KR101057365B1 - Preamble Detection Method for Burst Mode Packet Transmission System - Google Patents

Abstract

The present invention relates to a method for detecting a preamble in a burst mode packet transmission system. More particularly, the present invention relates to a method for detecting a preamble in a system for transmitting a frame or a packet in burst mode. It is more reliable by using the minimum misprobability characteristics in the data domain, the idle interval, the preamble interval, the dynamic threshold characteristics in the data interval, the minimum leakage probability characteristic at the minimum signal level, the maximum detection characteristic even when the preamble size changes and the frequency offset exist. Transmission frame reception is possible.

Burst Mode, Preamble, Cross Correlation, False Probability, Error Probability, Threshold

Description

Preamble Detection Method for Burst Mode Packet Transmission Systems

The present invention relates to a method for detecting a preamble in a burst mode packet transmission system, and more particularly, to a method for detecting a preamble in a system for transmitting a frame or packet in burst mode.

In general, the transmission frame configuration in burst mode communication includes a ramp-up signal for the stabilization of high-power amplifiers and AGC operation, and a signal known in advance to know the starting point of the frame, followed by data. In this case, the AGC loop of the receiver should be designed to have maximum gain in the idle period so that the presence of a burst can be known even at a small signal level. If there is a preamble at the start of the burst, if there is not enough AGC response speed in the burst period, the signal saturates for several symbol intervals at the start point of the preamble and fails to detect the preamble or, even if detected, is required for demodulation using the preamble. It adversely affects the estimation of various parameters. Therefore, in order to normally detect the preamble in burst mode, stable gain adjustment must be preceded at the start of the burst. That is, failure of the AGC operation leads to failure of the preamble detection and eventually fails to receive the burst.

For example, VDL MODE-2 uses a D8PSK modulation scheme and transmits a packet in burst mode. In this case, the AGC operating range of the receiver must be 100dB or more to satisfy the reception sensitivity required by the VDL MODE-2 standard. However, there was a problem that it is difficult to stabilize the gain of the 100dB range without invading the preamble region in the burst mode.

는 다음 수학식 1과 같이 나타낼 수 있다.As a result, a preamble detector capable of preamble detection even when AGC operation is unstable is required. Conventional preamble detection divides the magnitude of the cross-correlation value between the received signal and the preamble coefficient by the average power value of the received signal and determines that the preamble exists when the normalized value is larger than the predetermined threshold value TH. In this case, a result value corresponding to the correlation value

May be expressed as in Equation 1 below.

 However, this method is relatively accurate in the preamble period, but the false probability property is not good in the idle period without the burst, the data period after the preamble, and the end of the burst.

In addition, burst mode AGC typically attempts to gain gain at low gain to detect low-level bursts in idle sections, and attempts to adjust gain when bursts appear, resulting in high false probability as a result of noise processes during idle periods. do. In addition, there may be a pattern similar to the preamble in the data section, potentially having a high false probability, and a false probability due to AGC operation always exists at the end of the burst.

This preamble detection is a part that must be performed before initial frequency synchronization and symbol timing synchronization. Therefore, it is necessary to derive an algorithm that can operate stably even in the presence of carrier frequency, phase, and timing offset. In this case, in general, the larger the frequency offset, the smaller the magnitude of the correlation value, so that the preamble detection characteristic is worse.

For example, in the case of the VDL MODE-2 system, when the frequency offset between the transceiver and the Doppler frequency shift value are considered, there may be an initial frequency offset of up to 960 Hz. Since this is a frequency offset that can cause a phase rotation of about 33 degrees per D8PSK symbol, there is a problem that can cause severe performance degradation when the preamble detection method used in the prior art is directly applied.

An object of the present invention for solving the above-described conventional problem is to receive a transmission frame to ensure more reliability in a system for transmitting a frame or packet in a burst mode.

-DQPSK, DBPSK 등 다양한 변조방식에 적용 가능한 버스트 모드 패킷 전송 시스템을 위한 프리앰블 검출 방법을 제시하는 데 또 다른 목적이 있다.In addition, not only D8PSK

Another object of the present invention is to propose a preamble detection method for a burst mode packet transmission system applicable to various modulation schemes such as DQPSK and DBPSK.

