which are used for constructing an MIMO-OFDM system model of a mine tunnel; a deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system is obtained by establishing a deep neural network and carrying out effective training; and actually testing the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system until the original transmitting signal is recovered. The invention adopts the mine wireless communication system based on the MIMO-OFDM technology, improves the capacity of the system, reduces the error rate of the system and can effectively resist the multipath fading of the channel. The invention can reduce the design complexity of the mine communication system and effectively realize the data transmission under the mine.

Description

Mine communication method, system, computer equipment and medium based on depth receiver
Technical Field
The invention belongs to the technical field of mine communication, and particularly relates to a mine communication method and system based on a depth receiver, computer equipment and a medium.
Background
At present, a mine tunnel is a non-free propagation space with limited space, the propagation characteristics of electromagnetic waves are complex and variable, the transmission phase, frequency and amplitude characteristics of useful signals are seriously damaged by the multipath fading phenomenon generated by the propagation characteristics, the reliability of a mine wireless communication system is sharply reduced as a result of intersymbol interference (ISI), and the reliable transmission distance of the electromagnetic waves in the mine tunnel is only hundreds of meters. The special transmission environment and safe production in the underground coal mine require a mine wireless communication system to provide high data rate and high reliability, but the frequency spectrum resource shortage and the fading and interference caused by the special transmission environment of a roadway require measures to be taken to improve the frequency spectrum utilization rate and the transmission reliability.
A Multiple Input Multiple Output (MIMO) technique performs wireless transmission using multiple transmitting antennas and multiple receiving antennas, and can improve communication quality and rate of a wireless communication system through multiplexing and diversity techniques of multiple spatial channels. MIMO technology is used for quasi-flat fading channels for narrowband (signal bandwidth is smaller than channel coherence bandwidth) wireless communication, and shows a deficiency for frequency selective fading channels for wideband wireless communication. Orthogonal Frequency Division Multiplexing (OFDM) is a high-efficiency parallel multi-carrier transmission technology, high-speed serial data streams transmitted are distributed to a plurality of orthogonal sub-channels for parallel transmission through serial-to-parallel conversion, so that the symbol width of each sub-carrier channel is larger than the channel delay spread, each sub-carrier channel is a quasi-flat fading channel, and then cyclic spread is added to ensure that a communication system is not influenced by ISI (inter-symbol interference) caused by multipath interference, and the multi-path fading is effectively resisted.
In the field of communications, the increase in data volume, high-speed information transmission, and high-accuracy communication demand have posed major challenges to present-day communication systems. In addition, the existing communication system has limitations in the use of mass data processing and system configuration information, and based on these demands, a new communication system theory needs to be established. Artificial intelligence technology is widely applied to various industries and has achieved great success in various industries, and has great advantages in data processing and structural information utilization, so that artificial intelligence algorithms such as deep learning and the like slowly begin to replace traditional communication algorithms. At present, a large number of algorithms are used for optimizing a single module of a receiver for channel estimation, channel equalization and the like, and although the performance of the module can be improved by a modularization-based optimization design method, the optimal performance of the single module does not mean the optimal overall performance of a system. The neural network model is designed based on overall optimization, and after a complete neural network is used for realizing the receiver, the network model can be optimized through an overall optimization algorithm of the neural network, so that the overall performance of the system is further improved.
Through the above analysis, the problems and defects of the prior art are as follows:
in order to resist the influence of multipath fading in mine wireless communication, researchers have started from the aspect of coding modulation in recent years, and apply more excellent wireless technologies on the ground, such as UWB, SISO, MIMO, OFDM, MIMO-OFDM and the like, to mine communication to resist the multipath fading of the mine communication.
The ultra-wideband (UWB) technology has the characteristics of low power consumption, high data rate, frequency spectrum sharing, lower system complexity and stronger multipath resistance, but because the UWB transmitting signal frequency spectrum can be from 3.1Hz to 10.6GHz, the span is very large, the interference can be generated to other communication systems, and meanwhile, the UWB is only suitable for short-distance transmission, so that the long-distance communication of a mine tunnel cannot be met.
The channel capacity of the traditional communication system with Single Input and Single Output (SISO), namely a transmitting end and a receiving end, of which only one antenna is provided is limited, and the requirement of mine communication on the channel capacity cannot be met.
The Multiple Input Multiple Output (MIMO) technology can improve the anti-fading performance of the system and improve the capacity of the channel, but cannot resist frequency selective fading.
