CN113325375A - Self-adaptive cancellation method based on deep neural network - Google Patents
Self-adaptive cancellation method based on deep neural network Download PDFInfo
- Publication number
- CN113325375A CN113325375A CN202110569844.8A CN202110569844A CN113325375A CN 113325375 A CN113325375 A CN 113325375A CN 202110569844 A CN202110569844 A CN 202110569844A CN 113325375 A CN113325375 A CN 113325375A
- Authority
- CN
- China
- Prior art keywords
- signal
- model
- nonlinear
- self
- amplitude
- 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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/38—Jamming means, e.g. producing false echoes
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Noise Elimination (AREA)
Abstract
The invention provides a self-adaptive cancellation method based on a deep neural network, 1) a signal model received by a receiving antenna is defined, and the signal model comprises transmitted signal power PfNonlinear distortion function G [. for power amplifier]And carrier center frequency fc(ii) a 2) Defining a model of a non-linear power amplifier; 3) carrying out nonlinear modeling on a target signal, and training a DNN network by using a large amount of data; 4) inputting a signal generated after the original reference signal passes through a trained network as a new reference signal into an adaptive filter; 5) and comparing signals before and after cancellation by the adaptive filter. The invention utilizes a large amount of training prior information to simulate the nonlinear characteristic of a power amplifier of the radar jammer, solves the interference problem, directly estimates the amplitude of a signal by the method, and reduces algorithm steps by using a large amount of data.
Description
Technical Field
The invention belongs to the field of self-interference cancellation of radar jammers, and particularly relates to a self-adaptive cancellation method based on a deep neural network.
Background
The self-coupling interference elimination in the radar jammer is always a key technology and a hot topic, and in recent years, a self-adaptive algorithm is widely applied to the self-coupling interference elimination. With the increasing complexity of the electromagnetic environment, the self-interference signal is difficult to estimate, and the traditional algorithm is difficult to adapt to the nonlinear signal generated by the radar jammer power amplifier, so that the self-interference signal is effectively eliminated.
With the continuous and deep research of the adaptive algorithm, the application of the adaptive algorithm in the radar jammer is more and more mature. Currently, adaptive filtering algorithms, such as Normalized Least Mean Square Error (NLMS), are mainly used to eliminate self-interference signals. Lee and B.Min ("Results and trade-off of self-interference cancellation in a full-duplex radio front-end,"2015 International Workshop on Antenna Technology (iWAT), Seoul, pp.249-251,2015.) demonstrates that the error signal approaches the target value more and more as the adaptive filter weights are adaptively changed. The method only has a good experimental result for the radio frequency cancellation in the indoor electromagnetic environment, and the nonlinearity of the self-interference signal is not considered. The adaptive digital self-interference cancellation optimization algorithm based on the improved variable step length is provided by L.Sun, Y.Li, Y.ZHao, L.Huang and Z.Gao ('Optimized adaptive optimization of digital self-interference based on improved variable step length,' 2015 IEEE 9th International Conference on Anti-correlation, Security, and Identification (ASID), Xiamen, pp.176-179,2015.), a new nonlinear relation is established between the step factor and the error signal by using an iteration threshold, the problem that the error signal changes slowly when approaching zero is solved, and the convergence speed is accelerated. The method does not mention cancellation of radar jammer non-linear self-interference signals. Dani Korpi, Lauri Anttila, and Mikko Valkama ("Nonlinear selection-interference cancellation in MIMO full-duplex transceivers under cross talk." "Eurasip Journal on Wireless Communications & Networking 2017.1 (2017)") proposes a novel digital self-interference canceller for in-band multiple-input multiple-output (MIMO) full-duplex radios. It details various models of full duplex including analysis of the non-linear component of the self-interference signal, but does not suggest how to eliminate the non-linear self-interference signal. In summary, the cancellation methods above are all simple for electromagnetic environment, and have a good cancellation effect when the generated self-interference signal is linear, and when the non-linear self-interference signal is generated due to the radar jammer power amplifier, it is not easy to obtain a good cancellation result when the cancellation results are implemented by using these methods.
Because the neural network can well solve the nonlinear problem, recent studies of scholars show that the neural network can be used for channel modeling in full-duplex communication, and therefore reconstructed channel signals are more accurate. This method is based on adaptive filtering, where the weights of the filter are estimated by means of DNN (deep Neural network), and the non-linear performance of the adaptive filter itself is not particularly good, so DNN is introduced to optimize the performance of the adaptive filter.
Disclosure of Invention
The invention provides a self-adaptive cancellation method based on a deep neural network, which is used for solving the problem that the cancellation effect is poor when a traditional self-adaptive cancellation algorithm processes self-interference signals formed by a nonlinear power amplifier. The invention utilizes a large amount of training prior information to simulate the nonlinear characteristic of a power amplifier of the radar jammer, solves the interference problem, directly estimates the amplitude of a signal by the method, and reduces algorithm steps by using a large amount of data.
