METHOD OF PROCESSING WCDMA SIGNAL TIMING OFFSET FOR SIGNAL ANALYZING EQUIPMENT

Abstract

Provided is a method of processing a WCDMA signal timing offset for a signal analyzer. The method includes estimating an integer multiple timing offset of WCDMA baseband sample data corresponding to an amount of at least one frame; generating a frequency domain signal which is time delayed corresponding to a fractional timing offset estimation resolution after generating the frequency domain signal by performing an FFT calculation on an already-known reference signal; converting each time-delayed frequency domain signal into a time domain signal by performing an IFFT calculation on each time-delayed frequency domain signal and calculating a correlation between an input signal from a position of the integer multiple timing offset and the time domain signal; and estimating a delay time leading to a maximum correlation as a fractional timing offset.

Claims

1. A method of processing a WCDMA signal timing offset for a signal analyzer, the method comprising: estimating an integer multiple timing offset of WCDMA baseband sample data corresponding to an amount of at least one frame; generating a frequency domain signal which is time delayed corresponding to a fractional timing offset estimation resolution after generating the frequency domain signal by performing an FFT calculation on an already-known reference signal; converting each time-delayed frequency domain signal into a time domain signal by performing an IFFT calculation on each time-delayed frequency domain signal and calculating a correlation between an input signal from a position of the integer multiple timing offset and the time domain signal; and estimating a delay time leading to a maximum correlation as a fractional timing offset.

2. The method of claim 1, wherein the frequency domain signal R(ω) is obtained by performing the FFT calculation on a PSCH r(t) through a following equation,
R(ω)=FFT[r(t),N.sub.ft] wherein N.sub.ft is a number of FFT samples.

3. The method of claim 2, wherein the PSCH r(t) is a signal RRC-filtered at a sampling rate which is twice a WCDMA chip rate.

4. The method of claim 1, wherein the frequency domain signal R(ω) is obtained by performing the FFT calculation on a CPICH r(t) through a following equation,
R(ω)=FFT[r(t),N.sub.ft] wherein N.sub.ft is a number of FFT samples.

5. The method of claim 4, wherein the CPICH r(t) is a signal RRC-filtered at a sampling rate which is twice a WCDMA chip rate.

6. The method of claim 1, wherein each delayed time domain signal is obtained by a following equation,
r(t−τ.sub.1̂*i)=IFFT[R(ω)e.sup.j*ω*τ.sup.1.sup.̂*i)] wherein τ.sub.1̂=T.sub.c/N.sub.1, I=0, . . . , N.sub.1, r′(t−τ.sub.1̂*i) is a conjugate complex number of r(t−τ.sub.1̂*i), and N.sub.1 is a fractional timing offset estimation resolution.

7. The method of claim 6, wherein the correlation y(t) is obtained by a following equation,
y(t)=Σ[x(t)*r′(t−τ.sub.1̂*i)].

8. The method of claim 7, further comprising compensating for the fractional timing offset by applying the estimated fractional timing offset to a fractional RRC filter as a compensation coefficient.

9. The method of claim 8, wherein the RRC filter is operated according to a following equation, ${\mathrm{RC}}_0\left(t\right)=\frac{\sin \left(\pi .Math.\frac{t}{T_C}.Math.\left(1-\alpha \right)\right)+4.Math.\alpha .Math.\frac{t}{T_C}.Math.\cos \left(\pi .Math.\frac{t}{T_C}.Math.\left(1+\alpha \right)\right)}{\pi .Math.\frac{t}{T_C}.Math.\left(1-{\left(4.Math.\alpha .Math.\frac{t}{T_C}\right)}^2\right)}$ wherein t=t−τ.sub.1*i.sub.maxc.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0030] FIG. 1 is a view illustrating a structure of a wireless frame of a WCDMA system including a synchronization channel.

[0031] FIG. 2 is a flowchart illustrating a method of processing a WCDMA signal timing offset for a signal analyzer according to an embodiment.

[0032] FIG. 3 is a graph illustrating a correlation obtained with a fractional timing offset estimation resolution of N.sub.1=128 according to an embodiment.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] Hereinafter, a method of processing a WCDMA signal timing offset for a signal analyzer according to a preferable embodiment will be described in detail with reference to accompanying drawings.

[0034] First, to receive and analyze a WCDMA signal, timing, frequency and phase synchronizations are required and specifically, at the sampling frequency of 30.72 MHz (the LTE clock frequency is used to compatible with an LTE signal analyzer), the timing, frequency and phase synchronizations are required to be satisfied within 0.1 sample, 1 Hz and 1°.

