KR20020014259A - Tep calculation update method in channel equalizator using blind equalization - Google Patents

Abstract

PURPOSE: A tap coefficient updating method in a channel equalizer using a blind equalization is provided to improve an initial channel equalization capacity due to a serious channel distortion by increasing an initial convergence speed in a transient state according to a newly adopted error item e¬BG (n) by performing a blind equalization through a tap coefficient updating equation simultaneously using a decision directed error item and a Sato error item. CONSTITUTION: A blind equalization is performed according to a Sato error item of a tap coefficient updating equation at the initial state of equalization(S1). That is, by performing a blind equalization according to the Sato error item in the early state of equalization, an initial convergence speed is increased. Subsequently, when a channel equalizer is converted according to performing of the blind equalization by the Sato error item, a blind equalization is performed according to a decision directed error item of the tap coefficient updating equation(S2). That is, a residual error is reduced by performing a blind according to the decision directed error item in a stable state after the channel equalizer is converged.

Description

How to update tap coefficients in channel equalizer using blind equalization {TEP CALCULATION UPDATE METHOD IN CHANNEL EQUALIZATOR USING BLIND EQUALIZATION}

The present invention relates to a tap coefficient updating method in a channel equalizer using blind equalization, and in particular, a channel equalizer using blind equalization by a stop and go algorithm is determined by a tap coefficient updating equation. Tap coefficient in channel equalizer using blind equalization to improve initial channel equalization capability and reduce residual error after convergence by using both Decision Directed and Sato error terms It is about an update method.

In general, since the physical channel of a digital communication system is received through multiple time-varying multipaths, intersymbol interference (ISI) is generated. When a signal is transmitted to such a non-ideal channel, signal distortion occurs at the receiver side. It may cause a bit detection error and fail to recover the original signal.

In order to overcome such a phenomenon, an adaptive channel equalizer is used in a digital communication system. The adaptive channel equalizer processes a received signal that has passed through a distorted channel out of an ideal characteristic and compensates for the characteristics of the channel to compensate for the bit detection error at the receiving side. To reduce it.

That is, when the transmitting side transmits the data training sequence known to the receiving side for a predetermined period of time, the receiving side compares the distorted data with the original data through the channel, estimates the distortion level of the channel, and then taps the equalizer's tap. Adjusting the coefficient cancels out the distortion characteristics of the channel.

However, in many cases, it is necessary to allow channel equalization to be performed initially without such training sequence, that is, to compensate for channel distortion only with the received signal, and such adaptive channel equalization is called blind equalization. do.

Such blind equalization is mainly used in radio or TV broadcasting, that is, point to multi point communication, and is rarely used in point to point communication.

That is, in the one-to-one communication method, once a symbol error occurs on the receiving side according to a channel change, the tap coefficient of the equalizer is adjusted to cancel the changed channel characteristics by stopping data transmission and transmitting and receiving a training sequence. Data transmission can be resumed later, but in one-to-multipoint communication method, since the channel environment is different between broadcasting stations and receivers, it is impossible to use the same method as in the one-to-one communication method. Should use blind equalization.

On the other hand, in the blind equalization, a decision based algorithm using a data symbol sequence estimated from the output of the equalizer instead of a training sequence known in advance, and since the reliability of the discriminator is not very high at the initial stage of equalization, the quantization interval is broadened to make the sparse quantization more sparse There is a Sato algorithm for increasing the reliability of the algorithm and a stop-and-go algorithm for determining whether to update the tap coefficient of the equalizer by using the error signal of the Sato algorithm to improve the convergence performance of the decision-based algorithm. .

In this case, in the case of a channel equalizer using blind equalization by the stop and go algorithm, if the sato error is the same sign as the judgment error, the judgment error is judged to be sufficiently reliable, and the tap coefficient of the equalizer is updated. If the sign is signed, the tap coefficient of the equalizer is not updated.

However, in the channel equalizer using the blind equalization by the conventional stop and go algorithm as described above, when the two error codes used for the tap coefficient update are the same, the tap coefficient update is limited to allow the initial channel tracking capability such as the convergence speed. There was a problem that had a lot of influence.

That is, in the early stage of equalization, since the channel distortion is severe and the eye pattern is closed, the two error terms used are the same, so even if the tap coefficient is updated, the number of times of updating the tap coefficient is limited.