The purpose of the invention is realized as follows:
a self-adaptive cancellation method based on a deep neural network comprises the following steps:
step 1: defining a model of the signal received by the receiving antenna, including the power P of the transmitted signalfNonlinear distortion function G [. for power amplifier]And carrier center frequency fc；
The received signals of the receiving antenna comprise target signals, noise signals and self-interference signals; in the normal case, the noise signal is usually white Gaussian noise with zero mean, and is represented by n: (t) represents the power Pn(ω), i.e.
Defining the expected target signal r (t) received by the receiving antenna as:
wherein, PfIs the power of the radar transmitted signal, G [. cndot.)]Representing the nonlinear distortion function of the power amplifier, df(t) represents a modulated baseband waveform, fcRepresents the center frequency of the carrier;
the self-interference signal in the radar jammer is a delay of the transmitted signal, so the self-interference signal x (t) can be set as follows:
x(t)＝r(t-τ1)+r(t-τ2)+······+r(t-τn) (3)
wherein, tau1、τ2And τnIs to simulate the interference delay in the actual situation;
after the target signal is modulated by baseband, dn(t) is obtained from x (t), xPA(t) is obtained by power amplification; and finally, the transmitting antenna sends out the self-interference signal SI (t) formed by the receiving antenna, wherein the transmitting signal is inevitably input:
Pnpower of the signal transmitted by the jammer, fcRepresents the center frequency of the carrier, phi represents the initial phase of the carrier, and k represents a known coefficient;
therefore, the model of the signal y (t) actually received by the receiving antenna is:
y(t)＝r(t)+SI(t)+n(t) (5)
step 2: defining a model of a non-linear power amplifier;
the reason for causing the nonlinear distortion of the power amplifier is mainly AM/AM distortion, wherein the AM/AM distortion refers to the distortion of the amplitude of an output signal caused by the amplitude change of an input signal; the adopted power amplifier is a traveling wave tube amplifier, the nonlinear distortion of the traveling wave tube amplifier can be described by a saliche model, and the AM/AM and AM/PM of the saliche model have the characteristic functions as follows:
where r is the amplitude of the input signal; alpha is alphaa、βa、αφAnd betaφIs a model parameter; obtaining a proper fixed model by adjusting the four parameters; nonlinear characteristics of the salich model include amplitude and phase changes of the signal after passing through the salich model;
taking the derivative of equation (6) to obtain the input signal amplitude when the input signal reaches the input saturation amplitude of the amplifier model:
the model obtains the maximum output signal amplitude:
f(A)max＝αaAsat/2 (9)
the maximum output value of the power amplifier determines the maximum value that the linearization can correct; if the linear output value corresponding to the input amplitude value of the power amplifier is larger than the maximum output amplitude value of the power amplifier, the nonlinear distortion cannot be compensated;
and step 3: carrying out nonlinear modeling on a target signal, and training a DNN network by using a large amount of data;
3a) the training data takes the form of two signals, namely a Linear Frequency Modulation (LFM) signal and a BPSK signal, as target signals respectively; for the twoSignal, two data sets were made separately, each data set containing 10000 samples; each LFM sample is a pulse signal with the pulse width of 3 mu s, each BPSK sample comprises 13 symbols, and the number of sampling points of each symbol at the sampling frequency is 70; for the LFM signal, the LFM signal sampling frequency f is defineds300MHz, carrier frequency 100MHz, bandwidth 20MHz, signal amplitude 20; defining the sampling frequency of BPSK signals as fs300MHz, amplitude 20, carrier frequency 50 MHz;
3b) the target signal generated in 3a is a label of the training sample obtained by the salich model in step 2 (the saturation output amplitude of the model is 80); wherein the parameters of the saleach model are set as followsa＝2，αφ＝π/3，βa＝1.5625e-5，β φ1 is ═ 1; generating a target signal as a training signal of the DNN network, and taking a signal passing through a Saichh model as a label of training data; after nonlinear amplification, the data are obviously changed in both time domain and frequency domain;
3c) the input layer of the DNN network is x (t), the number of nodes of the hidden layer is respectively amplified and reduced, the final number of nodes of the output layer is the same as that of the input layer, and a nonlinear amplified output signal x is obtainedAP(t); in the experiment, the hidden layers use Relu (transformed linear units) activation functions, and the number of hidden layers of the deep neural network is Num, 1024, 2048, …, 1024, Num; wherein Num is the number of points of each sample, and the specific layer number can be adjusted; the loss quadratic cost function in a neural network is:
Loss＝mean(square(x-xAP)) (10)
the output of each node of the DNN network is a nonlinear activation function to which its inputs are applied; weights between layers in the neural network are optimized through extensive learning, and expected outputs of training samples containing known inputs are learned;