Orthogonal Frequency Division Multiplexing (OFDM) technology, carry on the high-speed data flow/parallel conversion to the low-speed data flow, and adopt and insert the method of the cyclic prefix, dispel the influence of ISI, its every subcarrier frequency spectrum is orthogonal to overlap each other at the same time, have higher frequency spectrum utilization rate, especially adaptive OFDM system has stronger anti-multipath ability, therefore OFDM can be used in the mine wireless communication, in order to overcome the serious multipath fading problem in the mine.
The MIMO-OFDM technology integrates the advantages of the MIMO technology and the OFDM technology, and the MIMO system can fully utilize multipath components, eliminate multipath effect and increase channel capacity. The OFDM technology divides a channel into a plurality of orthogonal sub-channels, so that fading on each sub-channel can be regarded as flat fading, and the influence caused by frequency selective fading is solved. The combination of the two is more suitable for the special environment of the mine laneway.
The difficulty in solving the above problems and defects is:
different from the ground wireless communication environment, the space in a mine tunnel is limited, the tunnel wall is rough, large-scale mining equipment is numerous, and the large-scale mining equipment contains a large amount of substances such as coal dust, water vapor and the like, so that the propagation environment is very complex. The multipath effect of the channel in the mine is severe, which causes severe fading of the signal during transmission in the mine, and therefore, the wireless communication system in the mine must have the function of overcoming the multipath fading. In addition, the influence of frequency selective fading on mine wireless communication cannot be ignored. Therefore, the existing MIMO system model on the ground cannot be directly applied to the mine wireless communication system. In recent years, a large number of expert scholars have devoted themselves to the research of the mine MIMO channel model and have obtained remarkable research results. However, most modeling methods are based on the propagation theory of electromagnetic waves in a complex mine, and a mine MIMO related channel model is established by calculating the space-time correlation of each propagation path in a multipath channel, and the modeling process is high in complexity and large in calculation amount. In addition, the deterministic modeling is a model established aiming at a specific environment, has no generality, and is more suitable for theoretical research and prediction of a mine channel model.
The significance of solving the problems and the defects is as follows:
in order to effectively solve the problems of serious multipath fading, poor anti-interference capability and the like in a mine, the MIMO-OFDM technology is applied to a mine wireless communication system, combines the advantages of the MIMO technology and the OFDM technology, can effectively resist the multipath fading and simultaneously utilize the multipath effect to realize the diversity and multiplexing of multiple antennas, not only increases the channel capacity and improves the frequency spectrum utilization rate, but also improves the transmission efficiency under the condition of not increasing the transmission power, and meets the requirement that the transmission power of communication equipment in the mine cannot be overlarge. Therefore, the research on the application of the MIMO-OFDM technology in the mine wireless communication system has very important significance.
The design scheme provided by the invention is an intelligent receiving system matched with a traditional communication transmitter. The entire information recovery process from the received signal samples to the information bit stream is achieved using a deep neural network. It does not optimize the local performance of one or more modules in the receiver, but rather the global performance of information recovery.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a mine communication method, a mine communication system, computer equipment and a medium based on a depth receiver.
The invention is realized in such a way that a mine communication method based on a depth receiver comprises the following steps:
step one, constructing an MIMO-OFDM system model of a mine roadway;
step two, a deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system is obtained by establishing a deep neural network and carrying out effective training;
and step three, actually testing the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system until the original transmitting signal is recovered.
Further, in the step one, the MIMO-OFDM system model of the mine roadway is as follows:
the MIMO-OFDM system model input-output relation is expressed as:
Figure BDA0002768486420000041
wherein, ykRepresents a k-th sub-channel received signal vector; h denotes the impulse response of each subchannel;
Figure BDA0002768486420000051
which transmitted signal vector is the k sub-carrier of the ith antenna; n iskRepresenting white gaussian noise on the kth sub-channel;
the impulse response of each subchannel is as follows:
Figure BDA0002768486420000052
further, in the second step, the obtaining of the effectively trained deep receiver model of the mine MIMO-OFDM communication receiving system by establishing the deep neural network and performing effective training includes:
(1) the transmitting terminal generates known random binary bit stream, and the known random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
(2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
(3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure BDA0002768486420000054
Wherein k is a carrier serial number;
(4) building a deep neural network; and optimizing the built deep neural network.
Further, in step (3), the received signal yCP(n) comprises:
the received signal of length N is:
Figure BDA0002768486420000053
wherein n is a discrete time index; h (n) is the impulse response function of the channel, and w (n) is additive white gaussian noise.
Further, in the step (4), the building of the neural network includes: establishing a depth receiver model on the basis of DenseNet;
the depth receiver model includes: the system comprises a convolutional layer, four conversion modules, four dense modules, a transition module, a global maximum pooling layer, a global average pooling layer and a depth connection layer;
each dense module comprises a different number of layers;
the transition module includes: processing the classified layer, the ReLu layer and the convolution layer;
the size of the global maximum pooling layer filter is 3 x 1, and the step length is 2; the global maximum pooling layer and the global average pooling layer are used for acquiring feature vectors;
the sizes of convolution kernels of all convolutions of the depth receiver model are one-dimensional.
Further, in step four, the optimizing the built deep neural network includes:
and (3) optimizing the neural network by using a random gradient descent method:
the loss function is determined using the intersection and determination of M classifiers: for a minimum batch containing N samples, the loss function is defined as:
Figure BDA0002768486420000061
wherein, cimkRepresenting the output probability of the ith classifier on the mth classifier on the kth class, dimkIndicating the kth true label corresponding to the mth bit of the ith sample; one-hot encoding is used for tags, i.e. when a bit in the real information bit stream is 0, the corresponding tag is [1,0 ]]TOtherwise the label is [0, 1]]T。
Further, in step three, the actually testing the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system includes:
1) the transmitting terminal generates known unknown random binary bit stream, and the known unknown random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure BDA0002768486420000071
Wherein k is a carrier serial number;
4) and inputting the obtained output data y (k) into the obtained deep receiver model of the mine MIMO-OFDM communication receiving system, and recovering the original transmitting signal.
The invention also aims to provide a mine communication system based on the depth receiver, which implements the mine communication method based on the depth receiver.
It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the depth receiver based mine communication method.
It is a further object of the present invention to provide a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform a method of depth receiver based mine communication.
By combining all the technical schemes, the invention has the advantages and positive effects that: the invention provides a mine MIMO-OFDM communication method based on a deep receiver, which adopts an MIMO-OFDM system to improve mine wireless communication performance, utilizes a deep neural network to recover original information bit streams from the receiver to a distorted IQ signal, replaces the whole information recovery process from carrier and symbol synchronization to channel estimation, equalization, demodulation and decoding of a traditional receiver, and overcomes the influence of wireless channel fading, noise and co-channel interference.
The invention adopts the mine wireless communication system based on the MIMO-OFDM technology, improves the capacity of the system, reduces the error rate of the system and can effectively resist the multipath fading of the channel.
The communication method of the invention replaces the traditional wireless communication receiver with a deep receiver model; different from the traditional mine communication which needs to carry out definite channel estimation and equalization, the deep neural network can learn and process the complex distortion caused by the mine channel through training and then directly recover the original transmitting signal from the receiving signal. The deep receiver can learn from the data to better match the non-idealities experienced by the communication system. The deep receiver has good performance under the non-ideal conditions of noise, radio frequency damage, multipath fading, co-channel interference and the like.
The invention can reduce the design complexity of the mine communication system and effectively realize the data transmission under the mine. Meanwhile, compared with the traditional mine communication method, the method has better robustness for mine communication under the conditions of less pilot frequency data amount and cyclic prefix missing.
The space-time coding scheme of the invention adopts orthogonal space-time block coding, because of the orthogonality of the coding matrix array and the column, the receiving end can use the maximum likelihood detection decoding, thereby greatly reducing the complexity of the decoding and still obtaining the maximum transmission diversity gain; meanwhile, due to the special environment of mine communication, the invention adopts orthogonal space-time block coding, and the orthogonal space-time coding not only can obtain the maximum diversity gain, but also has very low decoding complexity.
The invention realizes the modulation and demodulation of the system by adopting IFFT transformation and FFT transformation for each subchannel modulated by OFDM, and has small operation amount and simple realization when the number of the subchannels is large.
The capacity of the MIMO-OFDM system is simulated, the simulation condition assumes that the channel attenuation coefficient obeys Nakagami distribution, the simulation result fully shows that the channel capacity is increased along with the increase of the number of the antennas at the transmitting end and the receiving end, and the channel capacity of the MIMO system is also greatly increased compared with the SISO system.