The present invention is to solve the above problems, the object of the channel equalizer using the blind equalization by the stop and go algorithm in the tap coefficient update equation by using the determination error term and the sato error term at the same time severe channel distortion The present invention provides a method of updating tap coefficients in a channel equalizer using blind equalization that improves the initial channel equalization capability and reduces residual errors after convergence.

1 is a structural diagram of a channel equalizer to which the tap coefficient updating method of the present invention is applied;

2 is a view showing a data structure of the VSB method of the present invention;

3 is a diagram illustrating a tap coefficient update region in a channel equalizer using blind equalization according to the present invention.

4 is a flowchart illustrating a tap coefficient updating method in a channel equalizer using blind equalization according to the present invention.

<Explanation of symbols for main parts of drawing>

10 to 10-N: delay 20 to 20-N: multiplier

30: adder

The tap coefficient updating method in the channel equalizer using the blind equalization of the present invention for achieving the above object is, blind equalization by the Sato error term in the tap coefficient updating equation at the initial stage of equalization when the blind equalization is performed by the stop and go algorithm. If the channel equalizer converges by performing equalization by the Sato error term and performing blind equalization by the Sato error term during the equalization by the Sato error term, blind equalization by the error term according to the determination of the tap coefficient update equation is performed. Characterized by performing the equalization step by the error terms based on the determination performed.

Hereinafter, a tap coefficient updating method in a channel equalizer using blind equalization of the present invention will be described in detail with reference to the accompanying drawings.

1 is a structural diagram of a channel equalizer to which the tap coefficient updating method of the present invention is applied, that is, a channel equalizer using a finite impulse response (FIR) filter, and continuously delays the equalizer input signal x (n) at time n. The delays 10 to 10-N, the equalizer input signals x (n) to x (n-N + 1) delayed by the delays 10 to 10-N, and the filter at time n. An adder 30 that adds both the multipliers 20 to 20-N to multiply the first tap coefficient C_i (n) and the outputs of the multipliers 20 to 20-N, and outputs the output signal z (n) of the equalizer. It consists of.

In the present invention, a tap equalizer updating method is implemented in a channel equalizer using blind equalization, that is, a channel equalizer using an FIR filter, and the tap coefficients of the equalizer input signal, the equalizer reference signal, and the filter are defined as follows. .

x (n): equalizer input signal at time n,

d (n): equalizer reference signal at time n,

C_i (n): i-th tap coefficient of the filter at time n,

And, the error signal at the equalizer output is expressed by the following equation (1).

e (n) = d (n)-z (n)

Therefore, the tap coefficient update expression of the filter may be expressed as Equation 2 when N is the total number of taps of the channel equalizer.

C_i (n + 1) = C_i (n) + mu (n) x (n-i) for i = 0, 1,... , N-1

Where mu is the step size.

The output signal of the equalizer using this updated tap coefficient is expressed by Equation 3 below.

On the other hand, in the decision-based algorithm of the algorithm for blind equalization, the decision-based error term is given by the following equation (4), and the tap coefficient update equation of the equalizer is shown by the following equation (5).

e ^ D (n) = z (n)-x ^ Λ (n)

Here, x ^ Λ (n) is a symbol determined by a corresponding threshold, and is a value obtained by quantizing an equalizer output at a level interval of a transmission symbol.

C (n + 1) = C (n)-mu ｅ ^ D (n) x (n)

Here, μ is a factor for determining the convergence speed, and represents the size of adjusting the tap coefficient.

In the Sato algorithm, the estimated value x ^ Λ (n) of the transmitted data is as shown in Equation 6 below, and the error signal at this time is as shown in Equation 7, and the tap coefficient update equation of the equalizer is as shown in Equation 8. , Sato error constant is as shown in Equation (9).

x ^ Λ (n) = γ sgn [z (n)]

C (n + 1) = C (n)-mu ｅ ^ S (n) y (n)

Finally, in the stop and go algorithm, the tap update equation of the equalizer is as shown in Equation 10, and the flag f (n) for determining whether to allow tap equalization of the equalizer is determined by the following equation (11). do.

C (n + 1) = C (n)-mu (f (n) ｅ ^ D (n) x (n))

Here, sgn represents the sign of each error signal.

