TWI403132B - Demodulation module, signal analyzer and signal analyzing method - Google Patents

Description

Demodulation module, signal analysis device and signal analysis method

The present invention relates to a method for estimating carrier frequency drift, and more particularly to a carrier frequency drift estimation method suitable for multipath transmission channels.

In wireless communication systems, carrier frequency offset (CFO) may occur between the transmitting end and the receiving end due to inaccurate oscillation frequency and Doppler effect. Especially in the Orthogonal Frequency Division Multiplexing (OFDM) system, the impact of carrier frequency drift is more serious. Since the transmission system of orthogonal frequency division multiplexing is a multi-carrier system, it is very sensitive to carrier frequency drift, and the carrier frequency drift will destroy the orthogonality between subcarriers in the orthogonal frequency division multiplexing system, thus causing mutual carriers. Inter-Carrier Interference (ICI) makes the system performance of the orthogonal frequency division multiplexing system worse, and the error rate increases. Therefore, how to accurately estimate carrier frequency drift to solve inter-carrier interference (ICI) becomes an important issue that needs to be solved in orthogonal frequency division multiplexing systems.

According to an embodiment of the invention, a demodulation module includes an analog to digital converter, a baseband mixer, a timing recovery device, a signal analysis device, and a decoder. The analog to digital converter converts an analog IF signal to output a digital IF signal. The baseband mixer receives the digital intermediate frequency signal, and down-converts the digital intermediate frequency signal according to a carrier frequency to generate a baseband signal, wherein the baseband mixer further adjusts the carrier according to a first feedback control signal Frequency to compensate for carrier frequency drift of the carrier frequency. The timing recovery device resamples the baseband signal according to a second feedback control signal. The signal analysis device receives the baseband signal, analyzes the correlation between the baseband signal and a predetermined virtual noise sequence to obtain a complex correlation result, and enhances the correlation results to generate a correlation sequence, and according to the correlation The sequence generates the first feedback control signal and the second feedback signal. The decoder is configured to decode an output signal of the timing recovery device to generate a decoded output signal.

According to another embodiment of the present invention, a signal analysis apparatus includes a correlation calculation device, a symbol margin detection device, and a carrier frequency drift estimation device. The correlation computing device receives an input signal, delays the input signal according to a plurality of different delay amounts to obtain a complex delayed signal, and calculates a correlation between the delayed signals and a predetermined virtual noise sequence to obtain a complex correlation calculation result, And generating a correlation sequence based on the results of the correlation calculations. The symbol margin detecting means detects one of the sampling points having the largest correlation according to the correlation sequence, and generates a symbol margin indicating signal according to the correlation value of the sampling point. The carrier frequency drift estimating means estimates a carrier frequency drift amount based on the symbol marginal indication signal.

According to another embodiment of the present invention, a signal analysis method includes: receiving an input signal, wherein the input signal includes at least one data material and a virtual noise sequence; delaying the input signal according to a plurality of different delay amounts to obtain a complex number Delaying a signal; calculating a correlation between the delayed signals and a predetermined virtual noise sequence to obtain a complex correlation calculation result; generating a correlation sequence based on the correlation calculation results; and estimating the input signal according to the correlation sequence One carrier frequency drift amount.

In order to make the manufacturing, operating methods, objects and advantages of the present invention more apparent, the following detailed description of the preferred embodiments and the accompanying drawings

Example:

The digital terrestrial multimedia/television broadcasting (DTMB) system is a newly developed digital television broadcasting system specification in recent years. The DTMB technology involves filling a known Pseudo Noise (PN) sequence into the guard interval to protect the data signal and further identifying the starting position of the data. Figure 1a-1b shows the data structure of DTMB. As shown, the data 100 and 101 are the data of a DTMB of a frame, wherein the data 101 is the result of delaying the data 100 by D sampling points. The data 100 and 101 respectively include a portion of the PN sequence (labeled as PN code) and a portion of the data data (labeled as Data), wherein the tail portion of the PN sequence is repeatedly added to the beginning of the PN sequence (labeled as Pre PN). And the header of the PN sequence is repeatedly added to the end of the PN sequence (labeled Post PN). For example, the DTMB can use the PN420 specification, in which a 420-point PN sequence can be added before the data, including a complete 255-point PN sequence, and 82 points are added to the first half of the PN sequence to form the Pre PN portion, and 83 of them are added. The Post PN portion is formed at the second half of the PN sequence, thus forming a sequence of 420 points in length.