and 4, step 4: inputting a signal generated after the original reference signal passes through a trained network as a new reference signal into an adaptive filter;
the receiver receives signals including a target signal, a noise signal and a self-interference signal, the sum of the signals is a signal before cancellation, and a self-interference signal SI (t) in formula (5) is a target to be cancelled;
4a) taking LFM signal as an example, the received signal is denoised and sampled to become y (n), which is the cancellation signal input to the adaptive filter and has a duration of 3 μ s,
4b) delaying the y (n) signal in the first step by 0.5 mu s, then carrying out nonlinear processing by using the Sa-Hz model in the step 2 to simulate the nonlinear characteristic of the power amplifier, and taking the processed signal as a self-interference signal SI (n), and adding SI (n), y (n) and white Gaussian noise as a target cancellation signal;
4c) inputting the y (n) signal delayed by 0.5 mu s in the second step into DNN neural network as training sample, using the signal after nonlinear processing by the Saichh model as label, and estimating the signal x by the deep neural networkAP(n) as a reference signal for the adaptive filter;
4d) inputting the signals generated in the steps 4b and 4c into the adaptive filter of the LMS algorithm to obtain an output signal x (n) after cancellation of the adaptive filter, and achieving the result of cancellation of the LFM signal;
the cancellation result can also be achieved by changing the LFM signal in the above steps into a BPSK signal.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides an interference machine signal interference elimination method based on a deep neural network. On the basis of the existing interference elimination theory, the nonlinear characteristic of a radar interference power amplifier is considered, and nonlinearity is introduced into modeling of an interference signal. The method is obtained through experimental simulation, and if the adopted training data is rich enough and the sampling frequency is not high, the nonlinear component in the self-interference signal generated by the nonlinear characteristic of the power amplifier and received in the jammer can be eliminated.
Drawings
Fig. 1 is a DNN-based self-interference cancellation system model.
Fig. 2 is a nonlinear characteristic of the salich model.
FIG. 3 shows LFM signals before and after non-linear amplification
Fig. 4 is a BPSK signal before and after non-linear amplification.
Fig. 5 is a block diagram of a DNN network model.
Fig. 6 is the result of LFM signal cancellation based on the DNN method.
Fig. 7 is a BPSK signal cancellation result based on the DNN method.
Fig. 8 is a result of the LFM signal cancellation of the conventional method.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
The technical scheme of the invention is a DNN-based adaptive cancellation algorithm, which comprises the following steps:
1) defining a model of the signal received by the receiving antenna, including the power P of the transmitted signalfNonlinear distortion function G [. for power amplifier]And carrier center frequency fc。
2) A model of a non-linear power amplifier is defined.
3) The target signal is non-linearly modeled and the DNN network is trained using a large amount of data.
4) And inputting a signal generated after the original reference signal passes through the trained network into the adaptive filter as a new reference signal.
5) And comparing signals before and after cancellation by the adaptive filter.
Step 1: defining a model of the signal received by the receiving antenna, including the power P of the transmitted signalfNonlinear distortion function G [. for power amplifier]And carrier center frequency fc。
The received signal of the receive antenna includes a target signal, a noise signal, and a self-interference signal. In the normal case, the noise signal is usually white Gaussian noise with zero mean, denoted by n (t), and has a power Pn(ω), i.e.
Defining the expected target signal r (t) received by the receiving antenna as:
wherein, PfIs the power of the radar transmitted signal, G [. cndot.)]Representing the nonlinear distortion function of the power amplifier, df(t) represents a modulated baseband waveform, fcRepresenting the center frequency of the carrier.
The self-interference signal in the radar jammer is a delay of the transmitted signal, so the self-interference signal x (t) can be set as follows:
x(t)＝r(t-τ1)+r(t-τ2)+······+r(t-τn) (3)
wherein, tau1、τ2And τnIs to simulate the interference delay in real situations.
After the target signal is modulated by baseband, dn(t) is obtained from x (t), xPA(t) is obtained by power amplification. And finally, the transmitting antenna sends out the self-interference signal SI (t) formed by the receiving antenna, wherein the transmitting signal is inevitably input:
Pnpower of the signal transmitted by the jammer, fcRepresents the center frequency of the carrier, phi represents the initial phase of the carrier, and k represents a known coefficient.
Therefore, the model of the signal y (t) actually received by the receiving antenna is:
y(t)＝r(t)+SI(t)+n(t) (5)
step 2: a model of a non-linear power amplifier is defined.