The special production environment and safe production in the underground coal mine require that the mine wireless communication must work accurately and reliably, and the space-time coding gain directly influences the performance of the system. The mine tunnel is a strong correlation channel, and each coding scheme based on the STBC not only needs to fully consider the influence of compensating the channel correlation to cause the performance reduction of the system, but also needs to consider the low complexity of coding and decoding. The space-time coding gain will directly affect the performance of the system.
The influence of the channel correlation on the system performance is simulated, and the simulation result shows that the larger the correlation coefficient is, the larger the loss of the signal-to-noise ratio of the system is under the condition that the system allows transmission errors, and when the correlation coefficient approaches to 1, the coding diversity gain disappears completely, and the system performance is seriously deteriorated.
The simulation result shows that compared with SISO system, MIMO-OFDM system can increase transmission capacity by increasing number of receiving and transmitting antennas, and effectively resist the influence of multipath fading. The influence of the channel correlation on the system is analyzed, and the condition that the moderate space-time coding scheme is adopted as much as possible on the premise of reducing the decoding complexity is pointed out, so that the channel correlation of the system is reduced.
In the present invention, a deep receiver model is proposed that uses a deep neural network instead of a conventional receiver for the entire information recovery process from the received IQ signal to the recovered information bit stream. Recovery of the multi-bit information stream is achieved using multiple binary classifiers sharing the CNN. A 1D-Conv-DenseNet network structure was designed to implement this module. Through a large number of simulation experiments, the depth receiver has the following characteristics:
1) the performance of the deep receiver can be close to ideal soft decision, which is far superior to a hard decision method which is subjected to serial processing such as equalization, demodulation, decoding and the like, and the end-to-end information recovery capability is proved.
2) The deep receiver can learn from the data to better match the non-idealities experienced by the communication system. Simulation results prove that the receiver has good performance under the non-ideal conditions of noise, radio frequency damage, multipath fading and the like.
3) The deep receiver has strong interference resistance. Simulation results show that the deep receiver has good anti-interference capability under different ISRs, which shows that the deep receiver can be used as a new anti-interference communication method.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art that other drawings can be obtained from the drawings without creative efforts.
Fig. 1 is a schematic diagram of a mine communication method based on a depth receiver according to an embodiment of the present invention.
Fig. 2 is a flowchart of a mine MIMO-OFDM communication method based on a deep receiver according to an embodiment of the present invention.
Fig. 3 is a transmission flow chart of a MIMO-OFDM system according to an embodiment of the present invention.
Fig. 4 is a flowchart of a concept of a depth receiver according to an embodiment of the present invention.
Fig. 5 is a structural diagram of a 1D-conv-densenet according to an embodiment of the present invention.
Fig. 6 is a structural diagram of DenseBlock 1 according to an embodiment of the present invention.
Fig. 7 is a structural diagram of DenseBlock 2 and DenseBlock 4 according to an embodiment of the present invention.
Fig. 8 is a structural diagram of DenseBlock 3 according to an embodiment of the present invention.
Fig. 9 is a Nakagami distribution probability density graph provided by an embodiment of the present invention.
Fig. 10 is a simulation diagram of the influence of the spatial correlation of the channel on the channel capacity according to the embodiment of the present invention.
Fig. 11 is a diagram illustrating simulation of the performance of a deep receiver at different carrier frequency offsets according to an embodiment of the present invention.
Fig. 12 is a partial procedure for building a deep neural network according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Aiming at the problems in the prior art, the invention provides a mine MIMO-OFDM communication method based on a deep receiver, and the invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1 to fig. 3, a mine MIMO-OFDM communication method based on a deep receiver provided by an embodiment of the present invention includes the following steps:
s101, constructing an MIMO-OFDM system model of a mine roadway;
s102, obtaining a deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system by establishing a deep neural network and carrying out effective training;
and S103, carrying out actual test on the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system until the original transmitting signal is recovered.
In step S101, the MIMO-OFDM system model of the mine roadway provided in the embodiment of the present invention is as follows:
the MIMO-OFDM system model input-output relation is expressed as:
Figure BDA0002768486420000111
wherein, ykRepresents a k-th sub-channel received signal vector; h denotes the impulse response of each subchannel;
Figure BDA0002768486420000112
which transmitted signal vector is the k sub-carrier of the ith antenna; n iskRepresenting white gaussian noise on the kth sub-channel;
the impulse response of each subchannel is as follows:
Figure BDA0002768486420000113
in step S102, the depth receiver model for obtaining the effectively trained mine MIMO-OFDM communication receiving system by establishing the depth neural network and performing effective training provided in the embodiment of the present invention includes:
