CN106713196A - Receiving circuit capable of estimating frequency offset and related method - Google Patents

Abstract

The invention provides a receiving circuit capable of estimating frequency offset and a related method. The front-end circuit receives a far-end signal to generate a receiving signal. The calculation circuit includes: the first power calculation module calculates the power of an exponent of the received signal to generate a high-order signal; a frequency domain conversion module, which performs frequency domain conversion on the high-order signal to generate a frequency spectrum; a peak searching module for searching the peak value of the spectrum amplitude to generate a peak coordinate value reflecting the peak generation frequency; an offset estimation module adds the peak coordinate value and a compensation value to generate a sum, divides the sum by a first divisor to generate a remainder, subtracts the remainder from the compensation value to generate a difference, and divides the difference by a second divisor to generate an offset estimation value, which reflects the frequency offset.

Description

Receiving circuit capable of estimating frequency offset and related method

technical field

The present invention relates to a receiving circuit and related method capable of estimating frequency offset, especially a receiving circuit and related method capable of correctly estimating frequency offset under multipath interference.

Background technique

Receiving remote signals has become one of the most commonly used functions in modern information devices. In order to receive the wireless or wired remote signal sent by the remote transmitting circuit, the information device will be equipped with a receiving circuit. The transmitting circuit upconverts and modulates a baseband signal into a radio frequency transmitting signal with a local frequency of the transmitting end, and then transmits and propagates it. The transmitted signal propagates to the receiving circuit to become a radio frequency remote signal, which is received by the receiving circuit, and the remote signal is down-converted and demodulated at a local frequency of the receiving end to obtain the base frequency signal.

However, the local frequency of the transmitter and the local frequency of the receiver cannot be completely matched, and there will be a difference between the two, that is, a frequency offset. Therefore, the transmitting circuit needs to estimate the frequency offset in order to correctly perform down-conversion demodulation and signal restoration. Furthermore, when the transmitted signal propagates from the transmitting circuit to the receiving circuit, it will encounter various propagation interferences, including multipath interference. Propagation interference can affect the estimation of frequency offset and make the estimation inaccurate.

Contents of the invention

In order to prevent frequency offset estimation from being affected by propagation interference, one object of the present invention is to provide a receiving circuit capable of estimating frequency offset, which may include a front-end circuit and a calculation circuit. The front-end circuit is used to receive a remote signal yRF(t) from a transmitting circuit, and generate a received signal y(t) accordingly. The calculation circuit is coupled to the front-end circuit, which may include a power calculation module, a frequency domain conversion module, a peak search module and an offset estimation module. The power calculation module can calculate a power of an exponent P of the received signal to generate a high-order signal y ^p (t). The frequency domain conversion module can perform frequency domain conversion on the high-order signal to generate a frequency spectrum Z(f). The peak search module can search for the peak value max|Z(f)| of the amplitude |Z(f)| of the frequency spectrum, thereby generating a peak coordinate value fM, which can reflect the occurrence of the peak value max|Z(f)| frequency f. The offset estimation module can add the peak coordinate value and a compensation value f_half to generate a sum value (fM+f_half), divide the sum value by a first divisor d1 to generate a remainder ((fM+f_half )%d1), subtract the remainder from the compensation value to produce a difference {((fM+f_half)%d1)-f_half}, divide the difference by a second divisor d2 to produce an offset estimate Value {((fM+f_half)%d1)−f_half}/d2; wherein, the offset estimation value may reflect the frequency offset between the local frequency of the transmitting circuit and the local frequency of the receiving circuit.

In one example, the received signal includes a plurality of symbols, and the plurality of symbols has a symbol frequency Fs; for example, each symbol lasts for a symbol period T, and the symbol frequency Fs can be equal to the symbol period The reciprocal of T, that is, Fs=1/T. The offset estimation module can further set the first divisor d1 according to the symbol frequency Fs; for example, the offset estimation module can make the first divisor equal to the symbol frequency. In one example, the offset estimation module can further set the compensation value f_half according to the symbol frequency Fs; for example, the offset estimation module can set the compensation value according to half of the symbol period (ie Fs/2) . In one example, the offset estimation module can further set the second divisor d2 according to the exponent P; for example, the offset estimation module can make the second divisor equal to the exponent.

In one example, the transmitting circuit modulates the remote signal according to quadrature phase shift keying (QPSK), and the power calculation module can set the index P to 4. In one example, each symbol of the received signal is selected from a plurality of constellation points c[1] to c[N], and each constellation point c[n] includes a real part re(c[n]) and An imaginary part im(c[n]); the power calculation module can also set the index P, and the setting of the index P is the sum of the power of the index of these constellation points not equal to zero.