The preamble detection method in the burst mode packet transmission system according to the present invention for solving the above-described conventional problems and to achieve the technical problem, in order to minimize the influence of the carrier frequency offset in obtaining the cross-correlation value for the preamble detection A step of obtaining a correlation value after performing decoding; Obtaining a cross-correlation value C [n] and an average power S [n] of a signal as an initial variable for preamble detection after performing the differential decoding; C) calculating M values, U values, L values, and C values by combining the initial variables C [n] and S [n] and their delayed signals; A d step of adding the obtained M value, U value, L value, and C value to a threshold value; Determining that the preamble is received when the correlation value is greater than the set threshold value; Estimating a sample position having a maximum correlation value as an initial symbol timing when the preamble is received; And estimating an initial phase (or frequency) offset with a correlation phase value at a position having a maximum correlation value when the preamble is received.
In this case, n means a sample index of the current point in time.

Further, here, the moving average (1-symbol interval) of the cross-correlation value (| C [n] |) to have a deep null between two M-shaped peaks and the cross-correlation value (| C [ n] |) is delayed by two symbol intervals, and the moving average (1-symbol interval) is added to the moving average value to obtain the M value so that null is located between the peaks.

In addition, the moving average (1 / 2-symbol interval) is obtained by subtracting the cross-correlation value (| C [n] |) from the average power size (| S [n] |) so that a U-shaped deep null exists. After subtracting one symbol interval delayed cross-correlation value (| C [n] |) from the average power magnitude delayed by one and one symbol interval (| S [n] |), the moving average (1 / 2-symbol interval) The U value is obtained by adding the moving average value so that the null value of the U value is located at the null position of the M value.

The cross-correlation delayed at the cross-correlation value (| C [n] |) is a moving average (several symbol intervals) of the average power amount (| S [n] |) delayed so that nulls exist in the peak value of the correlation value. The difference between the value magnitudes (| C [n] |) is added, and if the result is negative, '0' is obtained. If the result is positive, the L value is obtained by taking the corresponding result (CLAMP).

In addition, the moving average of the average power magnitude (| S [n] |) delayed so that null exists in the cross-correlation peak portion is subtracted from the delayed cross-correlation value magnitude (| C [n] |), and the result is negative. In the case of '0', and if it is positive, the L value is obtained by taking the corresponding result (CLAMP).

The C value may be obtained by delaying the peak of the cross-correlation value magnitude | C [n] | to be located at the null center of the M, U, and L values.

In addition, by using the M value, U value, L value, C value, it is set as a threshold value by adding the following equation.

이때, M값(Mvalue)은 프리앰블 검출 윈도우 영역을 한 심벌시간만큼의 샘플수로 제한하기 위한 값이고, U값(Uvalue) 와 L값(Lvalue)는 데이터 영역 및 데이터 영역의 끝부분에서 오보확률을 최소화하기 위한 값이며, G1, G2, 및 G3은 상기 Mvalue, Uvalue, 및 Lvalue에 적용되는 이득 상수값이다.

At this time, M value is a value for limiting the preamble detection window area to the number of samples by one symbol time, and U value and L value are incorrect probability at the end of the data area and the data area. Is a value for minimizing and G1, G2, and G3 are gain constant values applied to the Mvalue, Uvalue, and Lvalue.

In addition, gains (G ₁ , G ₂ , G ₃ ) are set to have minimum leakage probability characteristics at the minimum signal level and minimum false probability characteristics at the rest period, the preamble interval, the data interval, and the end of the burst. Determining an optimal threshold value.

In addition, when the preamble is received, the preamble detection is recognized if the magnitude of the correlation phase at the position where the correlation value is maximum is within the symbol decision boundary value. By ignoring the detection, the false probability can be further reduced.

In addition, in delaying the number of moving average points, the magnitude of the cross-correlation value (| C [n] |), and the average power amount (| S [n] |), the number of delay elements is an oversampling rate and the number of symbols constituting the preamble. And it can be changed by the group delay according to the moving average.

In addition, the threshold value is characterized in that the search window for searching for the peak value of the correlation value after preamble detection is limited to one symbol period.

As described above, the preamble detection method according to the present invention improves reliability in receiving transmission packets by minimizing false probability and maximizing preamble detection probability in the idle period, data region, and burst end by the dynamic threshold characteristic. It becomes possible. In addition, the preamble detection has an advantage of not having to perform division for normalization.

Hereinafter, an embodiment of a preamble detection method according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a frame structure of a burst mode transmission system according to an embodiment of the present invention, FIG. 2 is a diagram showing a structure of a correlator according to an embodiment of the present invention, and FIG. 3 is according to an embodiment of the present invention. 4A is a block diagram illustrating a process for obtaining C [n] and S [n], and FIG. 4A is a block diagram for calculating M, U, L, and C values according to an embodiment of the present invention. FIG. 4B is an embodiment of the present invention. 4A shows a simplified block diagram, FIG. 4C shows a preamble detector according to the present invention, and FIG. 5 shows | C [n] | and | S [n in a preamble section according to an embodiment of the present invention. ] | FIG. 6 is a diagram illustrating an M value and a U value in a preamble section according to an embodiment of the present invention, and FIG. 7 is a diagram showing an L value and a preamble section according to an embodiment of the present invention. 8 is a diagram illustrating a C value, and FIG. 8 illustrates a threshold value TH, a C value, and correlation according to an embodiment of the present invention. It is a figure which plotted phase.