(1) the transmitting terminal generates known random binary bit stream, and the known random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
(2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
(3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure BDA0002768486420000123
Wherein k is a carrier serial number;
(4) building a deep neural network; and optimizing the built deep neural network.
In step (3), the received signal y provided by the embodiment of the present inventionCP(n) comprises:
the received signal of length N is:
Figure BDA0002768486420000121
wherein n is a discrete time index; h (n) is the impulse response function of the channel, and w (n) is additive white gaussian noise.
As shown in fig. 4 to 8, in step (4), the neural network building method provided by the embodiment of the present invention includes: establishing a depth receiver model on the basis of DenseNet;
the depth receiver model includes: the system comprises a convolutional layer, four conversion modules, four dense modules, a transition module, a global maximum pooling layer, a global average pooling layer and a depth connection layer;
each dense module comprises a different number of layers;
the transition module includes: processing the classified layer, the ReLu layer and the convolution layer;
the size of the global maximum pooling layer filter is 3 x 1, and the step length is 2; the global maximum pooling layer and the global average pooling layer are used for acquiring feature vectors;
the sizes of convolution kernels of all convolutions of the depth receiver model are one-dimensional.
In step S104, the optimizing the built deep neural network provided by the embodiment of the present invention includes:
and (3) optimizing the neural network by using a random gradient descent method:
the loss function is determined using the intersection and determination of M classifiers: for a minimum batch containing N samples, the loss function is defined as:
Figure BDA0002768486420000122
wherein, cimkRepresenting the output probability of the ith classifier on the mth classifier on the kth class, dimkIndicating the kth true label corresponding to the mth bit of the ith sample; one-hot encoding for tagsThat is, when the bit in the real information bit stream is 0, the corresponding tag is [1,0 ]]TOtherwise the label is [0, 1]]T。
In step S103, the actual test of the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system provided in the embodiment of the present invention includes:
1) the transmitting terminal generates known unknown random binary bit stream, and the known unknown random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure BDA0002768486420000132
Wherein k is a carrier serial number;
4) and inputting the obtained output data y (k) into the obtained deep receiver model of the mine MIMO-OFDM communication receiving system, and recovering the original transmitting signal.
The technical solution of the present invention is further described below with reference to specific examples.
Example 1:
a mine MIMO-OFDM communication method based on a depth receiver comprises the following steps:
step 1, building an MIMO-OFDM system model of a mine roadway.
The input-output relationship of the MIMO system is expressed as
Figure BDA0002768486420000131
In the formula x [ n ]]——NTTransmitting a signal vector in x 1 dimension;
y[n]——NRa x 1-dimensional received signal vector;
n (t) -additive white gaussian noise.
For frequency selective deep fading, the OFDM modulation is converted into flat fading of a plurality of subcarriers, and the impulse response of each subchannel is
Figure BDA0002768486420000141
Thus, for frequency fading, the input-output relationship of the MIMO-OFDM system is expressed as
Figure BDA0002768486420000142
Wherein
Figure BDA0002768486420000143
Which is the transmitted signal vector for the k-th subcarrier of the ith antenna.
In the formula yk-the kth sub-channel receives the signal vector;
nk-white gaussian noise on the kth sub-channel.
The MIMO-OFDM system can improve the transmission capacity by increasing the number of the receiving and transmitting antennas, and effectively resist the influence of multipath fading.
Step 2, training: a deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system is obtained by establishing a deep neural network and carrying out effective training;
step 3, a testing stage: and (3) putting the deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system obtained in the step (1) into practical test until the original transmitting signal is recovered.
The effectively trained deep receiver model of the mine MIMO-OFDM communication receiving system in the step 2 is obtained through the following steps:
step 2.1, the input data generated at the input of the communication system is channel coded, interleaved, QAM constellation modulated and mapped and then processedThe line space time block coding is changed into multi-path output, then each path of signal is processed with serial-parallel conversion and IFFT conversion to realize OFDM modulation, and then a guard interval Cyclic Prefix (CP) is inserted, finally the signal is modulated on a carrier wave and is amplified in power, and the signal is transmitted out through a transmitting antenna. Wherein the mathematical expression of the modulated signal is xCP(n), wherein n is a discrete time index;
step 2.2, modulated signal x of step 2.1CP(N) a received signal y of length N, which reaches the communication receiving end after passing through the channelCP(n) is:
Figure BDA0002768486420000151
wherein h (n) is the impulse response function of the channel, and w (n) is additive white gaussian noise;
step 2.3, preprocessing the received data, and sequentially performing serial-parallel conversion, cyclic prefix removal and Fourier transform to obtain output data
Figure BDA0002768486420000153
Where k is the carrier number.
