1. A zero attraction echo cancellation method based on an arc tangent function is characterized in that: the method comprises the following steps:

the first step is as follows: acquiring a signal, and sampling a voice signal transmitted from a far end to obtain a far end signal discrete value x (n) of the current moment n; meanwhile, the echo signal collected by the near-end microphone is sampled to obtain an expected signal d (n) at the current time n.

The second step is that: calculating the output y (n) of the adaptive filter, and forming the value of the remote signal discrete value x (n) from the current time n to the time n-L +1 into an adaptive filter input vector x (n) of the current time n, wherein x (n) is [ x (n), x (n-1).. once, x (n-L +1)]^TWherein, L represents the number of taps of the adaptive filter, L is 512 or 1024, and the superscript T represents transposition;

the output signal y (n) of the adaptive filter at the current time n, and y (n) W^T(n) x (n), where W (n) is the tap weight vector of the adaptive filter at the current time n, the length of the tap weight vector is equal to L, and the initial value is zero vector, that is, W (1) is 0;

the third step: echo cancellation, subtracting an output signal y (n) of an adaptive filter at the current time n from an expected signal d (n) at the current time n to obtain an error signal e (n) at the current time n, and transmitting the error signal e (n) to a far end as a near-end signal at the current time n after echo cancellation, namely e (n) ═ d (n) -y (n);

the fourth step: the update of the weight coefficient vector is performed,

(1) according to the error signal e (n) of the current time n, the error signal of the current time n for removing the impulse interference is calculated

Wherein arctan (·) represents an arctangent operation;is initially zero, i.e.

(2) Updating to obtain the tap weight vector W (n +1) of the next time n +1,

wherein mu is a step length parameter and takes a value of 0.5; sgn [. cndot ] represents a sign operation; rho is a zero attraction factor, and the value is 0.000004; epsilon is a positive parameter and takes the value of 1-20;

the fifth step: and (5) enabling n to be n +1, and repeating the steps from the first step to the fourth step until the call is ended.

Background

In a communication system, noise and echo interference cannot always be ignored. Among them, acoustic echo is the most important one affecting voice call quality. The acoustic echo refers to the sound played by a loudspeaker which is picked up by a microphone and sent back to a far-end, so that a far-end talker can hear the sound of the far-end talker. For example, acoustic echo is often generated when a multi-person network audio conference is held or a user uses a hands-free function of a communication device. The short echo delay can be hardly perceived, and can be understood as a form of spectral distortion. Conversely, the echo is clearly noticeable with a delay of more than a few tens of milliseconds. Since the human ear is extremely sensitive to echoes, the study of methods for eliminating acoustic echoes is still a popular topic. The traditional widely used adaptive echo cancellation method mainly includes methods such as Least Mean Square (LMS) algorithm and Affine Projection (APA) algorithm. In the echo channel, however, the echo is mainly caused by the time delay effect, which is shown in that most echo paths are sparse channels, and only a few pulse values of the sparse system are not zero, and the rest are zero or close to zero. In such systems the steady state error of conventional LMS and APA algorithms becomes significantly larger with a concomitant slowing down of convergence. Meanwhile, if there is interference noise such as impulse noise in the system, the above conventional algorithm becomes very unstable.

Document 1, "Using a kernel function in linear adaptive filters" (single, a.and j.c. principal, proceedings of the International Joint Conference on Neural Networks) (2009)), discloses a mature echo cancellation method for resisting impact interference in the current sparse system identification application.

Disclosure of Invention

The invention aims to solve the technical problem of overcoming the defects of the prior art and provides the zero-attraction echo cancellation method based on the arctan function, which has high convergence rate, low steady-state error and strong anti-interference capability.

The invention solves the technical problem by adopting the technical scheme that a zero-attraction echo cancellation method based on an arc tangent function comprises the following steps:

the first step is as follows: acquiring a signal, and sampling a voice signal transmitted from a far end to obtain a far end signal discrete value x (n) of the current moment n; meanwhile, the echo signal collected by the near-end microphone is sampled to obtain an expected signal d (n) at the current time n.

The second step is that: calculating the output y (n) of the adaptive filter, and forming the value of the remote signal discrete value x (n) from the current time n to the time n-L +1 into an adaptive filter input vector x (n) of the current time n, wherein x (n) is [ x (n), x (n-1).. once, x (n-L +1)]^TWherein, L represents the number of taps of the adaptive filter, L is 512 or 1024, and the superscript T represents transposition;

the output signal y (n) of the adaptive filter at the current time n, and y (n) W^T(n) x (n), where W (n) is the tap weight vector of the adaptive filter at the current time n, the length of the tap weight vector is equal to L, and the initial value is zero vector, that is, W (1) is 0;

the third step: echo cancellation, subtracting an output signal y (n) of an adaptive filter at the current time n from an expected signal d (n) at the current time n to obtain an error signal e (n) at the current time n, and transmitting the error signal e (n) to a far end as a near-end signal at the current time n after echo cancellation, namely e (n) ═ d (n) -y (n);

the fourth step: the update of the weight coefficient vector is performed,

(1) according to the currentError signal e (n) at time n, and error signal for removing impulse interference at current time n

Wherein arctan (·) represents an arctangent operation;is initially zero, i.e.