As shown in FIG. 1, in the preamble detection method according to the present invention, the structure of the transmission frame structure includes three 'zero' symbols for stabilization of the transmitter high power amplifier and AGC operation of the receiver, and 16 symbols for knowing the start of the frame. It consists of the preamble, the length field to know the length of the frame, and the data field.

In the structure of the cross correlator 200 to know the existence of the preamble according to an embodiment of the present invention, as shown in FIG. After the correlation value is calculated, the magnitude 22 of the corresponding signal is obtained, and the threshold value is compared with the comparator 23. When the correlation value is larger than the threshold value, it is determined that the preamble exists.

In this case, N in the N-delay element 21 means the over-sampling rate processed by the receiver, and in the embodiment of the present invention, the over-sampling rate is 16, which is applied to FIG. This oversampling rate may have various optimized settings by the user.

3 is a block diagram 300 for obtaining Equations 2 and 3 as an initial step for detecting a preamble according to an embodiment of the present invention. Here, the value obtained by mutually correlating the received signal and the preamble pattern after the differential decoding through the differential decoder 20 is defined as C [n], and the average power value of the received signal is referred to as S [n]. In this case, the average power value is obtained using a plurality of complex signal magnitude squarers 30. These values are used as an input value (see FIGS. 4A and 4B) in the present invention to calculate the threshold required for preamble detection in the embodiment.

In the preamble detection method according to the present invention, a differential decoding (C [n], Equation 2) between a received signal and a preamble pattern generated by a receiver is performed after performing differential decoding to remove the frequency offset effect. Step b and step b of obtaining an average power (S [n], Equation 3) of the received signal as an initial variable for preamble detection after performing the differential decoding, using C values as shown in Table 1, M using the values of the steps It consists of the c step of obtaining the value, the U value, and the L value.

In addition, step d of estimating a threshold value by calculating a threshold value as shown in Equation 4 using the values shown in Table 1 below, e that the preamble exists if the correlation value (C value) satisfies Equation 5 The method may further include a step f of estimating a sample corresponding to the peak value among the correlation values satisfying Equation 5 with an initial symbol timing synchronization (Equation 6).

The method may further include g step of estimating an initial phase or a frequency offset as a correlation phase value at the position having the maximum correlation value when the preamble is received.

(One).

Delayed (1.5 Symbol) Correlator Output (2).

Preamble Area Peak Detect Window (3).

Data area mask (4).

Data Region and Burst End Masks

In Equation 4, gains G1, G2, and G3 are used to determine an optimal threshold value, and have a minimum leakage probability characteristic at a minimum signal level, and a minimum at a rest period, a preamble section, a data section, and a burst end. Try to have misleading probabilities.

In FIG. 5, the average power magnitude (| S [n] |) and the cross-correlation value magnitude || (C [n] |) in the preamble section are shown. As shown, | C [n] | The value forms a large peak with a maximum at the center point of the time axis.

Next, referring to FIGS. 4A and 4B, Tables 1 and 6 to 8, the core of the present invention can be understood.

As shown in FIG. 4A, the M value is a cross-correlation of two symbol delays with a cross-average moving average (40) of a cross correlation value (| C [n] |) so as to have a deep null between two M-shaped peaks. It can be obtained by adding a value magnitude (| C [n] |) to one symbol interval moving average 40 so that null is located between the peaks. Here, in the image of FIG. 6 showing the M value in the preamble section, it can be seen that there are two M-shaped peaks and a deep null is located between the peaks.

In FIG. 4A, the U value is obtained by subtracting the cross-correlation value size | C [n] | from the average power size | S [n] | so that a U-shaped deep null exists. 2-symbol interval) and the average power magnitude (| S [n] |) subtracted from one symbol interval delayed | C [n] | minus one symbol interval delayed moving average (1 / 2-symbol interval). It can be obtained by adding the moving average value so that the null value of U is located at the null position of the M value.

For example, take | S [n-4] | minus | C [n-4] | and take the eight-point moving average (42) and | C [n-20 from | S [n-20] | The value obtained by subtracting the || can be obtained by adding the eight-point moving average (42). Here, the lower part of FIG. 6 showing the U value in the preamble section shows that a U-shaped deep null exists.