And 2.4, building a deep neural network, and building a deep receiver network on the basis of the DenseNet, wherein the sizes of convolution kernels of all convolutions are one-dimensional. The function in 1D-Conv-DenseNet includes three operations: batch normalization, ReLu linear units and 5 × 1 convolution. This basic operation is represented using basicblock (K), where K represents the number of convolution kernels in the convolutional layer of the module. In the 1D-Conv-DenseNet of the invention, the transition module comprises a processing layer, a ReLu layer, a maximum pooling layer with the filter size of 3 x 1 and the step length of 2, and a convolution layer with the kernel size of 5 x 1 (the number of kernels is determined according to the requirement). The transform block is represented using TransitionBlock (K), where k represents the number of convolution kernels in this block.
The general structure of 1D-Conv-DenseNet is shown in FIG. 4. It consists of a series of dense modules and conversion modules. As can be seen from the figure, the network has four transitionblocks and four denseblocks, each of which contains a different number of layers, as shown in fig. 5. After these transitionblocks and DenseBlock, feature vectors are obtained using a global maximum pooling layer and a global average pooling layer, from which each binary classifier computes the probability of each bit being 0 or 1.
The optimization method used by the present invention is the Stochastic Gradient Descent (SGD) method, which uses the cross-sums of M classifiers to design the penalty function. For a minimum batch containing N samples, the loss function is defined as
Figure BDA0002768486420000152
Wherein, cimkThe output probability of the ith classifier to the mth classifier to the kth class for the ith sample as input, dimkThe m bit of the ith sample corresponds to the k true label. One-hot encoding is used for tags, i.e. when a bit in the real information bit stream is 0, the corresponding tag is [1,0 ]]TOtherwise the label is [0, 1]]T。
The deep learning test process in the step 3 specifically comprises the following operation steps:
step 3.1, step 3.2, step 3.3 synchronization step 2.1, step 2.2, step 2.3 are consistent, wherein the input data of the communication system is changed from the known random binary bit stream to the unknown random binary bit stream, and the input data of the neural network is obtained;
and 3.4, inputting the y (k) obtained in the step 3.3 into the deep receiver model of the mine MIMO-OFDM communication receiving system obtained in the step 2.4, and further recovering the original transmitting signal.
Step 2.1 transmitting signals comprises the following specific operating scheme:
the space-time coding scheme of the system adopts orthogonal space-time block coding, because of the orthogonality of the coding matrix array and the column, the receiving end can use the maximum likelihood detection decoding, thereby greatly reducing the complexity of the decoding and still obtaining the maximum transmission diversity gain; meanwhile, due to the special environment of mine communication, the system adopts orthogonal space-time block coding, and the orthogonal space-time coding not only can obtain the maximum diversity gain, but also has very low decoding complexity.
Each sub-channel of the OFDM modulation of the system adopts IFFT transformation and FFT transformation to realize the modulation and demodulation of the system, and when the number of the sub-channels is large, the operation amount of the system is small, and the realization is simple.
Demonstration section (concrete examples/experiments/simulation/pharmacological analysis/positive experimental data capable of demonstrating the inventive aspects of the invention, etc.)
1) In order to fully embody the fading characteristics of the mine wireless channel, the amplitude distribution of the multipath signals in the mine roadway is described by the Nakagami distribution with variable parameters. The Nakagami distribution not only contains the characteristics of Rayleigh distribution and Rician distribution, but also can be converted into different distributions by changing the size of the parameter m, so that the amplitude statistical characteristics of multipath signals in the mine roadway can be better described. FIG. 9 is a Nakagami distribution probability density graph.
2) The existence of the spatial correlation of the channel reduces the channel capacity without changing the number of antennas at the transmitting and receiving ends. When the SNR is 10d B, the channel capacity of the correlated channel of the 2 × 2 antenna system is reduced by about 0.9bps/Hz compared to the channel capacity of the uncorrelated channel; the capacity of the associated channel of the 2 x 2 antenna system is still significantly increased compared to the channel capacity of the SISO system, with an increment of about 3bps/Hz, and the channel capacity increases with increasing SNR. In addition, as can be seen from fig. 10, the channel capacity of the correlated channel of the 3 × 3 antenna system is significantly increased compared to the channel capacity of the uncorrelated channel of the 2 × 2 antenna system. Therefore, even if the channel has spatial correlation, increasing the number of antennas at the transmitting and receiving ends can still increase the channel capacity. Fig. 10 is a diagram showing simulation results of the influence of the spatial correlation of the channel on the channel capacity.
3) In a wireless communication system, two independent local oscillations are used at the transmitting end and the receiving end. Their frequencies may have some deviation. The performance of the receiver in the presence of carrier frequency offset was analyzed. In the simulation, the normalized carrier frequency offset Δ f (relative to the symbol rate) was randomly generated within the range of [ -0.010.01 ]. As can be seen from the simulation results of fig. 11, the conventional hard decision is greatly affected by carrier frequency deviation. The performance is significantly reduced with increasing carrier frequency offset, especially when Δ f is 0.004, the BER of Eb/N0 is worse than 0.01 in the range of 0-8 dB. Such error rate performance will be difficult to meet the requirements of practical applications. However, the BER performance of the deep receiver is still very close to the ideal soft decision, which means that it can overcome the effect of carrier frequency offset to some extent.