[0035] To this end, according to an embodiment, there is proposed a scheme capable of rapidly securing the timing synchronization by using FFT/IFFT while the sampling rate (2*Fc=7.68 MHz), which is about twice the chip rate Fc of the WCDMA system, is used.

[0036] FIG. 2 is a flowchart illustrating a method of processing a WCDMA signal timing offset for a signal analyzer according to an embodiment.

[0037] First, a signal analyzer according to an embodiment receives a WCDMA signal. The signal analyzer down-converts the WCDMA signal into an analog baseband I/Q signal and converts the analog baseband I/Q signal into digital data again. For example, the signal analyzer converts the analog baseband I/Q signal into I/Q sample data of a sampling rate of 30.72 MHz and a resolution of 16-bit and captures the I/Q sample data. In the following description, the I/Q sample data may not be separately explained. In this case, in step S10, the captured sample data are matched-filtered. In this case, an RRC (Root Raised Cosine) filtering operation is performed.

[0038] As well known in the art, since the RRC filtering demodulates a signal by utilizing spectrum aliasing characteristics in digital domain, a receiver is required to perform over sampling. In this case, since the minimum over sampling rate is 2, after the original sampling rate is decimated to 4:1, the RRC filtering is performed to use the minimum over rate.

[0039] In this case, the first step described above, that is, the PSC synchronization is performed, and at the same time, the timing and frequency offset compensations are performed. To this end, in the step S10, the frequency offset F.sub.off is set into 0, and then, in step S20, a sampling value in units of slots is calculated. Preferably, to more stably perform PSC detection, after the sample values in units of slots in one wireless frame, in which total 15 slots exist, are calculated, the average value of them is used.

[0040] Next, in step S30, the correlation between the sample values in units of slots obtained in step S20 and a reference PSC sample (sampled at twice of the chip rate) defined in the standard is calculated to search for the maximum correlation value C.sub.max and the location thereof. Then, in step S40, the maximum correlation value C.sub.max and the location thereof are matched with the corresponding frequency offset F.sub.off to be stored.

[0041] Next, in steps S50 and S60, in a state that a new frequency offset is set by increasing or decreasing a current frequency offset by a predetermined increment Δf (increased in once, or decreased in another time), the steps S20 to S60 are repeated until the new frequency offset reaches the predetermined maximum frequency offset ±F.sub.max. In this case, preferably, the increment Δf of the frequency offset is set as 50 Hz which is half of 100 Hz in consideration of the time length 10 ms of the wireless frame and the maximum frequency offset F.sub.max is set as 750 Hz (15*100 Hz*½) in consideration of the number of PSCs of 15 included in one wireless frame.

[0042] As described above, in the state that the maximum correlation value C.sub.max with respect to all frequency offsets and the location thereof are calculated and stored, in step S70, after confirming the maximum correlation value C.sub.max and the location, a frequency offset matched with the maximum correlation value C.sub.max and the location is estimated as the frequency offset of a received signal. In addition, the integer multiple timing offset of the PSC is estimated based on the sample location of the maximum correlation value C.sub.max.

[0043] Meanwhile, as expressed as the following Equation 1 according to an embodiment, the correlation calculation of step S30 may be performed in frequency domain by using the Fourier transform characteristics. When the correlation calculation is performed in frequency domain, an FFT algorithm may be applied thereto. As a result, as the number of samples is increased, the calculation may be effectively performed with a small amount of computation.

[0044] The correlation calculation for estimating a timing offset may be performed by following Equation 1.

[00002]$\begin{array}{cc}\begin{array}{c}R\left(\tau \right)=.Math.E\left[x\left(t+\tau \right)\star {s\left(t\right)}^{\prime }\right]\\ =.Math.F^{-1}.Math.\left\{F\left[x\left(t+\tau \right)\right]\star F\left[s\left(t\right)\right]\right\}\end{array}\hspace{1em}& \left[\mathrm{Equation}.Math..Math.1\right]\end{array}$

[0045] In the Equation 1, R(τ) is a correlation value, x(t+τ) is an input signal delayed for time τ (corresponding to one sample when the embodiment is based on 512 samples), s(t)′ is a reference signal, that is, a conjugate complex number of PSC having 512 samples, F represents an FFT calculation, and F.sup.−1 represents an IFFT calculation.