In FIG. 4A, the L value is obtained by the | 512 average moving average 44 of | S [n-24] | -Can be obtained by adding | C [n-24] | 'and then' 0 'if the result is negative and taking the result if it is positive (CLAMP). In this case, it is preferable to use the clamp circuit 45.

That is, the moving average (several symbol intervals) of the | S [n] | delayed so that nulls exist in the correlation peak portion is added to the difference of | C [n] | If the result is negative, it is set to '0' and if it is positive, the L value is obtained by using the clamp circuit 45 which takes the result as it is.

Here, the meaning of the L value is derived from the long term MA. Referring to FIG. 7, it can be seen that null is also present in the L value at the peak portion of the C value. 4A in FIG. 4A takes a moving average corresponding to 32 symbol intervals and can be replaced with a moving average 44 'of efficient 32 Ts intervals of hardware, which is a structure for performing MA after I & D in terms of hardware efficiency.

On the other hand, the M value, U value, L values for determining the threshold value in the present invention can be seen that there is a deep null in the peak portion of the C value in common with reference to Figs.

In FIG. 4A, the C value becomes | C [n-24] | delayed so that the peak of | C [n] | is located at the center of the deep null of the M, U, and L values.

Here, the identification number 41 indicates a delay element of one symbol section Ts.

In addition, FIG. 4B illustrates an efficient hardware structure using the linearity of the moving averages 40 and 42 in FIG. 4A.

Meanwhile, 'the structure of a preamble detector for a burst mode packet or frame transmission system' according to the present invention is shown in FIG. 4C. As shown in FIG. 4C, the preamble detection method of the present invention calculates S [n] and C [n] in the same manner as in FIG. 3 after the received signal passes through the channel filter, and S [n]. ] To obtain M, U, L, and C values based on [] and C [n] in the same manner as in FIG. 4A (or FIG. 4B), and add the M, U, L, and C values as shown in Equation 4 above. Determining a threshold value, comparing the determined threshold value with the C value, and determining that the preamble exists when the C value is larger than the threshold value; when the preamble exists, the sample corresponding to the peak among the values satisfying Equation 5 is determined. Estimating a position as an initial timing synchronization point (Equation 6), and estimating a correlation phase corresponding to the maximum position among values satisfying Equation 5 as an initial phase (or frequency) offset.

As described above, the preamble detection method according to the present invention recognizes the preamble detection when the magnitude of the correlation phase at the position where the magnitude of the correlation value is maximum when the preamble is received is within a symbol decision boundary value. If the magnitude of the phase is greater than or equal to the symbol decision boundary value, the false probability is further reduced by ignoring the preamble detection.

The number of delay elements in delaying the number of moving average points, the magnitude of the cross-correlation value (| C [n] |), and the average power amount (| S [n] | And group delay according to the moving average.

In addition, since the threshold value obtained according to the present invention has a null characteristic, it has a characteristic of limiting a search window for finding the peak value of the correlation value after preamble detection to one symbol period.

According to the preamble detection method for the burst mode packet transmission system according to the present invention described in the above description, the minimum misprobability characteristic in the idle period, the minimum misprobability characteristic in the data region, the idle interval, the preamble interval, the data It is possible to derive a dynamic threshold value characteristic, a minimum leakage probability characteristic even at a minimum signal level, and a maximum detection characteristic even when a preamble size change and a frequency offset exist in the interval.

As described above, although the present invention has been described by way of limited embodiments and drawings, the present invention is not limited thereto and will be described below by the person skilled in the art to which the present invention pertains. Of course, various modifications and changes are possible within the scope of the claims.

1 is a diagram illustrating a frame structure of a burst mode transmission system according to an exemplary embodiment of the present invention.

2 illustrates a structure of a correlator according to an embodiment of the present invention.

3 is a diagram illustrating a process of obtaining C [n] and S [n] according to an embodiment of the present invention.

4A is a block diagram for calculating M, U, L and C values according to an embodiment of the present invention.

4B is a simplified block diagram of FIG. 4A in accordance with an embodiment of the present invention.

4C illustrates a preamble detector according to the present invention.

5 is a diagram illustrating | C [n] | and | S [n] | in a preamble section according to an embodiment of the present invention.

6 is a diagram illustrating M and U values in a preamble section according to an embodiment of the present invention.

7 is a diagram illustrating L and C values in a preamble section according to an embodiment of the present invention.