The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. A mine communication method based on a depth receiver is characterized by comprising the following steps:
constructing an MIMO-OFDM system model of a mine tunnel;
a deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system is obtained by establishing a deep neural network and carrying out effective training;
and actually testing the obtained deep receiver model of the effectively trained mine MIMO-OFDM communication receiving system until the original transmitting signal is recovered.
2. The method of claim 1, wherein the MIMO-OFDM system model of the mine roadway is as follows:
the MIMO-OFDM system model input-output relation is expressed as:
Figure FDA0002768486410000011
wherein, ykRepresents a k-th sub-channel received signal vector; h denotes the impulse response of each subchannel;
Figure FDA0002768486410000012
which transmitted signal vector is the k sub-carrier of the ith antenna; n iskRepresenting white gaussian noise on the kth sub-channel;
the impulse response of each subchannel is as follows:
Figure FDA0002768486410000013
3. the method of claim 1, wherein the obtaining the effectively trained deep receiver model of the mine MIMO-OFDM communication receiving system by establishing the deep neural network and performing effective training comprises:
(1) the transmitting terminal generates known random binary bit stream, and the known random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
(2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
(3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure FDA0002768486410000022
Wherein k is a carrier serial number;
(4) building a deep neural network; and optimizing the built deep neural network.
4. The method of claim 3, wherein in step (3), the received signal y is receivedCP(n) comprises:
the received signal of length N is:
Figure FDA0002768486410000021
wherein n is a discrete time index; h (n) is the impulse response function of the channel, and w (n) is additive white gaussian noise.
5. The mine communication method based on the depth receiver as claimed in claim 3, wherein in the step (4), the building of the neural network comprises: establishing a depth receiver model on the basis of DenseNet;
the depth receiver model includes: the system comprises a convolutional layer, four conversion modules, four dense modules, a transition module, a global maximum pooling layer, a global average pooling layer and a depth connection layer;
each dense module comprises a different number of layers;
the transition module includes: processing the classified layer, the ReLu layer and the convolution layer;
the size of the global maximum pooling layer filter is 3 x 1, and the step length is 2; the global maximum pooling layer and the global average pooling layer are used for acquiring feature vectors;
the sizes of convolution kernels of all convolutions of the depth receiver model are one-dimensional.
6. The method of claim 1, wherein the optimizing the constructed deep neural network comprises:
and (3) optimizing the neural network by using a random gradient descent method:
the loss function is determined using the intersection and determination of M classifiers: for a minimum batch containing N samples, the loss function is defined as:
Figure FDA0002768486410000031
wherein, cimkRepresents the ith sample as input to the ith classifier versus the mth classifierOutput probability for the kth class, dimkIndicating the kth true label corresponding to the mth bit of the ith sample; one-hot encoding is used for tags, i.e. when a bit in the real information bit stream is 0, the corresponding tag is [1,0 ]]TOtherwise the label is [0, 1]]T。
7. The method of claim 1, wherein the actual testing of the resulting deep receiver model of the actively trained mine MIMO-OFDM communication receiving system comprises:
1) the transmitting terminal generates known unknown random binary bit stream, and the known unknown random binary bit stream is subjected to channel coding, interleaving and QAM constellation modulation mapping, and is subjected to space-time block coding to be converted into multi-channel output;
2) after each path of signal is processed with serial-to-parallel conversion and IFFT conversion for OFDM modulation, a guard interval cyclic prefix is inserted and modulated to a carrier wave for power amplification, and a modulated and amplified signal x is processedCP(n) transmitting using a transmit antenna;
3) the communication receiving end receives the transmitted modulation signal; and performing serial-to-parallel conversion, cyclic prefix removal and Fourier transform on the received data to obtain output data
Figure FDA0002768486410000032
Wherein k is a carrier serial number;
4) and inputting the obtained output data y (k) into the obtained deep receiver model of the mine MIMO-OFDM communication receiving system, and recovering the original transmitting signal.
8. A mine communication system based on a depth receiver is characterized in that the mine communication system based on the depth receiver implements the mine communication method based on the depth receiver according to any one of claims 1 to 7.
9. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the depth receiver based mine communication method of any one of claims 1 to 7.
10. A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to carry out the method of depth receiver based mine communication of any one of claims 1 to 7.
CN202011241415.XA 2020-11-09 2020-11-09 Mine communication method, system, computer equipment and medium based on depth receiver Pending CN112383496A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011241415.XA CN112383496A (en) 2020-11-09 2020-11-09 Mine communication method, system, computer equipment and medium based on depth receiver