The reason for the nonlinear distortion of the power amplifier is mainly AM/AM distortion, which means that the amplitude variation of the input signal causes the distortion of the amplitude of the output signal. The power amplifier adopted by the invention is a traveling wave tube amplifier, the nonlinear distortion of the power amplifier can be described by a saliche model, and the characteristic functions of AM/AM and AM/PM of the saliche model are as follows:
where r is the amplitude of the input signal. Alpha is alphaa、βa、αφAnd betaφAre the model parameters. A suitable fixed model can be obtained by adjusting the four parameters. Fig. 2 shows the nonlinear characteristics of the salich model, which includes the amplitude and phase changes of the signal after passing through the salich model.
Taking the derivative of equation (6) can obtain the input signal amplitude when the input signal reaches the input saturation amplitude of the amplifier model as:
the model obtains the maximum output signal amplitude:
f(A)max＝αaAsat/2 (9)
the maximum output value of the power amplifier determines the maximum value that the linearization can correct. If the linear output value corresponding to the input amplitude value of the power amplifier is larger than the maximum output amplitude value of the power amplifier, the nonlinear distortion cannot be compensated.
And step 3: the target signal is non-linearly modeled and the DNN network is trained using a large amount of data.
3a) The training data is in the form of two signals, a Linear Frequency Modulated (LFM) signal and a BPSK signal, respectively, as target signals. For these two signals, two data sets were made, each containing 10000 samples. Each LFM sample is a pulse signal with a pulse width of 3 μ s, each BPSK sample contains 13 symbols, and the number of samples per symbol at the sampling frequency is 70. For LFM signals, determiningSampling frequency f of pseudo LFM signals300MHz, carrier frequency 100MHz, bandwidth 20MHz, signal amplitude 20; defining the sampling frequency of BPSK signals as fs300MHz, amplitude 20, and carrier frequency 50 MHz.
3b) The target signal generated in 3a is a label of the training sample obtained by the salich model in step 2 (the saturation output amplitude of the model is 80). Wherein the parameters of the saleach model are set as followsa＝2，αφ＝π/3，βa＝1.5625e-5，β φ1. The generated target signal serves as a training signal for the DNN network, and a signal passing through the salich model serves as a label for training data. As shown in fig. 3 and 4, after the nonlinear amplification, the data has a significant change in both time domain and frequency domain.
3c) FIG. 5 is a structural diagram of a DNN network model, wherein an input layer of the DNN network is x (t), the number of nodes of a hidden layer is respectively amplified and reduced, the final number of nodes of an output layer is the same as that of the input layer, and a nonlinear amplified output signal x is obtainedAP(t) of (d). In the experiment, the hidden layers used the relu (transformed linear units) activation function, and the number of hidden layers of the deep neural network was Num, 1024, 2048, …, 1024, Num, respectively. Where Num is the number of points per sample, the number of specific layers can be adjusted. The loss quadratic cost function in a neural network is:
Loss＝mean(square(x-xAP)) (10)
the output of each node of the DNN network is a nonlinear activation function that applies its inputs. The weights between layers in a neural network are optimized through extensive learning, and the expected output of training samples containing known inputs is learned.
And 4, step 4: and inputting a signal generated after the original reference signal passes through the trained network into the adaptive filter as a new reference signal.
Fig. 1 is a block diagram of an implementation of an adaptive filter system, where a receiver receiving signal includes a target signal, a noise signal, and a self-interference signal, and the sum of these signals is a signal before we cancel, as shown in equation (5), where the self-interference signal si (t) is the target to be cancelled.
4a) Taking LFM signal as an example, the received signal is denoised and sampled to become y (n), which is a cancellation signal input to the adaptive filter and has a duration of 3 μ s as shown in the figure,
4b) and (3) delaying the signal y (n) in the first step by 0.5 mu s, then carrying out nonlinear processing by using a Sa-Hz model in the step 2 to simulate the nonlinear characteristic of the power amplifier, and adding SI (n) and y (n) as well as white Gaussian noise as a target cancellation signal after the nonlinear processing is taken as a self-interference signal SI (n).
4c) Inputting the y (n) signal delayed by 0.5 mu s in the second step into DNN neural network as training sample, using the signal after nonlinear processing by the Saichh model as label, and estimating the signal x by the deep neural networkAPAnd (n) as a reference signal of the adaptive filter.
4d) The signals generated in steps 4b and 4c are input into the adaptive filter of the LMS algorithm to obtain an output signal x (n) after cancellation by the adaptive filter, and the result of the cancellation by the LFM signal is shown in fig. 6.
Similarly, the LFM signal in the above steps is changed into a BPSK signal, and the cancellation result is shown in fig. 7.
And 5: compared with the traditional self-adaptive method.
Fig. 6 is a cancellation result of the LFM signal using the new method, fig. 8 is a cancellation result of the existing LMS algorithm, and it is obvious by comparison that the DNN-based method has a better cancellation effect on the self-interference signal. Compared with the traditional self-adaptive algorithm, the new method is adopted to eliminate the self-interference signal, and the BPSK signal and the LFM signal have similar cancellation results.
