KR20040108213A - Symbol timing regeneration device for receiver - Google Patents

Abstract

PURPOSE: A symbol timing recovery apparatus of a receiver is provided to perform stably a timing control operation by using a Kaiser window interpolation filter or a synchronous interpolation filter utilizing a Kaiser window function. CONSTITUTION: A sampler(611) is used for sampling a symbol as a predetermined multiple. A matching filter(621) is used for removing a remaining side band from an output signal of the sampler and transmitting only a desired frequency. A Kaiser window interpolation filter(631) is used for interpolating an output signal of the matching filter by using a Kaiser window function. A timing error detector(641) is used for extracting a timing error of an output signal of the Kaiser window interpolation filter. A loop filter(651) is used for reducing a phase and a frequency error of an output signal of the timing error detector. A numerical controlled oscillator(661) is used for receiving an output signal of the loop filter and providing information necessary for an interpolating operation of the Kaiser window interpolation filter.

Description

Symbol timing regeneration device for receiver

The present invention relates to an apparatus for recovering symbol timing of a receiver, and in particular, when a sampling timing of a received symbol is not made in a digital communication system, an error probability increases. It is about.

When the transmitter modulates and transmits a signal, the receiver demodulates the modulated signal to reproduce the original signal. In the process of demodulating the received signal, the sampling timing of the signal must be restored, and the symbol timing recovery apparatus is used. As a symbol timing recovery device, a sync interpolation filter is generally used. However, the sync interpolation filter has a disadvantage in that the symbol timing offset convergence speed is slow. Therefore, the sync interpolation filter needs to be improved.

The present invention has been made to solve the disadvantages of the conventional sync interpolation filter, and an object thereof is to provide a symbol timing recovery apparatus having an improved symbol timing offset convergence speed.

BRIEF DESCRIPTION OF THE DRAWINGS In order to better understand the drawings cited in the detailed description of the invention, a brief description of each drawing is provided.

1 is a block diagram of a symbol timing recovery apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram of the Kaiser window interpolation filter shown in FIG. 1.

3 is a graph illustrating the frequency response of interpolation filters.

4 shows the envelope of the Kaiser window function.

FIG. 5 is a block diagram of the loop filter shown in FIG. 1.

6 is a block diagram of a symbol timing recovery apparatus according to another embodiment of the present invention.

7A and 7B show symbol timing variances for the initial timing offset of the sync interpolation filter and the Kaiser window interpolation filter, respectively.

8 shows the symbol timing variance for the initial timing offset of the interpolation filters for the Lysian fading channel.

9A to 9C show simulation results of symbol timing offset convergence processes of interpolation filters.

The present invention to achieve the above object

CLAIMS 1. A symbol timing recovery apparatus of a receiver for receiving a signal representing successive symbols, comprising: a sampler sampling the symbols by a predetermined multiple; A matched filter for removing residual sidebands of the signal output from the sampler and passing only a desired frequency band; A Kaiser window interpolation filter for interpolating and outputting the signal output from the matched filter using a Kaiser window function; A timing error detector for extracting a timing error of the signal output from the Kaiser window interpolation filter; A loop filter for reducing phase and frequency errors of the signal output from the timing error detector; And a numerically controlled oscillator receiving the output signal of the loop filter and providing information necessary for the interpolation operation of the Kaiser window interpolation filter.

In order to achieve the above object, the present invention also provides

CLAIMS 1. A symbol timing recovery apparatus of a receiver for receiving a signal representing successive symbols, comprising: a sampler sampling the symbols by a predetermined multiple; A matched filter for removing residual sidebands of the signal output from the sampler and passing only a desired frequency band; A sync interpolation filter for interpolating and outputting a signal output from the matched filter using a convolution of an impulse response for sync interpolation and a Kaiser window function; A timing error detector for extracting a timing error of the signal output from the sink interpolation filter; A loop filter for reducing phase and frequency errors of the signal output from the timing error detector; And a numerically controlled oscillator receiving the output signal of the loop filter and providing information necessary for the interpolation operation of the sync interpolation filter.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings denote like elements.

1 is a block diagram of a symbol timing recovery apparatus according to an embodiment of the present invention. Referring to FIG. 1, the symbol timing recovery apparatus 101 includes a sampler 111, a matched filter 121, a Kaiser window interpolation filter 131, a timing error detector 141, a loop filter 151, and a numerically controlled oscillator ( 161. The symbol timing recovery apparatus 101 receives a signal representing a continuous symbol from a transmitter (not shown).

The sampler 111 receives the signal r (t) and samples and outputs a symbol by a predetermined multiple, for example, four times. In other words, the input signal r (t) is sampled at a period Ts.

The matching filter 121 removes the residual sidebands of the signal mTs output from the sampler 111 and passes only a desired frequency band.