One of the purposes of the present invention is to provide a method for estimating the frequency offset, comprising: calculating the power of an exponent P of the received signal y(t) to generate a high-order signal y ^p (t); Perform frequency domain conversion to generate a spectrum Z(f); search for the peak value max|Z(f)| of the amplitude |Z(f)| of the spectrum, and generate a peak coordinate value fM, reflecting the frequency at which the peak occurs ; Produce a difference according to the difference between the peak coordinate value and an integer multiple of the first divisor d1, so that the difference is between a negative lower limit and a positive upper limit, wherein the absolute value of the positive upper limit and the negative lower limit equal to half the first divisor; and dividing the difference by a second divisor d2 to generate an offset estimate reflecting the frequency between the local frequency of the transmit circuit and the local frequency of the receive circuit offset.

Description of drawings

In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

FIG. 1 schematically shows a receiving circuit according to an example of the present invention.

FIG. 2 schematically illustrates an exemplary flow process according to the present invention.

FIG. 3 illustrates an example of distribution of constellation points.

4A and 4B respectively illustrate the power spectral density of the received signal and the spectral amplitude of the higher-order signal in the case of no frequency offset and no multipath interference.

FIG. 5A and FIG. 5B respectively illustrate the power spectral density of the received signal and the spectral amplitude of the higher-order signal in the case of frequency offset but no multipath interference.

6A and 6B respectively illustrate the power spectral density of the received signal and the spectral amplitude of the higher-order signal in the case of frequency offset and multipath interference.

Description of component numbers in the figure:

100: Transmitting circuit

102: Fundamental frequency plastic wave circuit

104: Modulation circuit

106: channel

110: Receive circuit

112: Front-end circuit

114: Calculation circuit

116: Power calculation module

118: Frequency domain conversion module

120: peak search module

122: Offset estimation module

200: Process

202 to 208: Steps

x(t), y(t), y ^p (t), xRF(t), yRF(t), ys(t): signal

g(t-k*T): time domain function

h(t): response

w(t), n(t): noise

a _k , c[1] to c[4]: constellation points

Z(f): Spectrum

P: index

fM: Peak coordinate value

f_half: compensation value

d1, d2: divisor

T: symbol period

dF: frequency offset

dT: time difference

: Phase

df: offset estimate

fLO1, fLO2: local frequency

A: constant

detailed description

Please refer to FIG. 1 and FIG. 2 ; FIG. 1 shows a receiving circuit 110 according to an example of the present invention, and FIG. 2 shows a process 200 according to an example of the present invention. The receiving circuit 110 can use the process 200 to estimate the frequency offset. As shown in FIG. 1 , the receiving circuit 110 can cooperate with a transmitting circuit 100 to form a transceiver system. For example, the transceiver system can be a satellite or cable digital video broadcasting (DVB, digital videobroadcasting) system, the transmitting circuit 100 can be set on a satellite or a server, and the receiving circuit 110 can be set on a satellite TV box or a cable TV The set-top box; the receiving circuit 110 can also be set on a TV or a portable smart device.

The transmitting circuit 100 may include a baseband shaping circuit 102 and a modulating circuit 104 . The base frequency wave shaping circuit 102 can generate a corresponding base frequency signal x(t) for the digital input to be transmitted, which can be expressed as ∑ _k a _k ^* g(tk ^* _T ); wherein, the constellation point a _k reflects the kth The digital content of the symbol, the time-domain function g(tk*T) is the waveform of the k-th symbol, for example, the waveform obtained by processing the square root raised cosine filter (squareroot raised cosine filter), and T is A symbol period is the duration of a symbol; the reciprocal 1/T of the symbol period T is the symbol rate (symbol rate) Fs. Each symbol constellation point a _k can be selected from a plurality of preset constellation points according to the digital content of each symbol, for example, the four constellation points c[1] to c[4] shown in FIG. 3 . Fig. 3 illustrates the constellation points under quadrature key shift (QPSK) modulation, and each constellation point c[n] in the constellation points c[1] to c[4] can be expressed as a complex variable, including a real part re( c[n]) and an imaginary part im(c[n]), respectively represent the in-phase (in-phase, marked as "I" in Figure 3) component and the quadrature-phase (quadrature-phase, marked as "Q") ) component. For example, constellation points c[1] to c[4] can be A*(1+j), A*(1-j), A*(-1-j) and A*(1-j) respectively, using To represent the two-bit digital content "00", "10", "11" and "01"; wherein, j is the square root of -1, and A is a constant.