(2) Updating to obtain the tap weight vector W (n +1) of the next time n +1,

wherein mu is a step length parameter and takes a value of 0.5; sgn [. cndot ] represents a sign operation; rho is a zero attraction factor, and the value is 0.000004; epsilon is a positive parameter and takes the value of 1-20;

the fifth step: and (5) enabling n to be n +1, and repeating the steps from the first step to the fourth step until the call is ended.

The value of the tangent function in the method of the invention is varied according to the state of the noise environment and, when no impulse noise is present,when there is an impact noise present, the noise,close to zero. Namely, when the impact noise exists, the algorithm is not updated, which shows that the algorithm has good capability of resisting the impact noise and can obtain smaller steady-state error; when there is no impulse noise, the formula is updatedThe term is close to e (n), and the algorithm is similar to NLMS. Therefore, the algorithm can obtain a fast convergence speed and has good noise resistance. When the system is a sparse system, the zero attraction item in the inventionNon-zero elements in the system can be well distinguished, namely, when the elements are zero, the zero attraction item is zero; when the elements are non-zero elements, the zero attraction is large, and the convergence speed is accelerated. The invention can reconcile the contradiction between the fast convergence speed and the low steady-state error and resist the impact noise.

Drawings

FIG. 1 is a channel diagram of a simulation experiment using an embodiment of the present invention;

fig. 2 shows a normalized steady-state offset curve of a simulation experiment when the real speech signal is the input signal according to the method of document 1(MCC) and the method of the embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples.

The embodiment comprises the following steps:

the first step is as follows: acquiring a signal

Sampling a voice signal transmitted from a far end to obtain a far end signal discrete value x (n) of the current moment n; meanwhile, the echo signal collected by the near-end microphone is sampled to obtain an expected signal d (n) at the current time n.

The second step is that: computing the output of an adaptive filter y (n)

The values of the far-end signal discrete value x (n) from the current time n to the time n-L +1 form an adaptive filter input vector x (n) of the current time n, wherein x (n) is [ x (n), x (n-1). ], x (n-L +1)]^TWherein, L represents the number of taps of the adaptive filter, L is 512 or 1024, and the superscript T represents transposition;

the output signal y (n) of the adaptive filter at the current time n, and y (n) W^T(n) x (n), where W (n) is the adaptation of the current time nThe tap weight vector of the filter has the length equal to L, and the initial value is zero vector, namely W (1) is 0;

the third step: echo cancellation

Subtracting an output signal y (n) of the adaptive filter at the current time n from an expected signal d (n) at the current time n to obtain an error signal e (n) at the current time n, and transmitting the error signal e (n) to a far end as a near-end signal at the current time n after echo cancellation, namely e (n) ═ d (n) -y (n);

the fourth step: updating of weight coefficient vectors

(1) According to the error signal e (n) of the current time n, the error signal of the current time n for removing the impulse interference is calculated

Wherein arctan (·) represents an arctangent operation;is initially zero, i.e.

(2) Updating to obtain the tap weight vector W (n +1) of the next time n +1,

wherein mu is a step length parameter and takes a value of 0.5; sgn [. cndot ] represents a sign operation; rho is a zero attraction factor, and the value is 0.000004; epsilon is a positive parameter and takes the value of 1-20;

the fifth step: and (5) enabling n to be n +1, and repeating the steps from the first step to the fourth step until the call is ended.

In order to verify the effectiveness of the present invention, simulation experiments were performed on the method of the present example, and compared with the method of the prior document 1.

The echo channel impulse response of the simulation experiment is obtained in a quiet closed room with the length of 6.25m, the width of 3.75m, the height of 2.5m, the temperature of 20 ℃ and the humidity of 50 percent, and the impulse response length, namely the number L of taps of the filter is 512. The background noise is zero-mean white gaussian noise with a 30dB signal-to-noise ratio. The sampling frequency is 8 KHz.

According to the above experimental conditions, the echo cancellation experiment is carried out by using the method of the present invention and the method of the prior document. The experimental optimal parameter values for each method are shown in table 1.

TABLE 1 values of the optimum parameters for the experiments of the methods

Wherein, fig. 1 is a channel diagram of a communication system consisting of a quiet enclosed room for experiment; fig. 2 shows a normalized steady-state imbalance curve obtained from simulation experiments when the real speech signal is the input signal according to the MCC method and the MCC method of the present invention.

As can be seen from fig. 2: the invention converges at about 15000 sampling moments, and the steady state error is about-18 dB; while document 1 also converges at about 15000 sample times, but the steady state error is about-12 dB; the steady state error of the present invention is reduced by-6 dB from that of document 1.

Various modifications and variations of the present invention may be made by those skilled in the art, and they are still within the scope of the present patent invention provided they are within the scope of the claims and their equivalents.

What is not described in detail in the specification is prior art that is well known to those skilled in the art.