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011241415.XA CN112383496A (en) 2020-11-09 2020-11-09 Mine communication method, system, computer equipment and medium based on depth receiver

Publications (1)

Publication Number Publication Date
CN112383496A true CN112383496A (en) 2021-02-19

Family

ID=74579100

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011241415.XA Pending CN112383496A (en) 2020-11-09 2020-11-09 Mine communication method, system, computer equipment and medium based on depth receiver

Country Status (1)

Country Link
CN (1) CN112383496A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113300788A (en) * 2021-04-19 2021-08-24 嘉兴学院 Blind receiver method and device based on Capsule network

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108062756A (en) * 2018-01-29 2018-05-22 重庆理工大学 Image, semantic dividing method based on the full convolutional network of depth and condition random field
CN109067688A (en) * 2018-07-09 2018-12-21 东南大学 A kind of OFDM method of reseptance of data model double drive
CN109246038A (en) * 2018-09-10 2019-01-18 东南大学 A kind of GFDM Receiving machine and method of data model double drive
JP2019016967A (en) * 2017-07-10 2019-01-31 Ｈｅｒｏｚ株式会社 Receiver and receiving method
CN109412996A (en) * 2018-12-11 2019-03-01 深圳大学 Chain circuit self-adaptive method, electronic device and computer readable storage medium
CN109474352A (en) * 2018-12-24 2019-03-15 哈尔滨工程大学 A kind of underwater sound orthogonal frequency division multiplexing communication method based on deep learning
KR101992053B1 (en) * 2018-03-22 2019-06-21 세종대학교산학협력단 SISO-OFDM channel estimation apparatus using deep neural network based on adaptive ensemble supervised learning and method thereof

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2019016967A (en) * 2017-07-10 2019-01-31 Ｈｅｒｏｚ株式会社 Receiver and receiving method
CN108062756A (en) * 2018-01-29 2018-05-22 重庆理工大学 Image, semantic dividing method based on the full convolutional network of depth and condition random field
KR101992053B1 (en) * 2018-03-22 2019-06-21 세종대학교산학협력단 SISO-OFDM channel estimation apparatus using deep neural network based on adaptive ensemble supervised learning and method thereof
CN109067688A (en) * 2018-07-09 2018-12-21 东南大学 A kind of OFDM method of reseptance of data model double drive
CN109246038A (en) * 2018-09-10 2019-01-18 东南大学 A kind of GFDM Receiving machine and method of data model double drive
CN109412996A (en) * 2018-12-11 2019-03-01 深圳大学 Chain circuit self-adaptive method, electronic device and computer readable storage medium
CN109474352A (en) * 2018-12-24 2019-03-15 哈尔滨工程大学 A kind of underwater sound orthogonal frequency division multiplexing communication method based on deep learning

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
YUE QUE等: "Densely_Connected_Convolutional_Networks_for_Multi-Exposure_Fusion", 《2018 INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE AND COMPUTATIONAL INTELLIGENCE (CSCI)》 *
孟宪猛等: "基于MIMO技术的矿井无线通信系统的设计", 《煤炭工程》 *
陈珊珊等: "MIMO-OFDM技术在矿井无线通信系统中的应用", 《煤矿机械》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113300788A (en) * 2021-04-19 2021-08-24 嘉兴学院 Blind receiver method and device based on Capsule network

Similar Documents

Publication Publication Date Title
CN106357311B (en) Detection method of MIMO-OFDM system based on carrier index modulation
CN101039137B (en) Method and system for reducing codebook search-based precoding feedback bits of MIMO-OFDM system
CN102624666B (en) Cyclic coding method of multi-channel transceiving orthogonal multi-carrier underwater acoustic communication in sparse channel model
CN104243069B (en) A kind of multiple antennas traffic communication network system and signal detecting method
CN101453444B (en) Constellation map simplifying apparatus, method and receiver
CN100571098C (en) The maximum likelihood detecting method of low complex degree and device in the communication system
CN104702390A (en) Pilot frequency distribution method in distributed compressive sensing (DCS) channel estimation
CN105141563A (en) Space frequency combined modulation design scheme used for MIMO-OFDM system
CN109167649B (en) Low-complexity detection method for GSM-MBM system
CN112215335B (en) System detection method based on deep learning
CN113114603B (en) Information recovery method and device for MIMO-OFDM system
CN105162527A (en) Underwater communication hybrid carrier method based on weighted fractional order Fourier transform
CN110932766A (en) Multi-carrier spatial modulation transmission method for reducing radio frequency switching times
CN112383496A (en) Mine communication method, system, computer equipment and medium based on depth receiver
CN107196880A (en) A kind of phase noise compensation method in differential space-time coding
CN102710567A (en) Part judgment method in interference elimination technology for multiple-input multiple-output (MIMO) wireless communication receiver
Li et al. Model-driven deep learning scheme for adaptive transmission in MIMO-SCFDE system
CN113938234B (en) Low-complexity sparse large-scale MIMO detection method
Zhao et al. Adaptive power allocation for non-coherent FSK in time-varying underwater acoustic communication channels
CN114745230B (en) OTFS signal receiving and recovering method based on deep neural network structure
CN113709075B (en) Method for realizing underwater acoustic communication receiver by using underwater acoustic channel multipath effect
CN102006250A (en) Turbo enhancement method for MIMO-SCFDE wireless communication receiver
CN1677892A (en) Space hour encoding method based on rotary planisphere
CN102035787A (en) Band sequencing Turbo enhancement method for multiple-input multiple-output-orthogonal frequency division multiplexing (MIMO-OFDM) wireless communication receiver
CN113556159A (en) Channel feedback method of large-scale MIMO multi-user system

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
AD01 Patent right deemed abandoned

Effective date of abandoning: 20230307

AD01 Patent right deemed abandoned