The Kaiser window interpolation filter 131 interpolates and outputs the signal x (mTs) output from the matched filter 121 using a Kaiser window function. A detailed block diagram of the Kaiser window interpolation filter is shown in FIG. Referring to FIG. 2, the Kaiser window interpolation filter 131 includes delayers D1 to Dn, multipliers MT1 to MTn + 1, and adders ADD1 to ADDn.

The function of the Kaiser window interpolation filter 131 is to generate a synchronized signal for accurate demodulation of the signal. The Kaiser window interpolation filter 131 may further include a decimator, but the decimator varies the sampling rate before and after interpolation, but the decimation process is not performed separately because the decimation process is performed by itself in the interpolation filter.

When the number of samples per symbol is small, a Kaiser window interpolation filter 131 is required for accurate timing correction. Since the sampling point of the sampler 111 may not be the Nyquist point, the Kaiser window interpolation filter 131 interpolates the timing error to have a sample value at the Nyquist point.

The band limited input signal may be recovered using a Kaiser window interpolation filter 131. In this case, the impulse response {h _I (t)} of the Kaiser window interpolation filter 131 is expressed by Equation 1 below.

Where α = Where I ₀ (·) is a modified zero-order Bessel function, where 0 ≦ t ≦ M. In fact, the integration of a typical sink interpolation filter (not shown) is an ideal filter when it extends from negative infinity to positive infinity. However, because this is not feasible, it is necessary to cover the window with an ideal impulse response to make it finite.

If the impulse response of the sync interpolation filter considering the zero crossings of the right and left third with respect to the actual pulse value is implemented, the frequency response may be expressed as 311 of FIG. 3. In terms of frequency, the stopband exhibits a sharp attenuation above -50 [dB]. A stopband of more than -80 [dB] may be indicated. Thus, the Kaiser window facilitates access to circular pulses due to the abrupt stopband and low side robe.

In the case of the Kaiser window interpolation filter 131, M is 60 and β is 10. For reference, in contrast to other window functions, the Kaiser window function has two parameters: length (M + 1) and shape parameter (β). By changing (M + 1) and β, i.e., the length and shape of the window, the width of the main lobe and the amplitude of the side lobe can be adjusted appropriately.

The envelope of the Kaiser window function is shown in FIG. 4. 4 shows a continuous envelope of the Kaiser window function with length M + 1 21 for β = 5 (411), β = 10 (421) and β = 15 (431). In Equation 1, (β = 0) is a rectangular window function. As the window function decreases, the side lobes of the Fourier transform become smaller, but the width of the main lobe becomes larger. Keeping β constant and increasing M decreases the width of the main lobe but does not affect the size of the side lobes. Through extensive numerical experiments, Kaiser obtained a pair of formulas to predict in advance the values of M and β that satisfy a given frequency selective filter characteristic. Kaiser found that the peak approximation error was determined by β over a wide range of conditions. If the peak approximation error is fixed, the passband cutoff frequency ω _P of the low pass filter is defined as the maximum frequency. The stopband cutoff frequency ω _S is defined as the lowest frequency. Thus, for low pass filter approximation, the width of the transition band is

Δω = ω _S -ω _P

Becomes

A = -20 log ₁₀ δ

Δ is the peak approximation error. The value of β needed to get a specific value of A

Kaiser experimentally found that (a rectangular window with A = 21 is the case of β = 0). To get the predetermined values of A and Δω, M is

Must be satisfied.

The timing error detector 141 extracts a timing error by searching for a transition of the input signal y _T , and the error signal may be obtained by the following method. x (kT _S ) is the input of the timing error detector at time kT _S , and ε (kT _S ) is the output, that is, error component, of the timing error detector at time kT _S.

The timing error detector 141 combines Gardner's method with a method using a left and right tilt to more stably control timing. If the timing error detector 141 is designed only in accordance with the Gardner method, there is no problem in the operation of the symbol timing recovery apparatus 101 as a whole, but there may be problems such as slow convergence and more jitter. To solve this problem, the peak value of the signal y _T output from the Kaiser window interpolation filter 131 is divided into a negative peak and a positive peak, and the sign of the peak value is used. Ensure proper timing control at all times. If the peak is not clear, the timing error is extracted by multiplying the difference between the early and late samples by the on-time sample value. Therefore, it is possible to overcome the problems caused by the Early Late Gate (ELG) clock recovery method and, unlike the Modified Weighted Early-Late Gate (MELG) clock recovery method, correct timing control is possible even when the peak is not clear. More than 2 samples per symbol ensures proper timing recovery. Where ε _{k, sym} is the timing error at time k, y _{k, sym} is the symbol value at time k (this is the Nyquist point for symbol determination), and y _{k + 1, sym} is at time k + 1 If the symbol value y _{k-1, sym} is a symbol value at time k-1, the operation of the timing error detector is as follows. First, the slope is extracted around the Nyquist point to determine whether the peak is positive or negative.