Since the receiving end knows the content of the PN sequence added by the transmitting end, the conjugate complex number of the DTMB data (for example, the data 100 shown in FIG. 1) received by the receiver is delayed by D sampling with the DTMB data. The result of the point (for example, the data 101 shown in FIG. 1) is multiplied, and then the operation result is correlated with the PN sequence stored at the local end of the receiver to find the correlation with the greatest correlation. Sample point max ( Corr _D ( n )) to achieve data synchronization. The sampling point with the largest correlation represents that the PN sequence stored at the local end matches the received DTMB data at this sampling point. Once the matching of the PN sequence is completed, the starting position of the PN sequence in the received signal can be obtained. Then, the data data synchronization can be achieved by recognizing the starting position of the data data by the position of the PN sequence. Examples of computational correlations can be further referenced by Ling-Long Dai et al. in November 2008 at the Institute of Electrical and Electronic Engineers (IEEE) International Conference on Communications Systems (ICCS). The published technical literature is entitled "A new frequency synchronization algorithm in the TDS-OFDM systems".

In addition to achieving the timing synchronization of the data, the receiver can further estimate the carrier frequency drift Δ f generated by the transmission channel by using the obtained maximum correlation calculation result max( Corr _D ( n )), wherein the carrier frequency drifts Δ f The estimation method is as follows:

Where fs is a sampling frequency (Fig. 4, analog to the sampling frequency of the digital converter 403).

However, in a multi-path transmission channel environment, since the signal received by the receiving end can contain DTMB data from more than one transmission path, in addition to the theoretical maximum correlation, the correlation operation should be performed. In addition to the Desired term, there are a number of additional False terms, and even the relevance of the error term is equal to or greater than the requested term. The term is the most closely matched sampling point of the PN sequence (with maximum correlation), that is, the result of multiplying the conjugate complex number of the theoretically received DTMB data with the delayed version of the DTMB data and the PN sequence at the local end. The sampling point that should have the greatest correlation after the correlation operation represents that at this sampling point, the PN sequence at the local end is synchronized with the PN sequence in the received DTMB data, and the other sampling points are error items. Figure 2 shows the correlation operation results of the received signals in the multipath transmission channel environment, where the horizontal axis represents the sampling point and the vertical axis represents the calculation result of the correlation. It is assumed that the transmission channel includes two paths, so the calculation result of correlation calculation It should include two more relevant items 201. However, once the correlation of the error term 202 is higher than the expected term due to the channel response of the multipath, it will result in an erroneous result when looking for the maximum correlation sampling point max( Corr _D ( n )). In this way, not only the erroneous data data is synchronized, but also the wrong carrier frequency drift estimation result.

