Echo elimination method suitable for aircraft communication

Abstract

The invention discloses an echo cancellation method suitable for aircraft communication. The echo cancellation method comprises the following steps: S1, constructing a signal receiving model of an aircraft during signal receiving; s2, setting a weight vector w (n) belonging to RM * 1 to represent an estimated value of wo at an instantaneous moment n, and initializing elements of w (n); s3, calculating an error vector e (n); updating the weight vector w (n) according to the error vector e (n); s4, enabling n to be n + 1, taking the updated weight vector as a new w (n), and then repeatedly executing the steps to obtain an identification result of the echo interference channel; and S5, echo cancellation is realized in the actual signal receiving process of the aircraft. According to the method, echo noise of aircraft communication can be effectively eliminated through identification of the echo interference channel, abnormal values can be effectively eliminated through error signal judgment factors during identification of the echo channel, and the convergence speed is improved.

Description

Technical Field

[0001] The present invention relates to aircraft communication, and in particular to an echo elimination method suitable for aircraft communication. Background Art

[0002] During the communication process of drones / aircraft, the echo signal of their own transmitted signal will be received, which affects the quality of communication. Therefore, the echo elimination method can be used to filter out the echo signal, thereby reducing the impact of the echo on drone / aircraft communication. In recent years, the most commonly used adaptive filtering algorithms are the least mean square algorithm (LMS) and the normalized least mean square algorithm (NLMS). The advantage of these two algorithms is that they are easy to implement. Later, in order to improve the anti-noise ability, some other algorithms emerged, such as the adaptive filtering algorithm based on the activation function, the robust mixed norm algorithm (RMNA) and the minimum mean logarithm square algorithm (LMLS). However, the above algorithms converge slowly when faced with complex inputs.

[0003] Ozeki et al. proposed the Affine Projection Algorithm (APA), which solves this problem by combining multiple input vectors to construct an affine orthogonal subspace. However, the APA algorithm is still severely affected by impulse noise. To overcome this shortcoming, the Affine Projection Sign Algorithm (APSA) enhances the algorithm's ability to resist impulse noise by using the vector 1 norm as the cost function. However, when complex input and impulse noise exist at the same time, the APA and APSA algorithms cannot cope well. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an echo elimination method suitable for aircraft communication, which can effectively eliminate the echo noise of aircraft communication by identifying the echo interference channel, and when identifying the echo channel, it can effectively eliminate outliers and improve the convergence speed through error signal judgment factors.

[0005] The object of the present invention is achieved through the following technical solution: an echo elimination method suitable for aircraft communication, comprising the following steps:

[0006] S1. Construct a signal receiving model when the aircraft is receiving signals:

[0007] d(n)=w o T u(n)+v(n);

[0008] Among them, d(n) represents the received signal of the aircraft, and u(n) represents the transmitted signal of aircraft n at the instant, which is a column vector of length M:

[0009] u(n)=[u(n),u(n-1),...,u(n-M+1)] T ∈R M×1

[0010] The superscript T represents the transpose of a matrix or vector, v(n) represents the target signal that the aircraft needs to receive; w o ∈R M×1 is the column vector of unknown weight coefficients to be estimated, representing the echo interference channel;

[0011] S2. Let the weight vector w(n)∈R M×1 Indicates w o Initialize the elements of w(n) at the estimated value at instantaneous time n so that all elements of the weight matrix w(n) are 0. The weights need to be made to approximate the echo interference channel to achieve the identification of the echo interference channel.

[0012] S3. Construct the input matrix U(n) and output matrix d(n) of the signal reception model, calculate the error vector e(n); according to the error vector e(n), update the weight vector w(n) at the instant n to obtain the weight vector w(n+1) at the instant n+1;

[0013] S4. Let n be n+1, take the updated weight vector as the new w(n), and then repeat step S3 until the error vector e(n) is less than the set threshold, it is considered that the convergence condition is met, and the weight vector at this time is recorded as w′ as the identification result of the echo interference channel;

[0014] S5. In the process of the aircraft actually receiving signals, the signal emitted by the aircraft itself is first multiplied by the transposed result of the echo interference channel identification result w′ to calculate the echo signal, and then the aircraft subtracts the echo signal from the received signal to achieve echo cancellation.

