Satellite signal nonlinear distortion compensation method based on neural network

Abstract

The invention provides a satellite signal nonlinear distortion compensation method based on a neural network, and the method comprises the steps: building a satellite signal nonlinear distortion model, and building a post-distortion compensation model for a zoom factor in the model; the method comprises the steps that a distortion compensation model is obtained, a BPNN is adopted to estimate a scaling factor of the distortion compensation model, a particle swarm method is introduced to optimize the initial weight and bias of the BPNN, a satellite signal training sample is generated to train the BPNN, the trained network is adopted to generate the scaling factor of the distortion compensation model, and then satellite signal nonlinear distortion compensation is achieved. By the adoption of the method, complex mathematical derivation and hypothesis are not needed, and the influence of nonlinear distortion on the system is reduced under the condition that the complexity is small.

Description

Technical Field

[0001] The invention belongs to the field of communication signal detection, and in particular relates to a satellite signal nonlinear distortion compensation method based on a neural network. Background Art

[0002] At present, the commonly used methods for compensating nonlinear distortion include pre-distortion of the transmitter and post-compensation of the receiver. At the transmitting end, signal pre-distortion can be used to linearize the characteristics of the amplifier, involving related analog electronic devices to achieve the inverse of the amplifier characteristics, but signal pre-distortion will introduce out-of-band extension. This method is not used in satellite communications due to the strong nonlinearity of the traveling wave tube amplifier TWTA. On the other hand, digital pre-distortion is a pure digital technology that retains the signal spectrum. By modifying the transmit constellation, the centroid of the reception coincides with the original transmit constellation. However, pre-distortion technology has strict requirements on the size and energy consumption of the transmitter. Therefore, in the research on compensation of nonlinear distortion of satellite return links, post-distortion compensation technology at the receiver end is often used.

[0003] In the research on nonlinear distortion solutions at the receiver end, there are many studies that use iterative algorithms to solve the nonlinear distortion problem of the system. However, they all require complex calculations and assumptions as a prerequisite, which leads to problems such as slow calculation speed and low calculation accuracy. Summary of the invention

[0004] Purpose of the invention: In order to solve the problems existing in the above-mentioned prior art, the present invention provides a satellite signal nonlinear distortion compensation method based on neural network.

[0005] Technical solution: The present invention provides a method for compensating satellite signal nonlinear distortion based on a neural network, which specifically includes the following steps:

[0006] Step 1: Establish satellite signal nonlinear distortion model A1;

[0007] Step 2: Establish a post-distortion compensation model A2 based on the scaling factor in model A1;

[0008] Step 3: Establish a back propagation neural network BPNN, and use the particle swarm optimization algorithm to calculate the initial weights and biases of the BPNN network;

[0009] Step 4: Solve model A2 to obtain the optimal scaling factor corresponding to the nonlinear distortion signal, and iteratively train the BPNN network;

[0010] Step 5: Use the scaling factor corresponding to the trained BPNN network output to perform nonlinear distortion compensation of the satellite signal based on the scaling factor.

[0011] Furthermore, the satellite signal nonlinear distortion model A1 is expressed as follows:

[0012] z(t)=ku(t)

[0013] Where k is the scaling factor of the signal u(t) at the transmitter, t is the time, Where j is an imaginary unit, r(t) represents the amplitude of the signal u(t), The expressions for the phase of the signal u(t), Α(r(t)), Φ(r(t)) and u(t) are as follows:

[0014]

[0015] Among them, α α , β α , and are all parameters, x n is the nth sampled signal at the transmitter, p(t) is the impulse response function of the pulse shaping filter, T s represents the sampling period, and N is the total number of sampled transmitter signals.

[0016] Furthermore, the expression of the post-distortion compensation model A2 is:

[0017]

[0018] Where K = [k0, k1, …, k n ,…,k N-1 ] T is the nonlinear distortion complex scaling factor vector, where k n For x n The corresponding nonlinear distortion complex scaling factor; x is the complex time domain signal sequence, x = [x0, x1, ..., x n ,…,x N-1 ] T ,m=[m0,m1,…,m n ,…,m N-1 ] T represents the signal vector after nonlinear distortion compensation, m n is the signal after the nth nonlinear distortion compensation, m n Expressed as:

[0019]

[0020] Among them, y n is the nth signal sampled at the receiver.

[0021] Furthermore, the step five is specifically as follows: inputting the signal of the receiving end into the trained neural network, the trained neural network outputs the corresponding nonlinear distortion complex scaling factor, and according to the formula Perform nonlinear distortion compensation on the signal received by the receiving end.