The 3a-3b diagram shows the estimation of the carrier frequency drift in the multipath transmission channel environment and the corresponding compensation result, wherein the horizontal axis represents the frame index and the vertical axis represents the method according to the equation (1). The result of carrier frequency drift estimation is Δ f . In more detail, Fig. 3a is a result of estimating the carrier frequency drift based on the estimated term, and Fig. 3b is a result of estimating the carrier frequency drift based on the error term. As shown in Figure 3a, if the correct term can be found, the carrier frequency drift Δ f can be accurately estimated. In this example, when the frame index is 0, the actual carrier estimated based on the found term is obtained. The frequency drift is 100kHz. The carrier is then estimated frequency shift Δ f to compensate for carrier frequency compensation over time, such that the actual carrier frequency may drift converges to 0kHz. However, if the carrier frequency drift estimate of the error term is not possible to estimate the actual carrier frequency drift Δ f. That is, if to compensate for the error term estimated carrier frequency drift Δ f, the actual carrier frequency drift can not converge to 0kHz. As shown in Figure 3b, when the frame index is 0, the carrier frequency estimated according to the error term drifts to 200 kHz (instead of the actual carrier frequency drifting 100 kHz). If the carrier frequency is shifted by 200 kHz to compensate, after a period of compensation. So that the actual carrier frequency drifts to -100 kHz instead of converge to 0 kHz. In view of this, the present invention proposes a new carrier frequency drift estimation method to accurately estimate the correct request. This carrier frequency drift estimation method is not only suitable for a single path transmission channel environment, but also suitable for a multipath transmission channel environment. Used to accurately complete data synchronization and accurately estimate carrier frequency drift.

Figure 4 is a diagram showing a receiver 400 in accordance with an embodiment of the present invention. The receiver 400 includes a tuner 401 and a demodulation module 402. The tuner 401 converts the RF signal S _RF received by the antenna into an analog intermediate frequency signal. The demodulation module 402 can be integrated into a demodulator IC for receiving an analog intermediate frequency signal from the tuner 401 and demodulating the signal to produce an output signal S _OUT (Transport Stream).

According to an embodiment of the present invention, the demodulation module 402 can include an analog to digital converter (ADC) 403, a baseband mixer 404, a timing recovery device 405, an equalizer 406, and decoding. The device 407 and the signal analysis device 408. The analog to digital converter 403 is configured to sample the analog intermediate frequency signal according to a sampling frequency (equations (1), fs ) to output a digital intermediate frequency signal. The baseband mixer 404 downconverts the digital intermediate frequency signal according to a carrier frequency to generate a digital baseband signal S _B . The signal analysis device 408 is configured to analyze the characteristics of the virtual noise sequence in the digital baseband signal S _B , and generate the feedback control signals S _C and S _T to the fundamental frequency mixer 404 and the timing respectively according to the characteristics of the virtual noise sequence. The recovery device 405, wherein the baseband mixer 404 changes the frequency of the carrier according to the information provided by the feedback control signal S _C to compensate for the carrier frequency offset, and the timing recovery device 405 provides the feedback control signal S _T according to The timing synchronization information resamples the digital baseband signal S _B to restore the signal to the timing synchronized with the transmitting end. The equalizer 406 equalizes the output signal of the timing recovery device 405 to compensate for the frequency response of the transmission channel to remove the effects of the transmission channel. Equalizer 406 is an optional element in this embodiment. The decoder 407 finally decodes the equalized signal to output the signal decoded signal S _OUT .

According to an embodiment of the present invention, the signal analysis device 408 can include a correlation calculation device 411, a symbol margin detection device 412, a timing error estimation device 413, and a carrier frequency drift estimation device 414. Correlation computing means 411 for the conjugated complex digital and digital-yl group obtained from the baseband frequency mixer 404 frequency signal S _B of different delayed version of the signal S _B for the multiplication result to obtain a complex operation, Further, the complex operation result and the PN sequence stored in the correlation calculating means 411 are respectively subjected to a correlation operation to obtain a complex correlation calculation result, and a correlation sequence Corr(n) is generated based on the complex correlation calculation result. ). The symbol margin detecting device 412 receives the correlation sequence from the correlation computing device 411, detects the most relevant sampling point in each frame, and records the correlation value and position of the sampling point to generate a symbol marginal indication. Signal S _IND . The timing error estimating device 413 is coupled to the symbol margin detecting device 412 for determining the position and correlation value of the sampling point having the greatest correlation indicated by the symbol margin indicating signal S _IND (ie, the amplitude of the sampling point) The information generates a feedback control signal S _T such that the timing recovery means 405 can resample the digital baseband signal S _B according to the feedback control signal S _T for returning to the timing synchronized with the transmitting end. The carrier frequency drift estimating device 414 is also coupled to the symbol margin detecting device 412 for correlating the correlation value of the sampling point having the greatest correlation indicated by the symbol margin indicating signal S _IND (ie, the amplitude of the sampling point) ) through methods such as formula (1) according to the estimated carrier frequency drift, and generate a feedback control signal S _C, such baseband mixer 404 can be transported in accordance with the estimated value Δ f feedback control signal S _C is provided by the carrier frequency Compensate for carrier frequency offset.