In the transmitting circuit 100 , the modulating circuit 104 can upconvert and modulate the baseband signal x(t) into a radio frequency transmitting signal xRF(t) according to a local frequency fLO1 of the transmitting end, and then transmit and propagate the signal xRF(t). For example, the signal xRF(t) can be expressed as Σ _k {re(a _k ) ^* cos(2 ^* π ^* fLO1 ^* t)+im(a _k ) ^* sin(2 ^* π ^* fLO1 ^* t)} _.

The signal xRF(t) propagates to the receiving circuit 110 and becomes the radio frequency remote signal yRF(t), which is received by the receiving circuit 110 . The process of signal xRF(t) propagating into signal yRF(t) can be modeled as a channel 106, and its effect on signal xRF(t) can be expressed as an impulse response h(t), so signal yRF(t) can be expressed as for Wherein, the item w(t) may be additive white Gaussian noise (additive white Gaussian noise).

The receiving circuit 110 may include a front-end circuit 112 and a computing circuit 114 . The front-end circuit 112 can receive the remote signal yRF(t), downconvert and demodulate the signal yRF(t) according to a local frequency fLO2 of the receiving end, and filter the signal to generate a baseband received signal y(t). In the receiving circuit 110, the front-end circuit 112 may include (not shown) a demodulation down conversion circuit, a low-pass filter (such as a filter against adjacent channel interference (ACI, adjacent channel interference)), an analog-to-digital converter , a downsampler (decimator), a symbol detection circuit, and the like. Since the local frequency fLO1 of the transmitting end and the local frequency fLO2 of the receiving end cannot perfectly match, there will be a frequency offset dF (not shown in FIG. 1 ) between the two, and the calculation circuit 114 can calculate an offset according to the signal y(t) The estimated value df is used to reflect the actual frequency offset dF. The calculation circuit 114 may include a power calculation module 116, a frequency domain conversion module 118, a peak search module 120 and an offset estimation module 122, respectively corresponding to steps 202, 204, 206 and 208 in the process 200, which can be described as follows .

Step 202: Set an exponent P by the power calculation module 116, and calculate the P power of the received signal y(t) to generate a high-order signal y ^p (t). Similar to signal x(t), signal y(t) also has real part re(y(t)) and imaginary part im(y(t)), that is, y(t)=re(y(t))+j* im(y(t)), and the high-order signal y ^p (t) calculated by the calculation circuit 114 can be expressed as {re(y(t))+j*im(y(t))} ^P . In an example, since the transmitting circuit 100 modulates the signal xRF(t) according to the four-phase key shift, the power calculation module 116 can set the index P to 4; that is:

y ^p (t) = y ⁴ (t) = re(y(t)) ⁴ +4*j*re(y(t)) ³ *im(y(t))-6*re(y(t) ) ² *im(y(t)) ² -4*j*re(y(t))*im(y(t)) ³ +im(y(t)) ⁴

. Under the four-phase bond shift, the exponent P can also be a multiple of 4.

In one example, the constellation point of each symbol is selected from N constellation points c[1] to c[N], and the setting of the index P by the power calculation module 116 is to make each constellation point c[ The sum after the Pth power of n] not equal to zero. For example, if the transmitting circuit 100 adopts 8PSK (eight-phase key shift), the index P can be set as a multiple of 8, such as 8 or 16. If the transmitting circuit 100 adopts 16PSK (Sixteen Phase Shift Keying), the index P can be set as a multiple of 16, such as 16 or 32. If the transmitting circuit 100 uses 4QAM (quadrature amplitude modulation), 16QAM, 64QAM or 256QAM, etc., the index P can also be set as a multiple of 4, such as 4 or 8.

Step 204: Perform frequency domain conversion (such as fast Fourier transform, FFT) on the high-order signal y ^p (t) by the frequency domain conversion module 118 to generate a frequency spectrum Z(f).

Step 206: The peak search module 120 searches for the global peak value max|Z(f)| of the amplitude |Z(f)| of the spectrum Z(f), thereby generating a peak coordinate value fM in the frequency domain, reflecting The frequency at which the peak value max|Z(f)| occurs; that is, the peak coordinate value fM can reflect argmax _f |Z(f)|. In another example, the peak search module 120 may also search for a peak occurrence frequency of spectrum amplitude |Z(f)| ² .