Claims (1)
1. A self-adaptive cancellation method based on a deep neural network is characterized by comprising the following steps:
step 1: defining a model of the signal received by the receiving antenna, including the power P of the transmitted signalfNonlinear distortion function G [. for power amplifier]And carrier center frequency fc；
The received signals of the receiving antenna comprise target signals, noise signals and self-interference signals; is just goingIn normal conditions, the noise signal is usually white Gaussian noise with zero mean, denoted by n (t), and has a power Pn(ω), i.e.
Defining the expected target signal r (t) received by the receiving antenna as:
wherein, PfIs the power of the radar transmitted signal, G [. cndot.)]Representing the nonlinear distortion function of the power amplifier, df(t) represents a modulated baseband waveform, fcRepresents the center frequency of the carrier;
the self-interference signal in the radar jammer is a delay of the transmitted signal, so the self-interference signal x (t) can be set as follows:
x(t)＝r(t-τ1)+r(t-τ2)+······+r(t-τn) (3)
wherein, tau1、τ2And τnIs to simulate the interference delay in the actual situation;
after the target signal is modulated by baseband, dn(t) is obtained from x (t), xPA(t) is obtained by power amplification; and finally, the transmitting antenna sends out the self-interference signal SI (t) formed by the receiving antenna, wherein the transmitting signal is inevitably input:
Pnpower of the signal transmitted by the jammer, fcRepresents the center frequency of the carrier, phi represents the initial phase of the carrier, and k represents a known coefficient;
therefore, the model of the signal y (t) actually received by the receiving antenna is:
y(t)＝r(t)+SI(t)+n(t) (5)
step 2: defining a model of a non-linear power amplifier;
the reason for causing the nonlinear distortion of the power amplifier is mainly AM/AM distortion, wherein the AM/AM distortion refers to the distortion of the amplitude of an output signal caused by the amplitude change of an input signal; the adopted power amplifier is a traveling wave tube amplifier, the nonlinear distortion of the traveling wave tube amplifier can be described by a saliche model, and the AM/AM and AM/PM of the saliche model have the characteristic functions as follows:
where r is the amplitude of the input signal; alpha is alphaa、βa、αφAnd betaφIs a model parameter; obtaining a proper fixed model by adjusting the four parameters; nonlinear characteristics of the salich model include amplitude and phase changes of the signal after passing through the salich model;
taking the derivative of equation (6) to obtain the input signal amplitude when the input signal reaches the input saturation amplitude of the amplifier model:
the model obtains the maximum output signal amplitude:
f(A)max＝αa Asat/2 (9)
the maximum output value of the power amplifier determines the maximum value that the linearization can correct; if the linear output value corresponding to the input amplitude value of the power amplifier is larger than the maximum output amplitude value of the power amplifier, the nonlinear distortion cannot be compensated;
and step 3: carrying out nonlinear modeling on a target signal, and training a DNN network by using a large amount of data;
3a) the training data takes the form of two signals, namely a Linear Frequency Modulation (LFM) signal and a BPSK signal, as target signals respectively; for these two signals, two data sets were made, each containing 10000 samples; each LFM sample is a pulse signal with the pulse width of 3 mu s, each BPSK sample comprises 13 symbols, and the number of sampling points of each symbol at the sampling frequency is 70; for the LFM signal, the LFM signal sampling frequency f is defineds300MHz, carrier frequency 100MHz, bandwidth 20MHz, signal amplitude 20; defining the sampling frequency of BPSK signals as fs300MHz, amplitude 20, carrier frequency 50 MHz;
3b) the target signal generated in 3a is a label of the training sample obtained by the salich model in step 2 (the saturation output amplitude of the model is 80); wherein the parameters of the saleach model are set as followsa＝2，αφ＝π/3，βa＝1.5625e-5，βφ1 is ═ 1; generating a target signal as a training signal of the DNN network, and taking a signal passing through a Saichh model as a label of training data; after nonlinear amplification, the data are obviously changed in both time domain and frequency domain;
3c) the input layer of the DNN network is x (t), the number of nodes of the hidden layer is respectively amplified and reduced, the final number of nodes of the output layer is the same as that of the input layer, and a nonlinear amplified output signal x is obtainedAP(t); in the experiment, the hidden layers use Relu (transformed linear units) activation functions, and the number of hidden layers of the deep neural network is Num, 1024, 2048, …, 1024, Num; wherein Num is the number of points of each sample, and the specific layer number can be adjusted; the loss quadratic cost function in a neural network is:
Loss＝mean(square(x-xAP)) (10)
the output of each node of the DNN network is a nonlinear activation function to which its inputs are applied; weights between layers in the neural network are optimized through extensive learning, and expected outputs of training samples containing known inputs are learned;
and 4, step 4: inputting a signal generated after the original reference signal passes through a trained network as a new reference signal into an adaptive filter;
the receiver receives signals including a target signal, a noise signal and a self-interference signal, the sum of the signals is a signal before cancellation, and a self-interference signal SI (t) in formula (5) is a target to be cancelled;
4a) taking LFM signal as an example, the received signal is denoised and sampled to become y (n), which is the cancellation signal input to the adaptive filter and has a duration of 3 μ s,
4b) delaying the y (n) signal in the first step by 0.5 mu s, then carrying out nonlinear processing by using the Sa-Hz model in the step 2 to simulate the nonlinear characteristic of the power amplifier, and taking the processed signal as a self-interference signal SI (n), and adding SI (n), y (n) and white Gaussian noise as a target cancellation signal;
4c) inputting the y (n) signal delayed by 0.5 mu s in the second step into DNN neural network as training sample, using the signal after nonlinear processing by the Saichh model as label, and estimating the signal x by the deep neural networkAP(n) as a reference signal for the adaptive filter;
4d) inputting the signals generated in the steps 4b and 4c into the adaptive filter of the LMS algorithm to obtain an output signal x (n) after cancellation of the adaptive filter, and achieving the result of cancellation of the LFM signal;
the cancellation result can also be achieved by changing the LFM signal in the above steps into a BPSK signal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110569844.8A CN113325375B (en) | 2021-05-25 | 2021-05-25 | Self-adaptive cancellation method based on deep neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110569844.8A CN113325375B (en) | 2021-05-25 | 2021-05-25 | Self-adaptive cancellation method based on deep neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113325375A true CN113325375A (en) | 2021-08-31 |
| CN113325375B CN113325375B (en) | 2022-12-13 |
Family
ID=77416659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110569844.8A Active CN113325375B (en) | 2021-05-25 | 2021-05-25 | Self-adaptive cancellation method based on deep neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113325375B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114221667A (en) * | 2021-12-08 | 2022-03-22 | 哈尔滨工程大学 | Method and system for eliminating known signals at receiving end of communication system |
| CN115061099A (en) * | 2022-07-28 | 2022-09-16 | 南京华成微波技术有限公司 | Method and device for radar non-stationary team following interference cancellation and terminal equipment |
| CN115842566A (en) * | 2022-11-28 | 2023-03-24 | 哈尔滨工程大学 | CNN-Bi-LSTM-based interference machine self-interference digital cancellation method |
Citations (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101527584A (en) * | 2009-03-16 | 2009-09-09 | 哈尔滨工程大学 | Spread-spectrum pseudo code detection method based on frequency domain amplitude processing |
| US20150043323A1 (en) * | 2013-08-09 | 2015-02-12 | Kumu Networks, Inc. | Systems and methods for non-linear digital self-interference cancellation |
| CN104678365A (en) * | 2015-03-20 | 2015-06-03 | 河海大学 | Adaptive interference cancellation method for radar jammer |
| US20160071003A1 (en) * | 2014-09-10 | 2016-03-10 | Qualcomm Incorporated | Multilayer Perceptron for Dual SIM Dual Active Interference Cancellation |
| WO2016040590A1 (en) * | 2014-09-10 | 2016-03-17 | Qualcomm Incorporated | Methods and systems for multi-model, multi-layer perceptron based non-linear interference management in multi-technology communication devices |
| CN108599809A (en) * | 2018-03-14 | 2018-09-28 | 中国信息通信研究院 | Full duplex self-interference signal number removing method and device |
| CN109412993A (en) * | 2018-11-27 | 2019-03-01 | 天津理工大学 | A kind of having based on deep learning inhibits the full duplex cognition water sound communication signal detection method of self-interference |
| CN109871882A (en) * | 2019-01-24 | 2019-06-11 | 重庆邮电大学 | Method of EEG signals classification based on Gauss Bernoulli convolution depth confidence network |
| CN109921822A (en) * | 2019-02-19 | 2019-06-21 | 哈尔滨工程大学 | The method that non-linear, digital self-interference based on deep learning is eliminated |