8 is a diagram illustrating a threshold value TH, a C value, and a correlation phase according to an embodiment of the present invention.

<Explanation of Symbols for Main Parts of the Drawings>

20: differential decoder

21: delay elements equal to oversampling rate

22: magnitude of the signal

23: comparator

30: magnitude squarer of a complex signal

40: moving average (16 points MA)

41: delay element of one symbol interval (Ts)

42: moving average (8 points MA)

43: delay element of 1 symbol interval (Ts)

44: moving average (512 points MA)

44 ': Moving average over hardware-efficient 32Ts

45: Clamp circuit that takes negative value as zero and positive value as it is

50: delay element for delaying the received sample equal to the processing delay required for calculating the threshold and peak values

Claims (11)

A preamble detection method in a burst mode packet transmission system, A step of obtaining a correlation value after performing differential decoding to minimize the influence of a carrier frequency offset in obtaining a cross correlation value for preamble detection; Obtaining a cross-correlation value C [n] and an average power S [n] of a signal as an initial variable for preamble detection after performing the differential decoding; C) calculating M values, U values, L values, and C values by combining the initial variables C [n] and S [n] and their delayed signals; A d step of adding the obtained M value, U value, L value, and C value to a threshold value; Determining that the preamble is received when the correlation value is greater than the set threshold value; Estimating a sample position having a maximum correlation value as an initial symbol timing when the preamble is received; And And estimating an initial phase (or frequency) offset with a correlation phase value at a position having a maximum correlation value when the preamble is received. In this case, n means a sample index of the current point in time. The method of claim 1, Moving average (1-symbol interval) of cross-correlation magnitude (| C [n] |) to have a deep null between two M-shaped peaks and the cross-correlation magnitude (| C [n] |) And delaying two symbol intervals, moving average (1-symbol interval), and then adding the moving average value to obtain an M value so that a null is located between the peaks. The method of claim 1, The average power magnitude (| S [n] |) minus the cross-correlation value (| C [n] |) is equal to the moving average (1 / 2-symbol interval) so that a U-shaped deep null exists. The moving average after subtracting one symbol interval delayed cross correlation value | C [n] | from a symbol interval delayed average power magnitude (| S [n] |) and then moving average (1 / 2-symbol interval). And adding a value to obtain a U value so that a null value of U is located at a null position of the M value. The method of claim 1, The cross-correlation magnitude delayed at the cross-correlation magnitude (| C [n] |) to the moving average (several symbol intervals) of the average power magnitude (| S [n] |) delayed so that nulls exist in the correlation peak portion. (| C [n] |) is added, and if the result is negative, the result is '0', and if it is positive, the L value is obtained by obtaining the corresponding result (CLAMP). The method of claim 1, Subtract the delayed cross-correlation value (| C [n] |) from the moving average of the average power amount (| S [n] |) delayed so that nulls exist in the peak portion of the correlation value, and if the result is negative, '0. And if it is positive, take the corresponding result (CLAMP) to obtain the L value. The method of claim 1, And obtaining a C value by delaying the peak of the cross-correlation value (| C [n] |) to be located at a null center of the M, U, and L values. The method of claim 1, And using the M value, the U value, the L value, and the C value as shown in the following equation to set a threshold value.

At this time, M value is a value for limiting the preamble detection window area to the number of samples by one symbol time, and U value and L value are incorrect probability at the end of the data area and the data area. Is a value for minimizing and G1, G2, and G3 are gain constant values applied to the Mvalue, Uvalue, and Lvalue. The method of claim 7, wherein The gain constants (G ₁ , G ₂ , G ₃ ) are set to have minimum leakage probability characteristics at the minimum signal level and minimum misprobability characteristics at the rest period, the preamble interval, the data interval, and the end of the burst. And determining a threshold of the preamble. The method of claim 1, When the preamble is received, the preamble detection is recognized if the magnitude of the correlation phase at the position where the correlation value is maximized is within the symbol decision boundary value. If the magnitude of the correlation phase is equal to or greater than the symbol decision boundary value, the preamble detection is performed. A method of detecting a preamble, characterized in that the error probability is further reduced by ignoring. The method according to any one of claims 2, 3, 4, 5, and 6, In delaying the number of moving average points, the magnitude of the cross-correlation value (| C [n] |) and the average power amount (| S [n] |), the number of delay elements is the number of symbols and shifts constituting the oversample rate and the preamble. Preamble detection method characterized in that it can be changed by the group delay according to the average The method of claim 1, The threshold value preamble detection method characterized in that to limit the search window to find the peak value of the cross-correlation value to one symbol interval after preamble detection

Images