Step 208: The offset estimation module 122 calculates an offset estimation value df according to the peak coordinate value fM, the symbol frequency Fs (provided by the front-end circuit 112) and the index P (provided by the power calculation module 116, step 202), To reflect the frequency offset dF between the local frequency fLO1 of the transmitting end and the local frequency fLO2 of the receiving end. In one example, to calculate the offset estimation value df, the offset estimation module 122 can add the peak coordinate value fM and a compensation value f_half to generate a sum value (fM+f_half), wherein the offset estimation module 122 can be based on The symbol frequency Fs sets the compensation value f_half, for example, f_half=(1/2)*Fs. Next, the offset estimation module 122 calculates the remainder (fM+f_half)%d1 of dividing the sum (fM+f_half) by a first divisor d1, wherein the offset estimation module 122 can set the divisor d1 according to the symbol frequency Fs, For example d1=Fs. Next, the offset estimation module 122 subtracts the remainder (fM+f_half)%d1 from the compensation value f_half to generate a difference {(fM+f_half)%d1-f_half}, and then the difference {(fM+f_half)% d1−f_half} is divided by a second divisor d2 to generate the offset estimation value df, wherein the offset estimation module 122 can set the divisor d2 according to the exponent P, for example, d2=P. That is, in one example, df={(fM+Fs/2)%Fs−Fs/2}/P.

As will be explained later, after the P power processing in step 202 and the frequency domain conversion in step 204, the peak coordinate value fM (step 206) at which the amplitude |Z(f)| peak occurs can be expressed as (L*Fs+ P*dF); wherein, Fs is the symbol frequency; L is an integer, and dF is the frequency offset. In other words, the peak coordinate value fM will be related to an integer multiple of the symbol frequency Fs plus P times the frequency offset dF; where the frequency offset dF can be positive or negative, and the product P*dF is in a frequency domain Between the lower limit (-Fs/2) and a frequency domain upper limit (+Fs/2).

When calculating {(fM+Fs/2)%Fs-Fs/2}/P to generate the offset estimate df, since the peak coordinate value fM can be expressed as (L*Fs+P*dF), the sum value (fM+Fs/2) can be expressed as {L*Fs+(P*dF+Fs/2)}, and the remainder (fM+Fs/2)%Fs can be expressed as (P*dF+Fs/2). Since P*dF is between the lower limit (-Fs/2) and the upper limit (Fs/2) of the frequency domain, adding the compensation value f_half=Fs/2 can make the value of (P*dF+Fs/2) fall in the frequency domain The range is between 0 and Fs. Therefore, dividing the sum (fM+Fs/2) by Fs and taking the remainder can just remove the item L*Fs and keep the item (P*dF+Fs/2). Then, P*dF can be obtained by subtracting the compensation value Fs/2 from the remainder (P*dF+Fs/2), and the frequency offset dF can be obtained by dividing P*dF by P.

Continuing with FIG. 1 and FIG. 2, please refer to FIG. 4A and FIG. 4B, which illustrate the power spectral density of the received signal y(t) and the higher-order signal y ^p (t ) spectrum amplitude |Z(f)|, the horizontal axis of the two graphs is the frequency (in MHz). In the case of no frequency offset and no multipath interference, the signal y(t) can be expressed as That is, the convolution of the signal x(t) and the channel response h(t) plus the noise n(t), the noise n(t) is the noise w(t) is the noise after low-pass filtering by the front-end circuit 112 ; The power spectral density shown in FIG. 4A is related to the spectrum of each symbol in the signal y(t). In the example of Fig. 4A and Fig. 4B, the transmitting circuit 100 shifts the modulated signal according to the four-phase key, the symbol frequency Fs is 20MHz, the signal-to-noise ratio of the noise w(t) (Fig. 1) of the channel 106 is 10dB, and the front-end circuit 112 The low-pass filtering pass band against adjacent frequency band interference is 15MHz. In response to the four-phase key shift modulation, the power calculation module 116 in the calculation circuit 114 sets the exponent P to 4. Therefore, after the front-end circuit 112 receives the signal yRF(t) and down-converts and filters the signal y(t), the power calculation module 116 will calculate the high ^- order signal y4(t) (step 202), and the frequency domain conversion module 118 performs frequency conversion on the signal y ⁴ (t) to calculate the spectrum Z(f) (step 204), and the peak search module 120 searches for the peak value max|Z(f)| in the spectrum amplitude |Z(f)| to obtain Find the peak coordinate value fM (step 206).