[0046] The following Table 1 is a table of comparing an amount of computation according to the correlation calculation scheme of the embodiment with that of the related art. According to the embodiment, as shown in Table 1, when based on 512 samples, the amount of computation is reduced to 57 times. When based on 4096 samples, the amount of computation is reduced to 341 times

TABLE-US-00001 TABLE 1 Related art Embodiment Remarks 512 * 512 512 * log.sub.2(512)  57 times 4096 * 4096 4096 * log.sub.2(4096) 341 times

[0047] As described above, when the frequency offset estimation and the integer multiple timing offset estimation are completed, in step S90, a fractional timing offset is estimated based on the location of the estimated integer multiple timing offset. To this end, FFT characteristics expressed as following Equations 2 and 3 are used.

FFT[r(t−t.sub.0)]=FFT[r(t)]e.sup.j*ω*(−t.sup.0.sup.)=R(ω)e.sup.j*ω*(−t.sup.0.sup.)  [Equation 2]

r(t−t.sub.0)=IFFT[R(ω)e.sup.j*ω*(−t.sup.0.sup.)]  [Equation 3]

[0048] When using Equations 2 and 3, if a signal R(w) is previously generated and the sample interval to is set as T.sub.c(=1/F.sub.c)/N.sub.1, only a required time delay signal may be rapidly generated through the IFFT calculation, where N.sub.1 is a fractional timing offset estimation resolution which may be predetermined as 128 in the embodiment. When the location of the maximum correlation value is determined by calculating the correlation between the time-delayed signal and the input signal, the location may be immediately estimated as the fractional timing offset. This relation may be expressed as follows with PSCH as a reference signal.

[0049] First, a frequency domain signal R(ω) for PSCH r(t) previously defined in the standard is generated according to following Equation 4.

R(ω)=FFT[r(t),N.sub.ft]  [Equation 4]

[0050] In the Equation 4, N.sub.ff is the number of samples of FFT. For example, N.sub.ff may be set as 4096. In addition, a PSCH signal, which is RRC filtered at the sampling rate of twice the chip rate, is used as r(t).

[0051] Next, the correlation between the input signal x(t) RRC-filtered at the sampling rate of twice the chip rate and r(t−τ.sub.1̂*i) is calculated. r(t−τ.sub.1̂*i) and the correlation may be obtained through following Equations 5 and 6.

r(t−τ.sub.1̂*i)=IFFT[R(ω)e.sup.j*ω*τ.sup.1.sup.̂*i)],τ.sub.1̂=T.sub.c/N.sub.1,i=0, . . . ,N.sub.1  [Equation 5]

y(t)=Σ[x(t)*r′(t−τ.sub.1̂*i)]  [Equation 6]

[0052] In Equation 6, r′(t−τ.sub.1̂*i) is the conjugate complex number of r(t−τ.sub.1̂*i). The location i(i.sub.maxc) at which the correlation obtained through Equation 6 is maximized is obtained. This location is the fractional timing offset.

[0053] FIG. 3 is a graph illustrating a correlation obtained with a fractional timing offset estimation resolution of N.sub.1=128 through Equation 6 according to an embodiment. It may be understood through FIG. 3 that the fractional timing offset is about 60τ.sub.1.

[0054] When the fractional timing offset is estimated as described above, the fractional timing offset is applied to the RRC filter as a compensation coefficient, so that the fractional timing offset may be exactly compensated.

[00003]$\begin{array}{cc}{\mathrm{RC}}_0\left(t\right)=\frac{\sin .Math.\left(\pi .Math.\frac{t}{T_C}.Math.\left(1-\alpha \right)\right)+4.Math..Math.\alpha .Math.\frac{t}{T_C}.Math.\cos \left(\pi .Math.\frac{t}{T_C}.Math.\left(1+\alpha \right)\right)}{\pi .Math.\frac{t}{T_C}.Math.\left(1-{\left(4.Math.\alpha .Math.\frac{t}{T_C}\right)}^2\right)}& \left[\mathrm{Equation}.Math..Math.7\right]\end{array}$

[0055] The Equation 7 is a mathematical expression of the RRC filtering defined in 3GPP TS.25.101/104. In the Equation 7, when t is replaced with t−τ.sub.1*i.sub.maxc, the fractional timing offset may be exactly compensated.

[0056] Meanwhile, since the fractional timing offset may not be exactly detected through the PSCH due to the phase offset, if CPICH is used, the fractional timing offset may be more exactly detected. Even in this case, only the reference signal is replaced with the CPICH and the fractional timing offset may be estimated through the same method as that described above.

[0057] The method of processing a WCDMA signal timing offset for a signal analyzer described with reference to accompanying drawings in this disclosure is for an illustrative purpose only, and the embodiment is not limited thereto. Thus, it should be understood that numerous other modifications and embodiments can be devised by those skilled in the art within the spirit and scope of the embodiment and they will fall within the scope of the embodiment.