| CN110190920A (en) * | 2019-06-10 | 2019-08-30 | 中国人民解放军陆军工程大学 | The reference signal extracting method and system of algorithm are offseted based on least mean-square error |
| US10581469B1 (en) * | 2017-04-17 | 2020-03-03 | DeepSig Inc. | Machine learning-based nonlinear pre-distortion system |
| CN111200470A (en) * | 2020-01-10 | 2020-05-26 | 东南大学 | High-order modulation signal transmission control method suitable for being interfered by nonlinearity |
| CN111490737A (en) * | 2019-01-28 | 2020-08-04 | ***通信有限公司研究院 | Nonlinear compensation method and device for power amplifier |
| US20200292660A1 (en) * | 2019-03-14 | 2020-09-17 | Infineon Technologies Ag | Fmcw radar with interference signal suppression using artificial neural network |
| CN111884602A (en) * | 2020-06-29 | 2020-11-03 | 北京邮电大学 | Power amplifier predistortion method based on single-output-node neural network |
| CN111983581A (en) * | 2020-08-26 | 2020-11-24 | 中国人民解放军军事科学院国防科技创新研究院 | Cognitive radar system, waveform generation method and device thereof and readable storage medium |
| CN112034429A (en) * | 2020-01-06 | 2020-12-04 | 中国航天科工集团八五一一研究所 | Self-adaptive digital cancellation method for eliminating interference self-excitation |
| CN112053676A (en) * | 2020-08-07 | 2020-12-08 | 南京时保联信息科技有限公司 | Nonlinear adaptive active noise reduction system and noise reduction method thereof |
| CN112383322A (en) * | 2020-10-23 | 2021-02-19 | 北京大学 | Regularization-based full-duplex system combined self-interference elimination method and electronic device |
| CN112491442A (en) * | 2020-11-17 | 2021-03-12 | 中山大学 | Self-interference elimination method and device |
-
2021
- 2021-05-25 CN CN202110569844.8A patent/CN113325375B/en active Active
Patent Citations (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101527584A (en) * | 2009-03-16 | 2009-09-09 | 哈尔滨工程大学 | Spread-spectrum pseudo code detection method based on frequency domain amplitude processing |
| US20150043323A1 (en) * | 2013-08-09 | 2015-02-12 | Kumu Networks, Inc. | Systems and methods for non-linear digital self-interference cancellation |
| CN105556860A (en) * | 2013-08-09 | 2016-05-04 | 库姆网络公司 | Systems and methods for non-linear digital self-interference cancellation |
| US20160071003A1 (en) * | 2014-09-10 | 2016-03-10 | Qualcomm Incorporated | Multilayer Perceptron for Dual SIM Dual Active Interference Cancellation |
| WO2016040590A1 (en) * | 2014-09-10 | 2016-03-17 | Qualcomm Incorporated | Methods and systems for multi-model, multi-layer perceptron based non-linear interference management in multi-technology communication devices |
| CN104678365A (en) * | 2015-03-20 | 2015-06-03 | 河海大学 | Adaptive interference cancellation method for radar jammer |
| US10581469B1 (en) * | 2017-04-17 | 2020-03-03 | DeepSig Inc. | Machine learning-based nonlinear pre-distortion system |
| CN108599809A (en) * | 2018-03-14 | 2018-09-28 | 中国信息通信研究院 | Full duplex self-interference signal number removing method and device |
| CN109412993A (en) * | 2018-11-27 | 2019-03-01 | 天津理工大学 | A kind of having based on deep learning inhibits the full duplex cognition water sound communication signal detection method of self-interference |
| CN109871882A (en) * | 2019-01-24 | 2019-06-11 | 重庆邮电大学 | Method of EEG signals classification based on Gauss Bernoulli convolution depth confidence network |
| CN111490737A (en) * | 2019-01-28 | 2020-08-04 | ***通信有限公司研究院 | Nonlinear compensation method and device for power amplifier |
| CN109921822A (en) * | 2019-02-19 | 2019-06-21 | 哈尔滨工程大学 | The method that non-linear, digital self-interference based on deep learning is eliminated |
| US20200292660A1 (en) * | 2019-03-14 | 2020-09-17 | Infineon Technologies Ag | Fmcw radar with interference signal suppression using artificial neural network |
| CN110190920A (en) * | 2019-06-10 | 2019-08-30 | 中国人民解放军陆军工程大学 | The reference signal extracting method and system of algorithm are offseted based on least mean-square error |
| CN112034429A (en) * | 2020-01-06 | 2020-12-04 | 中国航天科工集团八五一一研究所 | Self-adaptive digital cancellation method for eliminating interference self-excitation |
| CN111200470A (en) * | 2020-01-10 | 2020-05-26 | 东南大学 | High-order modulation signal transmission control method suitable for being interfered by nonlinearity |
| CN111884602A (en) * | 2020-06-29 | 2020-11-03 | 北京邮电大学 | Power amplifier predistortion method based on single-output-node neural network |
| CN112053676A (en) * | 2020-08-07 | 2020-12-08 | 南京时保联信息科技有限公司 | Nonlinear adaptive active noise reduction system and noise reduction method thereof |
| CN111983581A (en) * | 2020-08-26 | 2020-11-24 | 中国人民解放军军事科学院国防科技创新研究院 | Cognitive radar system, waveform generation method and device thereof and readable storage medium |
| CN112383322A (en) * | 2020-10-23 | 2021-02-19 | 北京大学 | Regularization-based full-duplex system combined self-interference elimination method and electronic device |