[0022] Furthermore, the particle swarm optimization algorithm is used in step 3 to calculate the initial weights and biases of the BPNN network as follows:

[0023] Step 3.1: Initialize the initial position P of a population of l particles = [p1, p2, ..., p i …,p l ], the position of the i-th particle is represented by p i =(p i1 ,p i2 ,…,p iD ), p iD The D-th dimension data representing the position of the ith particle, D = D w +D d , D w is the number of weights of the neural network, D w is the number of biases of the neural network; initialize the global optimal value and the individual optimal value;

[0024] Step 3.2: Update the particle velocity at the current iteration;

[0025] Step 3.3: Update the particle position according to the updated particle velocity;

[0026] Step 3.4: Calculate the fitness of the particle at the current iteration and update the individual optimal value and the global optimal value; if the number of iterations reaches the maximum number of iterations, output the optimal particle position information to obtain the initial weight and bias of the neural network; otherwise, proceed to the next iteration; go to step 3.2.

[0027] Furthermore, the particle speed is updated in step 3.2 as follows: first, the particle speed v at the current tpth iteration is calculated. id,tp , if v id,tp Greater than the preset maximum speed v max , then v id,tp =v max , if v id,tp Less than the preset minimum speed w min , then v id,tp =v min , i represents the i-th particle, d represents the dimension, d = 1, 2,…, D.

[0028] Beneficial effects: The present invention is based on data drive, does not require complicated mathematical derivation and assumptions, and reduces the impact of nonlinear distortion on the system with less complexity. DETAILED DESCRIPTION

[0029] The exemplary embodiments of the present invention and their description are used to explain the present invention and do not constitute improper limitations of the present invention.

[0030] In order to better illustrate the method of the present invention, it is described below in conjunction with more detailed examples:

[0031] In a satellite return link, the nonlinear distortion caused by the traveling wave tube amplifier (TWTA) on the satellite is considered and compensated at the receiver end.

[0032] A method for compensating satellite signal nonlinear distortion based on a neural network comprises the following steps:

[0033] Step A: Establish satellite signal nonlinear distortion model A1, model A1 is:

[0034] z(t)=ku(t)

[0035] in,

[0036]

[0037] Among them, r(t) represents the amplitude of the signal u(t), The expressions for the phase of the signal u(t), Α(r(t)) and Φ(r(t)) are:

[0038]

[0039] Among them, α α , β α , and are the parameters of the Saleh model, which are set to 2.1587, 1.1517, 4.0033 and 9.104 in this embodiment.

[0040] u(t) is the transmitting end signal, and its expression is:

[0041]

[0042] Among them, x n is the nth sampling symbol (that is, the nth sampling signal, which is the sampling of the transmitting end signal), N is the symbol sequence length, and the symbol period is T s , p(t) is the impulse response function of the pulse shaping filter; k is a complex scaling factor related to the input signal amplitude; in this embodiment, the nonlinear distortion complex scaling factor corresponding to the i-th symbol is k i .

[0043] Step B: Establish the post-distortion compensation model A2, which is:

[0044]

[0045] Among them, ||·|| is the 2-norm, which is the square root of the sum of the squares of the absolute values ​​of the vector elements. In order to make ||mx|| 2 The K with the smallest value; expand the above model in detail and get:

[0046]

[0047] Where K = [k0, k1, …, k i ,…,k N-1 ] T is the nonlinear distortion complex scaling factor vector corresponding to the symbol sequence, k n is the nonlinear distortion complex scaling factor corresponding to the nth symbol, x is the complex time domain signal sequence, x=[x0,x1,…,x n ,…,x N-1 ] T ,m=[m0,m1,…,m n ,…,m N-1 ] T represents the symbol vector after nonlinear distortion compensation, m n is the symbol after the nth nonlinear distortion compensation, m n Expressed as:

[0048]

[0049] Among them, y n It is the nth symbol after sampling at the receiver end (that is, the nth signal at the sampled receiver end).

[0050] Step C: Establish a back propagation neural network BPNN. The input layer of the network sets two input nodes, corresponding to the real and imaginary parts of the complex received signal. The output layer sets two output nodes, corresponding to the real and imaginary parts of the complex scaling factor k. The particle swarm algorithm is used to optimize the initial weights and biases of the neural network. The specific process is as follows:

[0051] Step C-1: Initialize in a D-dimensional search space, D = D w +D d , D w is the number of weights of the neural network, D w is the number of biases of the neural network; the initial position of the population consisting of l particles P = [p1, p2, ..., p l ], the position of the i-th particle is represented by p i =(p i1 ,p i2 ,…,p iD ), p iDis the D-th dimension data of the position of the i-th particle, and initializes the global optimal value G best =inf, inf is infinite, initialize the individual optimal value G id =inf.

[0052] Step C-2: Iteratively update the particle position and velocity. The particle velocity update formula is:

[0053] v i,tp =w p ×v i,tp-1 +c1×ψ1×(G i,tp-1 -p i,tp-1 )+c2×ψ2×(G best,tp -p i,tp-1 )

[0054] Where i represents the i-th particle, tp represents the tp-th iteration, and w p is the inertia weight coefficient, c1 and c2 are population learning factors, ψ1 and ψ2 are random numbers on [0,1], and w p The expression is:

[0055]

[0056] T max is the maximum number of iterations, w min and w max are the minimum inertia factor and the maximum inertia factor respectively.