FIG. 2 is a diagram illustrating a data structure of a VSB method of the present invention, and illustrates a data structure of a high definition television terrestrial broadcasting mode of a residual side band modulation method in a digital communication system. It can be seen that it is composed of the levels (-7, -5, -3, -1, + 1, + 3, + 5, + 7).

As described above, in the case of VSB-type data transmitted at eight levels, when blind equalization is performed by the stop-and-go algorithm, the Sato error constant may be calculated as in Equation 12 below.

Therefore, the tap coefficient update area of the stop and go algorithm using the error terms of the decision based algorithm and the Sato algorithm is shown in FIG. 3.

Here, the "+,-" symbol indicates the sign of the judgment error or the sato error term, and when the sign of the judgment error and the sato error is the same by the stop-and-go algorithm, It is higher than the probability that the signs are equal.

On the other hand, the Sato algorithm has a blind equalization capability in the early stage of equalization, but has a disadvantage that the residual error becomes a problem after the equalizer converges, and the determination algorithm is equalized when the eye model is closed due to severe channel distortion at the beginning of the equalization. There is a disadvantage of not being able to do it, but there is an advantage that the residual error is much smaller in the open state after the initial convergence and the tracking ability of the channel situation is good.

Therefore, using the stop-and-go algorithm for channel equalization, in the initial converged state where the eye model is closed, the probability of equality of the sign of the decision error and the sato error is relatively reduced, resulting in a large delay in updating the tap coefficient. do.

Accordingly, in the present invention, channel equalization is performed using a tap coefficient update equation as shown in Equation (13).

for i = 0,1,... , N-1

Here, e ^ BG (n) combines the merits of the decision-based algorithm and the Sato algorithm, and is limited by Benvenist and Goursat, and k_1 and k_2 are positive values determined by experimental values.

Therefore, when tap coefficient update is allowed by the tap coefficient update flag f (n), the value of the error term e ^ D (n) will be relatively large at the initial stage of equalization in the transient state. (n) | · e ^ S (n) |

When the equalizer becomes stable as the convergence progresses, the error term of the decision becomes smaller, so that the influence of residual error due to the Sato error on e ^ BG (n) is reduced, so that the transition between the decision-based algorithm and the Sato algorithm is naturally performed.

That is, the channel equalizer using blind equalization by the stop and go algorithm according to the present invention allows tap coefficient updating by combining two error terms used in the decision based algorithm and the Sato algorithm. The operation flow chart of FIG. 4 will be described below.

In performing equalization by the Sato error term when performing the blind equalization by the stop and go algorithm, blind equalization by the Sato error term in the tap coefficient update equation of Equation 13 is performed at the initial stage of the equalization.

That is, the initial convergence speed is increased by performing blind equalization by the Sato error term at the initial stage of equalization as in step S1.

Subsequently, after the equalization step S1 of the Sato error term is performed, the process proceeds to the equalization step S2 of the error term based on the determination and the channel equalizer converges by performing the blind equalization by the Sato error term in the step S1. The blind equalization based on the error term in the tap coefficient update equation of Equation 13 is performed.

That is, as in the step S2, the channel equalizer performs the blinding by the error term based on the determination in the converged and stable state so as to reduce the residual error.

As described above, according to the present invention, the channel equalizer using the blind equalization by the stop and go algorithm is newly applied by performing the blind equalization through the tap coefficient update equation using the error term and the sato error term. The error term e ^ BG (n) increases the initial convergence speed in the transient state, improving the initial channel equalization ability due to severe channel distortion, and reducing the residual error even in the stable state after convergence. There is an effect that can be further increased.

Claims (2)

Performing equalization by the Sato error term for performing blind equalization by the Sato error term in the tap coefficient update equation at the beginning of the equalization when performing blind equalization by the stop and go algorithm; If the channel equalizer converges to perform the blind equalization by the Sato error term in the equalization by the Sato error term, the equalization by the error term is performed by performing the blind equalization according to the error term based on the determination of the tap coefficients. A tap coefficient updating method in a channel equalizer using blind equalization, characterized in that the. The method of claim 1, wherein the tap coefficient update expression, for i = 0,1,... , N-1 tap method in the channel equalizer using the blind equalization, characterized in that.