[0015] The beneficial effects of the present invention are as follows: the present invention greatly improves the robustness in pulse noise by eliminating abnormal values ​​that do not meet the conditions in the error signal. At the same time, the computational complexity of the algorithm is low. Compared with related algorithms, the algorithm has a faster convergence speed and a smaller steady-state mismatch, has better performance, and can achieve more accurate echo elimination. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of the method of the present invention;

[0017] Figure 2 Schematic diagram of simulation results of this application. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0019] like Figure 1 As shown, an echo elimination method applicable to aircraft communication includes the following steps:

[0020] S1. Construct a signal receiving model when the aircraft is receiving signals:

[0021] d(n)=w o T u(n)+v(n);

[0022] Among them, d(n) represents the received signal of the aircraft, and u(n) represents the transmitted signal of aircraft n at the instant, which is a column vector of length M:

[0023] u(n)=[u(n),u(n-1),...,u(n-M+1)] T ∈R M×1

[0024] The superscript T represents the transpose of a matrix or vector, v(n) represents the target signal that the aircraft needs to receive; w o ∈R M×1 is the column vector of unknown weight coefficients to be estimated, representing the echo interference channel;

[0025] S2. Let the weight vector w(n)∈R M×1 Indicates w o Initialize the elements of w(n) at the estimated value at instantaneous time n so that all elements of the weight matrix w(n) are 0. The weights need to be made to approximate the echo interference channel to achieve the identification of the echo interference channel.

[0026] S3. Construct the input matrix U(n) and output matrix d(n) of the signal reception model, calculate the error vector e(n); according to the error vector e(n), update the weight vector w(n) at the instant n to obtain the weight vector w(n+1) at the instant n+1;

[0027] S4. Let n be n+1, take the updated weight vector as the new w(n), and then repeat step S3 until the error vector e(n) is less than the set threshold, it is considered that the convergence condition is met, and the weight vector at this time is recorded as w′ as the identification result of the echo interference channel;

[0028] S5. In the process of the aircraft actually receiving signals, the signal emitted by the aircraft itself is first multiplied by the transposed result of the echo interference channel identification result w′ to calculate the echo signal, and then the aircraft subtracts the echo signal from the received signal to achieve echo cancellation.

[0029] In an embodiment of the present application, in step S3, constructing an input matrix U(n) and an output matrix d(n) of a signal receiving model, and calculating an error vector e(n) includes:

[0030] According to the aircraft transmission signals at the current instant n and the previous L consecutive moments, the input matrix U(n) of the signal receiving model is constructed:

[0031] U(n)=[u(n),u(n-1),...,u(n-L+1)]∈R M×L ;

[0032] The received signals of the aircraft at the current instant n and the previous L consecutive moments are recorded as the output matrix d(n):

[0033] d(n)=[d(n),d(n-1),...,d(n-L+1)] T ∈R L×1

[0034] Multiply the input matrix U(n) by the weight vector w(n), and then calculate the error vector e(n) by subtracting the product of U(n) and w(n) from the output matrix d(n):

[0035] e(n)=d(n)-U T (n)w(n)∈R L×1 .

[0036] In step S3, the weight vector w(n) at instant n is updated according to the error vector e(n), and the weight vector w(n+1) at instant n+1 is obtained as follows:

[0037] w(n+1)=w(n)+μU(n)P(n)e(n)

[0038] Where, P(n)=diag[p(n),p(n-1),…,p(n-L+1)]; μ represents the step size, 0<μ<1;

[0039] Since e(n)∈R L×1 , the specific form of the error vector is expressed as:

[0040] e(n)=[e(n),e(n-1),...,e(n-L+1)] T ,

[0041] p(n) is calculated as follows:

[0042]

[0043] θ 1=med(e(n))-k·med(abs(e(n)-med(e(n))))

[0044] θ 2 =med(e(n))+k·med(abs(e(n)-med(e(n))))

[0045] p(n) is the judgment factor for the error signal. When the error signal e(n) does not satisfy θ 1 <e(n)<θ 2 When , p(n) is set to 0 to remove abnormal values, k is the preset adjustment factor and k>0, med represents the median operation, and abs represents the absolute value operation. Similarly, when n is n-1,…,n-L+1, the calculation of all elements in P(n) can be completed.