SL (Left Slope) = (y _{k, sym} -y _{k-1, sym} ),

SR (Right Slope) = (y _{k + 1, sym} -y _{k, sym} )

Positive Peak (PP): (SL> 0) & (SR <0)

NP (Negative Peak): (SL <0) & (SR> 0)

Once the peak is determined, the timing is controlled by the following equation (7).

(1) Positive peak ([SL> 0] & [SR <0])

(a) y _{k, sym} > 0: ε _{k, sym} = y _{k, sym} (y _{k + 1, sym} -y _{k-1, sym} )

(b) y _{k, sym} <0: ε _{k, sym} = -y _{k, sym} (y _{k + 1, sym} -y _{k-1, sym} )

(2) For negative peaks ([SL <0] & [SR> 0])

(a) y _{k, sym} > 0: ε _{k, sym} = -y _{k, sym} (y _{k + 1, sym} -y _{k-1, sym} )

(b) y _{k, sym} <0: ε _{k, sym} = y _{k, sym} (y _{k + 1, sym} -y _{k-1, sym} )

(3) If there is no peak

ε _{k, sym} = y _{k, sym} (y _{k + 1, sym} -y _{k-1, sym} )

The timing error detection algorithm is very simple and only requires two samples per symbol. The detection algorithm is as follows and represents the case of QPSK.

Here, y _TI and y _TQ denote decimated symbols of the virtual and real axes, and u _T (r) denotes an output value of the timing error detector 141.

The loop filter 151 reduces the phase and frequency errors of the signal u _T output from the timing error detector 141. A block diagram of the loop filter is shown in FIG. Referring to FIG. 5, the loop filter 151 includes multipliers 511 and 512, adders 521 and 522, and a feedback part T. The loop filter 151 receives an estimated phase error value and serves to reduce phase and frequency errors. In FIG. 5, K ₁ and K ₂ are coefficients of the loop filter 151, respectively. The reason for the two coefficients of the loop filter 151 is that the weights of the currently detected phase error value and the previously detected phase accumulation value are different from each other. Increasing the gain of the currently detected phase error increases the range of phase tracking in one sample processing in the carrier recovery circuit, but has a disadvantage of being sensitive to noise. On the other hand, if the gain of the accumulated value of the phase error is increased, there is a trade-off relationship in which the effect of noise is reduced but the tracking range is reduced. Therefore, it is very important to understand the characteristics of the loop filter 151 because the coefficient of the loop filter 151 determines the stability and tracking range of a phase locked loop (PLL) phase. If the transfer function of the loop filter 151 shown in Fig. 5 is expressed in the Z region,

Becomes Substituting the above equation into Equation 4 and arranging

We can get the transfer function of PLL. It is difficult to analyze the transfer function of the PLL in the Z region, so consider converting it to the s region. The relationship between Z and S

z = e ^ST

Becomes The bandwidth of the loop filter 151 is very small compared to the clock speed

It can be used as Equation 10 is replaced with a transfer function of the s region as in Equation 13.

To extract characteristic parameters in equation (13)

Normalize to something like In Equation 14, ω _n is a natural frequency and ζ is a damping factor. The natural frequency ω _n and the braking element ζ can be obtained by comparing Equation 13 and Equation 14.

The coefficients k ₁ and k ₂ used in the loop filter 151 are shown in Table 1 below.

Preamble (255) Preamble (31) k ₁ k ₂ k ₁ k ₂ Sink interpolation filter 1/4 1/64 1/2 1/64 Kaiser Window Interpolation Filter 1/4 1/64 1/2 1/64 Sync interpolation filter using Kaiser window 1/4 1/64 1/2 1/64

The number controlled oscillator 161 serves to provide information u _{k and} m _k necessary for operation in the Kaiser window interpolation filter 131. The corrected value of the timing signal may be represented by a basic integer value u _k and a fine value m _k . Therefore, the basic integer value and the fine value may be expressed as in Equation 16 below. Where int (z) is the largest integer that does not exceed z.

m _K = int [kT _i / T _s ]

u _K = kT _i / T _s -m _K

Therefore, 0 ≦ u _K <1.

6 is a block diagram of a symbol timing recovery apparatus according to another embodiment of the present invention. Referring to FIG. 6, the symbol timing recovery apparatus 601 includes a sampler 611, a matched filter 621, a sync interpolation filter 631 using a Kaiser window, a timing error detector 641, a loop filter 651, and a numerical value. A control oscillator 661 is provided.

The sampler 611, the matched filter 621, the timing error detector 641, the loop filter 651, and the numerically controlled oscillator 661 are respectively the sampler 111, the matched filter 121, and the timing error shown in FIG. 1. Since it is the same as the detector 141, the loop filter 151, and the numerically controlled oscillator 161, duplication description is abbreviate | omitted.