[0057] Speed ​​update: If v id,tp >v max , then v id,tp =v max ; if v id,tp <v min , then v id,tp =v min ; where d represents the dimension, d = 1, 2, ..., D, v id,tp is the vector v i,tp The dth element in .

[0058] The particle position update formula is:

[0059] p i,tp =p i,tp-1 +v i,tp ,

[0060] Calculate the fitness of the current particle and update the individual optimal value and the global optimal value; when the number of iterations reaches the maximum number of iterations, output the optimal particle position information to obtain the initial weight and bias of the neural network; otherwise, proceed to the next iteration.

[0061] Step C-3: Iteratively train the BPNN network; calculate the mean square error E t , update the network weights and biases; when E t When the error is less than the target error or the number of iterations reaches the maximum, the iteration ends.

[0062] Step D: Detect satellite signal distortion and perform satellite signal distortion compensation. The specific steps are as follows:

[0063] Step D-1: Use the trained network and input the nonlinear distortion symbol y at the receiving end to be detected n , the network output is the corresponding nonlinear distortion complex scaling factor;

[0064] Step D-2: According to Compensation is performed on each received symbol.

[0065] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

Claims

1. A method for compensating satellite signal nonlinear distortion based on neural network, characterized in that: The specific steps include: Step 1: Establish satellite signal nonlinear distortion model A1; Step 2: Establish a post-distortion compensation model A2 based on the scaling factor in model A1; Step 3: Establish a back propagation neural network BPNN, and use the particle swarm optimization algorithm to calculate the initial weights and biases of the BPNN network; Step 4: Solve model A2 to obtain the optimal scaling factor corresponding to the nonlinear distortion signal, and iteratively train the BPNN network; Step 5: Use the scaling factor corresponding to the trained BPNN network output to perform nonlinear distortion compensation of the satellite signal based on the scaling factor.

2. The method for compensating satellite signal nonlinear distortion based on a neural network according to claim 1, characterized in that: The expression of the satellite signal nonlinear distortion model A1 is as follows: z(t)=ku(t) Where k is the scaling factor of the signal u(t) at the transmitter, t is the time, Where j is an imaginary unit, r(t) represents the amplitude of the signal u(t), The expressions for the phase of the signal u(t), Α(r(t)), Φ(r(t)) and u(t) are as follows: Among them, α α , β α , and are all parameters, x n is the nth sampled signal at the transmitter, p(t) is the impulse response function of the pulse shaping filter, T s represents the sampling period, and N is the total number of sampled transmitter signals.

3. The method for compensating satellite signal nonlinear distortion based on a neural network according to claim 2, characterized in that: The expression of the post-distortion compensation model A2 is: Where K = [k0, k1, …, k n ,…,k N-1 ] T is the nonlinear distortion complex scaling factor vector, where k n For x n The corresponding nonlinear distortion complex scaling factor; x is the complex time domain signal sequence, x = [x0, x1, ..., x n ,…,x N-1 ] T ,m=[m0,m1,…,m n ,…,m N-1 ] T represents the signal vector after nonlinear distortion compensation, m n is the signal after the nth nonlinear distortion compensation, m n Expressed as: Among them, y n is the nth signal sampled at the receiver.

4. The method for compensating satellite signal nonlinear distortion based on a neural network according to claim 3, characterized in that: The step five is specifically as follows: inputting the signal of the receiving end into the trained neural network, the trained neural network outputs the corresponding nonlinear distortion complex scaling factor, and according to the formula Perform nonlinear distortion compensation on the signal received by the receiving end.

5. The method for compensating satellite signal nonlinear distortion based on neural network according to claim 1, characterized in that: In step 3, the particle swarm optimization algorithm is used to calculate the initial weights and biases of the BPNN network as follows: Step 3.1: Initialize the initial position P of a population of l particles = [p1, p2, ..., p i …,p l ], the position of the i-th particle is represented by p i =(p i1 ,p i2 ,…,p iD ), p iD The D-th dimension data representing the position of the ith particle, D = D w +D d , D w is the number of weights of the neural network, D w is the number of biases of the neural network; initialize the global optimal value and the individual optimal value; Step 3.2: Update the particle velocity at the current iteration; Step 3.3: Update the particle position according to the updated particle velocity; Step 3.4: Calculate the fitness of the particle at the current iteration and update the individual optimal value and the global optimal value; if the number of iterations reaches the maximum number of iterations, output the optimal particle position information to obtain the initial weight and bias of the neural network; otherwise, proceed to the next iteration; go to step 3.

2.

6. The method for compensating satellite signal nonlinear distortion based on neural network according to claim 5, characterized in that: The specific updating of particle velocity in step 3.2 is as follows: firstly, the velocity v of the particle at the current tpth iteration is calculated. id,tp , if v id,tp Greater than the preset maximum speed v max , then v id,tp =v max , if v id,tp Less than the preset minimum speed w min , then v id,tp =v min , i represents the i-th particle, d represents the dimension, d = 1, 2,…, D.

Images