[0046] In the embodiment of the present application, through the first-order system The white zero-mean Gaussian signal with a power of 1 is filtered to obtain the color input signal. We set the signal-to-noise ratio (SNR) to 30 dB and add it to the output signal d(n) = w o T In u(n), we use the alpha-stable distribution to generate impulse noise.

[0047] The characteristics are in:

[0048] is a positional parameter, is a discrete parameter, is a characteristic factor, is the symmetry parameter.

[0049] make yes In all experiments, we set We use the normalized MSD (NMSD), i.e. To measure the performance of the algorithms, all simulation results are averaged over 1000 independent trials. We test SA, APSA and APRODA (the algorithms proposed in this application) from Figure 2 It can be seen that the filtering performance of APRODA of the present application is better.

[0050] The above is a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not depart from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.

Claims

1. An echo elimination method suitable for aircraft communication, characterized in that: The following steps are involved: S1. Construct a signal receiving model when the aircraft is receiving signals: d(n)=w o T u(n)+v(n); Among them, d(n) represents the received signal of the aircraft, and u(n) represents the transmitted signal of aircraft n at the instant, which is a column vector of length M: u(n)=[u(n),u(n-1),...,u(n-M+1)] T ∈R M×1 The superscript T represents the transpose of a matrix or vector, v(n) represents the target signal that the aircraft needs to receive; w o ∈R M×1 is the column vector of unknown weight coefficients to be estimated, representing the echo interference channel; S2. Let the weight vector w(n)∈R M×1 Indicates w o Initialize the elements of w(n) at the estimated value at instantaneous time n so that all elements of the weight matrix w(n) are 0. The weights need to be made to approximate the echo interference channel to achieve the identification of the echo interference channel. S3. Construct the input matrix U(n) and output matrix d(n) of the signal reception model, calculate the error vector e(n); according to the error vector e(n), update the weight vector w(n) at the instant n to obtain the weight vector w(n+1) at the instant n+1; S4. Let n be n+1, take the updated weight vector as the new w(n), and then repeat step S3 until the error vector e(n) is less than the set threshold, it is considered that the convergence condition is met, and the weight vector at this time is recorded as w′ as the identification result of the echo interference channel; S5. In the process of the aircraft actually receiving signals, the signal emitted by the aircraft itself is first multiplied by the transposed result of the echo interference channel identification result w′ to calculate the echo signal, and then the aircraft subtracts the echo signal from the received signal to achieve echo cancellation.

2. The echo elimination method for aircraft communication according to claim 1, characterized in that: In step S3, constructing the input matrix U(n) and the output matrix d(n) of the signal receiving model, and calculating the error vector e(n) includes: According to the aircraft transmission signals at the current instant n and the previous L consecutive moments, the input matrix U(n) of the signal receiving model is constructed: U(n)=[u(n),u(n-1),...,u(n-L+1)]∈R M×L ; The received signals of the aircraft at the current instant n and the previous L consecutive moments are recorded as the output matrix d(n): d(n)=[d(n),d(n-1),...,d(n-L+1)] T ∈R L×1 Multiply the input matrix U(n) by the weight vector w(n), and then calculate the error vector e(n) by subtracting the product of U(n) and w(n) from the output matrix d(n): e(n)=d(n)-U T (n)w(n)∈R L×1 。 3. The echo elimination method for aircraft communication according to claim 1, characterized in that: In step S3, the weight vector w(n) at instant n is updated according to the error vector e(n), and the weight vector w(n+1) at instant n+1 is obtained as follows: w(n+1)=w(n)+μU(n)P(n)e(n) Where, P(n)=diag[p(n),p(n-1),…,p(n-L+1)]; μ represents the step size, 0<μ<1; Since e(n)∈R L×1 , the specific form of the error vector is expressed as: e(n)=[e(n),e(n-1),...,e(n-L+1)] T , p(n) is calculated as follows: θ1=med(e(n))-kmed(abs(e(n)-med(e(n)))) θ2=med(e(n))+kmed(abs(e(n)-med(e(n)))) p(n) is the judgment factor for the error signal. When the error signal e(n) does not satisfy θ1<e(n)<θ2, p(n) is set to 0 to eliminate the abnormal value. k is the preset adjustment factor and k>0. med represents the median operation and abs represents the absolute value operation.

Images