| CN112491442A (en) * | 2020-11-17 | 2021-03-12 | 中山大学 | Self-interference elimination method and device |
Non-Patent Citations (9)
| Title |
|---|
| XIAOYU LI: "Self-Interference Cancellation in Radar Jammer Based on Deep Neural Networks", 《ICDSP 2020: PROCEEDINGS OF THE 2020 4TH INTERNATIONAL CONFERENCE ON DIGITAL SIGNAL PROCESSING》 * |
| YITAO CHEN: "MIMO Full Duplex Radios with Deep Learning", 《2020 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS WORKSHOPS (ICC WORKSHOPS). PROCEEDINGS》 * |
| 侯爱霞: "OFDM-SDMA***中非线性功率放大器的影响", 《智能机器人》 * |
| 唐成凯等: "卫星星载后置非线性失真消除算法", 《西安电子科技大学学报》 * |
| 杨家兴等: "神经网络自适应滤波器", 《信息工程大学学报》 * |
| 王少丹: "收发同时DRFM干扰机中自干扰对消技术的研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 王玉军等: "一种LFM雷达回波对消干扰算法", 《西安电子科技大学学报》 * |
| 耿敏: "RLS算法在同时同频全双工***中的自干扰抑制性能分析", 《成都信息工程大学通信工程学院》 * |
| 蒋威等: "基于量子神经网络的无线电引信干扰对消", 《微计算机信息》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114221667A (en) * | 2021-12-08 | 2022-03-22 | 哈尔滨工程大学 | Method and system for eliminating known signals at receiving end of communication system |
| CN114221667B (en) * | 2021-12-08 | 2023-08-18 | 哈尔滨工程大学 | Method and system for eliminating known signal of receiving end of communication system |
| CN115061099A (en) * | 2022-07-28 | 2022-09-16 | 南京华成微波技术有限公司 | Method and device for radar non-stationary team following interference cancellation and terminal equipment |
| CN115842566A (en) * | 2022-11-28 | 2023-03-24 | 哈尔滨工程大学 | CNN-Bi-LSTM-based interference machine self-interference digital cancellation method |
| CN115842566B (en) * | 2022-11-28 | 2023-09-15 | 哈尔滨工程大学 | CNN-Bi-LSTM-based self-interference digital cancellation method for jammers |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113325375B (en) | 2022-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113325375B (en) | Self-adaptive cancellation method based on deep neural network | |
| Duarte et al. | Experiment-driven characterization of full-duplex wireless systems | |
| CN109450486B (en) | Digital self-interference cancellation method for asynchronous simultaneous same-frequency full-duplex underwater acoustic communication system | |
| Qiao et al. | Self-interference channel estimation algorithm based on maximum-likelihood estimator in in-band full-duplex underwater acoustic communication system | |
| Komatsu et al. | Basis function selection of frequency-domain Hammerstein self-interference canceller for in-band full-duplex wireless communications | |
| Erdem et al. | Integrated linear and nonlinear digital cancellation for full duplex communication | |
| Shi et al. | Digital self-interference cancellation for full duplex wireless communication based on neural networks | |
| CN106685867A (en) | Upstream noise inhibiting method automatically adapting to noise thresholds in LTE-distributed relay system | |
| CN113726350A (en) | Deep neural network-based strong correlation self-interference cancellation method | |
| Muranov et al. | On deep learning assisted self-interference estimation in a full-duplex relay link | |
| He et al. | Frequency-domain successive cancellation of nonlinear self-interference with reduced complexity for full-duplex radios | |
| Ploder et al. | An adaptive machine learning based approach for the cancellation of second-order-intermodulation distortions in 4G/5G transceivers | |
| CN102761673B (en) | Method for multichannel echo cancellation of LTE (Long Term Evolution) relay system | |
| CN101283510B (en) | Interference canceling matched filter (ICMF) and related methods | |
| Balatsoukas-Stimming | Joint detection and self-interference cancellation in full-duplex systems using machine learning | |
| Erdem et al. | Nonlinear digital self-interference cancellation for full duplex communication | |
| CN113872898A (en) | Method and system for digital domain self-interference suppression based on bounded component analysis | |
| Kurt et al. | An adaptive-iterative nonlinear interference cancellation in time-varying full-duplex channels | |
| Ploder et al. | A neural network approach for the cancellation of the second-order-intermodulation distortion in future cellular RF transceivers | |
| He et al. | Denoising CNN based channel estimation for vehicular OTFS communication system | |
| Li et al. | Self-interference cancellation in radar jammer based on deep neural networks | |
| CN111181891A (en) | Improved power line impulse noise suppression method based on compressed sensing | |
| Wang et al. | Neural-network-based nonlinear self-interference cancelation scheme for mobile stations with dual-connectivity | |
| CN115001913A (en) | Full-duplex frequency domain self-interference elimination method based on digital assistance | |
| CN113747467A (en) | Design method of synchronous and asynchronous NOMA system of nonlinear waveform |
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 |