As mentioned above, since the signals received from the multipathed transmission channel environment produce a correlation item with a large correlation when calculating the correlation (as shown in Fig. 2), the sampling point for finding the maximum correlation is caused. Max( Corr _D ( n )) results in erroneous results, resulting in erroneous data synchronization, as well as incorrect carrier frequency drift estimation results (as shown in Figure 3). Therefore, according to an embodiment of the present invention, the correlation computing device 411 further enhances the correlation of the found items when calculating the correlation, so that the items found in the obtained correlation calculation result may have the largest correlation value. Thus, the symbol margin detecting means 412 can accurately find the position of the requested item, and the timing error estimating means 413 and the carrier frequency drift estimating means 414 can correctly obtain the estimated values of the timing synchronization information and the carrier frequency drift.

5a-5c are diagrams showing correlation calculation results for a multipath transmission channel environment in accordance with an embodiment of the present invention. The concept of the reinforcement term described in one embodiment of the present invention can be clearly understood from Figures 5a-5c. Figure 5a shows the result of multiplication between the conjugate complex number of the received DTMB data and the version of the D1 sample point (as shown in Figure 1) of the DTMB data, and then the result of the operation and the local end of the receiver. The result of the operation of the correlation performed by the stored PN sequence. As shown in the figure, it is assumed that the sampling points n2 and n4 are the obtained items, that is, the sampling points that are theoretically the most matching of the PN sequence, and the sampling points n1, n3 and n5 are the error items. It can be seen from Fig. 5a that due to the effect of the channel response of the multiple paths, the correlation of the error term at the sampling point n1 is quite close to the correlation of the term at the sampling point n2, and even greater than the sampling point. The correlation of the terms of n4, so if the correlation calculation result as shown in Fig. 5a is not able to find the exact terms (at sampling points n2 and n4) for timing synchronization and frequency offset estimation And thus an error occurred. Figure 5b shows another correlation operation result. This result is based on the multiplication of the conjugate complex number of the received DTMB data with the delay of the D2 data points of the DTMB data (as shown in Figure 1). The result of this operation is calculated by performing a correlation operation with the PN sequence stored at the local end of the receiver. It can be seen from Fig. 5b that due to the effect of the channel response of the multiple paths, the correlation of the error items located at the sampling point n5 is largely at the correlation of the terms of the sampling point n2, and is greater than the sampling point n4. The correlation of the terms is sought, so that the requested term (at sampling points n2 and n4) cannot be distinguished. Therefore, if timing synchronization and frequency offset estimation are performed based on the correlation calculation result as shown in Fig. 5b, an error occurs.

However, as shown in Figures 5a and 5b, since the position at which the error term occurs changes with the length of the delay D, and the position at which the term occurs does not change with the delay D, In an embodiment of the invention, the correlation calculation result obtained by accumulating different delays is used to strengthen the correlation of the obtained items, so that the correlation calculation result obtained by the obtained item can be clearly distinguished from the error item, such as As shown in Fig. 5c, the correlation between the accumulated terms n2 and n4 is significantly different from the error term, so that the signal analyzing means 408 can accurately find the desired result based on the enhanced correlation calculation result. The correct position of the term, which in turn leads to an accurate estimate of timing synchronization and carrier frequency drift.

Figure 6 is a diagram showing a correlation computing device in accordance with an embodiment of the present invention. According to an embodiment of the present invention, the correlation computing device 611 may include two or more sets of computing modules. In order to clearly illustrate the concept of the present invention, FIG. 6 shows two sets of computing modules 612 and 613. It should be noted that the present invention can also use more than two sets of computing modules, and is not limited to the use of two sets of computing modules. Anyone skilled in the art can make some changes without departing from the spirit and scope of the present invention. The scope of protection of the present invention is defined by the scope of the appended claims.