The impulse response of the sync interpolation filter 631 using the Kaiser window is defined as in Equation 17 below.

here, ego, Means convolution.

In the case of the sync interpolation filter 631 using the Kaiser window, M is 30 and β is 10. Since M is 30, the result of convolution of n and n is (2n-1), so we apply 30 instead of 60.

7A and 7B are symbol timing variances for the initial timing offset of the sync interpolation filter and the Kaiser window interpolation filter 131, respectively. The modulation method at this time is QPSK, and the initial timing offset is to be. The coefficient of the loop filter 151 when using the sink function is 0.25 and K2 is 0.125, and the coefficient of the loop filter 651 when using the Kaiser window function is 0.25 and K2 is 0.015625. The lines 711a, 721a, 731a, 741a shown in FIG. 7A and the lines 711b, 721b, 731b, 741b shown in FIG. 7B have gains of 0 dB, 5 dB, and 10 dB, respectively. ], 15 [dB].

As shown in FIGS. 7A and 7B, the symbol timing variance of the Kaiser window interpolation filter 631 is much smaller than the sync interpolation filter.

8 shows the symbol timing variance for the initial timing offset of the interpolation filters for the Lysian fading channel. The symbol timing variance of the sync interpolation filter 631 using the Kaiser window shows better performance than the sync interpolation filter when the initial timing offset is small due to the high main lobe (below g point).

9A to 9C show simulation results of symbol timing offset convergence processes of interpolation filters. 9A to 9C show a timing error convergence process of an m-sequence preamble in which m of interpolation filters is 8 and 255 bits. The modulation scheme is QPSK, and the initial sampling clock error is , Eb / N0 is 3 dB. The symbol timing recovery process is shown. In the case of the sync interpolation filter (FIG. 9A), the larger the influence of the noise, the slower the convergence speed of the initial timing error is, and the stability of the timing error is not guaranteed even after the correction. have. The results of the Kaiser window interpolation filter (FIG. 9B) or the sync interpolation filter (FIG. 9C) using the Kaiser window function are faster than the sync interpolation filter (FIG. 9A). .

The best embodiments have been disclosed in the drawings and the specification. Although specific terms have been used herein, they are used only for the purpose of describing the present invention and are not intended to limit the scope of the invention as defined in the claims or the claims. Therefore, those skilled in the art will understand that various modifications and equivalent other embodiments are possible from this. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

As described above, according to the present invention, by providing a Kaiser window interpolation filter or a Sink interpolation filter using a Kaiser window function, the frequency response is easy to access circular pulses due to the stop band of abrupt attenuation and the low side lobe. The timing error detector combines Gardner's method with the method of using the left and right slopes to achieve more stable timing control, resulting in faster convergence and reduced jitter.

Claims (6)

In the symbol timing recovery apparatus of the receiver for receiving a signal representing a continuous symbol, A sampler sampling the symbol by a predetermined multiple; A matched filter for removing residual sidebands of the signal output from the sampler and passing only a desired frequency band; A Kaiser window interpolation filter for interpolating and outputting the signal output from the matched filter using a Kaiser window function; A timing error detector for extracting a timing error of the signal output from the Kaiser window interpolation filter; A loop filter for reducing phase and frequency errors of the signal output from the timing error detector; And And a numerically controlled oscillator receiving the output signal of the loop filter and providing information necessary for the interpolation operation of the Kaiser window interpolation filter. The method of claim 1, wherein the Kaiser window function Where α is ( ), M is 60, ₀ ≦ t ≦ M, I ₀ (·) is a modified 0th order Bessel function, and β is 10. The receiver of claim 1, wherein the timing error detector finds the highest values among the values output from the Kaiser window interpolation filter, calculates the average of the highest and the highest of them, and outputs them as an output signal. Timing Restoration Device. In the symbol timing recovery apparatus of the receiver for receiving a signal representing a continuous symbol, A sampler sampling the symbol by a predetermined multiple; A matched filter for removing residual sidebands of the signal output from the sampler and passing only a desired frequency band; A sync interpolation filter for interpolating and outputting a signal output from the matched filter using a convolution of an impulse response for sync interpolation and a Kaiser window function; A timing error detector for extracting a timing error of the signal output from the sink interpolation filter; A loop filter for reducing phase and frequency errors of the signal output from the timing error detector; And And a numerically controlled oscillator receiving the output signal of the loop filter and providing information necessary for the interpolation operation of the sync interpolation filter. The impulse response function h _I (t) of claim 4, wherein Where α is ( ), M is 30, I ₀ (·) is a modified zero-order Bessel function, and β is 10. The receiver of claim 5, wherein the timing error detector finds the highest values among the values output from the Kaiser window interpolator, calculates the average of the highest positive and the highest negative among them, and outputs them as an output signal. Timing Restoration Device.