According to an embodiment of the invention, each computing module can include two sets of First In First Out (FIFO) registers respectively. The lengths of the first set of FIFO registers are respectively designed according to different delay requirements. For example, the length of the FIFO register 621 can be designed as D1 to delay the input digital baseband signal S _B by D1 samples. The length of the FIFO register 631 can be designed as D2 to delay the input digital baseband signal S _B by D2 sampling points. After the conjugate complex unit (conjugator, simply referred to as conj) 622 and 632 converts the input digital baseband signal S _{B into} a complex conjugate, the conjugate complex number of the digital baseband signal S _B is delayed by D1. The version (or D2) is multiplied by 633 by multipliers 623. The multiplied result is further input to the second set of FIFO registers 624 and 634. The second set of FIFO registers 624 and 634 are located in correlation calculation units 625 and 635, respectively, to aid in the operation of the correlation, and thus the lengths can be designed according to the required correlation lengths C1 and C2, respectively. The result of multiplying the conjugate complex number of the digital baseband signal S _B by the delayed version is stored in the correlation computing units 625 and 635 through the second set of FIFO registers 624 and 634 and the PN stored at the local end of the receiver, respectively. The sequence is multiplied point by point, and the multiplication results are respectively accumulated to obtain correlation calculation results Corr_D1(n) and Corr_D2(n). Finally, the correlation calculation results Corr_D1(n) and Corr_D2(n) obtained by the adder 614 are accumulated to further enhance the correlation of the found terms to output the correlation sequence Corr(n).

As shown in Figures 5a-5c, the position at which the error term occurs will change with the length of the delay D, and the position at which the term occurs will not change with the delay D, and different delays will be used. The resulting correlation calculation results can enhance the correlation of the items sought, so that the correlation of the items found in the obtained correlation sequence can be clearly distinguished from the error items. Therefore, the symbol margin detecting means 421 can accurately find the sampling point having the greatest correlation, and generate the feedback control signals S _C and S _T to the fundamental frequency mixer 404 and the timing recovery means 405 as described above for performing Carrier frequency drift estimation and timing recovery.

Figure 7 is a diagram showing a signal analysis method according to an embodiment of the present invention. First, an input signal is received (step S701), wherein the input signal may be a DTMB signal as shown in FIG. 1a, including a virtual noise sequence (PN Code) and a data data (Data), and the input signal may be The baseband mixer 404 performs the down-converted baseband signal. Then, the input signal is delayed according to a plurality of different delay amounts (step S702), for example, the input signals are delayed according to two or more different delay amounts to respectively obtain input signals having different delay amounts. Next, the correlation between the input signals having the different delay amounts and a virtual noise sequence is calculated to obtain a complex correlation calculation result (step S703). Next, a correlation sequence is generated based on the correlation calculation results (step S704). Finally, a starting position of a data material and a carrier frequency drift amount are estimated based on the correlation sequence (step S705). Referring to the structure of the correlation computing device as shown in FIG. 6, step S702 and step S703 may further include: multiplying the conjugate complex number of the input signal by the input signals having different delay amounts to obtain a complex phase Multiplying the result; and calculating a correlation of the virtual noise sequence with the multiplied result to obtain the correlation calculation results, such as the delay amounts D1 and D2 shown in FIG. 6, and the correlation calculation result Corr_D1 ( n) with Corr_D2(n). In step S704, the correlation sequence Corr(n) is generated by accumulating Corr_D1(n) and Corr_D2(n).

In addition, referring to the correlation computing device structure as shown in FIG. 4, step S705 may further include: detecting one of the correlation sequences Corr(n) having the largest correlation; according to the sampling point having the largest correlation A position and a correlation value generate a symbol marginal indication signal S _IND ; and a starting position of the data data and a carrier frequency drift amount are estimated based on the symbol marginal indication signal.

The present invention has been described above with reference to the preferred embodiments thereof, and is not intended to limit the scope of the present invention, and the invention may be modified and modified without departing from the spirit and scope of the invention. The scope of the invention is defined by the scope of the appended claims.

100, 101. . . data

201. . . Item requested

202. . . Error item

400. . . Receiver

401. . . tuner

402. . . Demodulation module

403. . . Analog to digital converter

404. . . Base frequency mixer

405. . . Timing recovery device

406. . . Equalizer

407. . . decoder

408. . . Signal analysis device

411, 611. . . Correlation computing device

412. . . Symbol edge detection device

413. . . Timing error estimation device

414. . . Carrier frequency drift estimation device

612, 613. . . Computing module

621, 631, 624, 634, FIFO. . . FIFO register

622, 632. . . Conjugate complex unit

623, 633. . . Multiplier

625, 635. . . Correlation calculation unit

Corr(n), Corr_D1(n), Corr_D2(n). . . Correlation

D. . . delay

Data. . . data

PN, PN code, Pre PN, Post PN. . . Virtual noise sequence

S _B , S _C , S _IND , S _OUT , S _RF , S _T . . . signal

Figure 1a-1b shows the data structure of digital television broadcast data.

Figure 2 shows the results of the correlation operation of the received signals in the multipath transmission channel environment.

Figure 3a-b shows the estimation of carrier frequency drift in the multipath transmission channel environment and the corresponding compensation results.

Figure 4 is a diagram showing a receiver in accordance with an embodiment of the present invention.

5a-5c are diagrams showing correlation calculation results for a multipath transmission channel environment in accordance with an embodiment of the present invention.

Figure 6 is a diagram showing a correlation computing device in accordance with an embodiment of the present invention.

Figure 7 is a diagram showing a signal analysis method according to an embodiment of the present invention.

400. . . Receiver

401. . . tuner

402. . . Demodulation module

403. . . Analog to digital converter

404. . . Base frequency mixer

405. . . Timing recovery device

406. . . Equalizer

407. . . decoder

408. . . Signal analysis device

411. . . Correlation computing device

412. . . Symbol edge detection device

413. . . Timing error estimation device

414. . . Carrier frequency drift estimation device

S _B , S _C , S _IND , S _OUT , S _RF , S _T . . . signal

Claims (15)

A demodulation module includes: a analog to digital converter that converts an analog IF signal to output a digital intermediate frequency signal; a fundamental frequency mixer that receives the digital intermediate frequency signal and is based on a carrier frequency Down-clocking the digital intermediate frequency signal to generate a baseband signal, wherein the baseband mixer further adjusts the carrier frequency according to a first feedback control signal to compensate for carrier frequency drift of the carrier frequency; a timing recovery device, Resampling the baseband signal according to a second feedback control signal; receiving a signal of the baseband signal, analyzing the correlation between the baseband signal and a predetermined virtual noise sequence to obtain a complex correlation result, and enhancing The correlation results are used to generate a correlation sequence, and the first feedback control signal and the second feedback signal are generated according to the correlation sequence; and a decoder for decoding an output signal of the timing recovery device To generate a decoded output signal. The demodulation module of claim 1, further comprising: a first equalizer, equalizing the output signal of the timing recovery device before the decoder decodes the output signal of the timing recovery device, Wherein the decoder decodes an output signal of one of the equalizers. The demodulation module of claim 1, wherein the signal analysis device comprises: a correlation calculation device, receiving the baseband signal, delaying the baseband signal according to a plurality of different delay amounts to obtain the same Delaying the signals, and calculating correlations between the delayed signals and the predetermined virtual noise sequence to obtain the correlation calculation results, and accumulating the correlation calculation results to obtain the correlation sequence; a symbol margin detection The device detects one of the sampling points having the largest correlation according to the correlation sequence, and generates a symbol marginal indication signal according to a position of the sampling point and a correlation value; a carrier frequency drift estimating device according to the symbol The margin indication signal estimates a carrier frequency drift amount, and generates the first feedback signal according to the carrier frequency drift amount; and a timing error estimating device that generates the second feedback signal according to the symbol margin indication signal. The demodulation module of claim 3, wherein the carrier frequency drift estimating means estimates the carrier frequency drift amount based on the correlation value of the sampling point. The demodulation module of claim 3, wherein the timing error estimating device generates the second feedback signal according to the position of the sampling point and the correlation value. The demodulation module of claim 3, wherein the correlation computing device comprises: a plurality of computing modules, wherein each computing module comprises: a first in first out register, for temporarily storing an established a delay amount of the baseband signal; a conjugate complex unit for performing conjugate complex conversion; a multiplier coupled to the FIFO register and the conjugate complex unit for One output of the first-in first-out register and one output of the conjugate complex unit perform a multiplication operation to obtain a multiplication result; and a correlation calculation unit for calculating the multiplication result and the predetermined virtual noise Correlation of the sequence to obtain the correlation calculation result; and an adder for summing the correlation calculation result output by the operation module to generate the correlation sequence. A signal analyzing device for estimating carrier frequency drift, comprising: a correlation computing device, receiving an input signal, delaying the input signal according to a plurality of different delay amounts to obtain a complex delayed signal, and calculating the delayed signal and a predetermined Correlation of virtual noise sequences to obtain complex correlation calculation results, and generating a correlation sequence according to the correlation calculation results; a symbol margin detection device, which detects the maximum correlation according to the correlation sequence a sampling point, and generating a symbol marginal indication signal according to a correlation value of the sampling point; and a carrier frequency drift estimating means for estimating a carrier frequency drift amount according to the symbol marginal indication signal. The signal analysis device of claim 7, wherein the correlation calculation device accumulates the correlation calculation results to generate the correlation sequence. The signal analysis device of claim 7, wherein the symbol margin detecting device generates the symbol margin indication signal according to the position of the sampling point, and further includes: a timing error estimating device, according to the The symbol marginal indication signal estimates the starting position of one of the input signal data. The signal analysis device of claim 7, wherein the correlation computing device comprises: a plurality of computing modules, wherein each computing module comprises: a first-in first-out register for temporarily storing a predetermined delay amount The input signal; a conjugate complex unit for converting the input signal into a conjugate complex number; a multiplier coupled to the FIFO register and the conjugate complex unit for using the first in first out One of the registers outputs a multiplication operation with one of the conjugate complex units to obtain a multiplication result; and a correlation calculation unit calculates a correlation between the multiplication result and the predetermined virtual noise sequence And obtaining an approximation calculation result; and an adder for summing the correlation calculation result output by the operation module to generate the correlation sequence. A signal analysis method for estimating carrier frequency drift includes: receiving an input signal, wherein the input signal includes at least one data data and a virtual noise sequence; delaying the input signal according to a plurality of different delay amounts to obtain a complex delay a signal; calculating a correlation between the delayed signals and a predetermined virtual noise sequence to obtain a complex correlation calculation result; generating a correlation sequence based on the correlation calculation results; and estimating the input signal according to the correlation sequence A carrier frequency drift amount. The signal analysis method of claim 11, wherein the step of generating the correlation sequence comprises accumulating the correlation calculation results to generate the correlation sequence. The signal analysis method according to claim 11, wherein the estimating step of the carrier frequency drift comprises: detecting a sampling point having a maximum correlation according to the correlation sequence; and correlating according to the sampling point The value is estimated by the carrier frequency drift. The signal analysis method of claim 13, further comprising: estimating a starting position of the data according to a position of the sampling point and the correlation value. The signal analysis method of claim 11, wherein the calculating step of the correlation calculation result further comprises: multiplying a conjugate complex number of the input signal by the delay signal to obtain a complex multiplication result; And calculating a correlation of the predetermined virtual noise sequence with the multiplied results to obtain the correlation calculation results.