As shown in FIG. 4B, if there is no frequency offset and no multipath interference, the peak coordinate value fM of the spectrum amplitude |Z(f)| will be located at 0 MHz, and the offset estimation value df calculated by the offset estimation module 122 ={(fM+Fs/2)% Fs-Fs/2}/P={(0+10)%20)-10}/4={10-10}/4=0 (step 208), can be correct Reflects the case of no frequency offset.

Continuing from FIG. 1 and FIG. 2, please refer to FIG. 5A and FIG. 5B, which illustrate the power spectral density of the received signal y(t) and the higher-order signal y ^p (t ) spectrum amplitude |Z(f)|, the horizontal axis of the two graphs is the frequency (in MHz). In the case of frequency offset but no multipath interference, the signal y(t) can be expressed as where dF is the actual frequency offset. Same as the example in FIG. 4A and FIG. 4B , in the example in FIG. 5A and FIG. 5B , the transmitting circuit 100 shifts the modulated signal according to the four-phase key, the symbol frequency Fs is 20 MHz, and the signal-to-noise ratio of the channel noise w(t) is 10 dB , the low-pass passband against adjacent frequency band interference in the front-end circuit 112 is 15 MHz. Furthermore, the frequency offset dF is equal to 2 MHz. In response to the four-phase key shift modulation, the power calculation module 116 can set the exponent P to 4. Therefore, after the front-end circuit 112 receives the signal yRF(t) and down-converts and filters the signal y(t), the power calculation module 116 will calculate the high ^- order signal y4(t) (step 202), and the frequency domain conversion module 118 Frequency conversion will be performed on the signal y ⁴ (t) to calculate the spectrum Z(f) (step 204), and the peak search module 120 searches for peaks in the spectrum amplitude |Z(f)| to find the peak coordinate value fM (step 206).

As shown in FIG. 5B, if the actual frequency offset dF=2MHz but there is no multipath interference, the peak coordinate value fM of the spectral amplitude |Z(f)| will be located at 8MHz, and the offset calculated by the offset estimation module 122 Estimate df={(fM+Fs/2)%Fs-Fs/2}/P={(8+10)%20)-10}/4={18-10}/4=8/4=2 (Step 208), the frequency offset dF of 2MH can be correctly reflected.

On the other hand, if the frequency offset dF is -2MHz (not shown) but there is no multipath interference, the peak coordinate value fM of the spectral amplitude |Z(f)| will be located at -8MHz (not shown), and the calculation circuit 114 Calculated offset estimate df={(fM+Fs/2)%Fs-Fs/2}/P={(-8+10)%20)-10}/4={2-10}/4 =-8/4=-2 (step 208), which can also correctly reflect the actual frequency offset dF of -2MH.

Continuing with FIG. 1 and FIG. 2, please refer to FIG. 6A and FIG. 6B, which illustrate the power spectral density of the received signal y(t) and the higher-order signal y ^p (t ) spectrum amplitude |Z(f)|, the horizontal axis of the two graphs is the frequency (in MHz). In the case of frequency offset and multipath interference, the signal y(t) can be expressed as ys(t)+exp(j*φ)*ys(t-dT)+n(t), where ys(t ) can be expressed as That is, a signal with a single path but with a frequency offset, exp(j*φ)*ys(t-dT) represents the signal of another path, φ represents the additional phase on this path, and dT represents the time difference between different paths. Same as the example in FIG. 4A and FIG. 4B, in the example in FIG. 6A and 6B, the transmitting circuit 100 shifts the modulated signal according to the four-phase key, the symbol frequency Fs is 20 MHz, and the signal-to-noise ratio of the channel noise w(t) is 10 dB. The low-pass passband against adjacent frequency band interference in the front-end circuit 112 is 15 MHz. Furthermore, the actual frequency offset dF is equal to 2 MHz, the phase φ is equal to 1.2, and the time difference dT is 0.01 μs.

In response to the four-phase key shift modulation, the power calculation module 116 can set the exponent P to 4. Therefore, after the front-end circuit 112 receives the signal yRF(t) and down-converts and filters the signal y(t), the power calculation module 116 will calculate the high ^- order signal y4(t) (step 202), and the frequency domain conversion module 118 Frequency conversion will be performed on the signal y ⁴ (t) to calculate the spectrum Z(f) (step 204), and the peak search module 120 searches for peaks in the spectrum amplitude |Z(f)| to find the peak coordinate value fM (step 206).

As shown in FIG. 6B, if the frequency offset dF=2MHz and there is multipath interference, the peak coordinate value fM of the spectrum amplitude |Z(f)| will be located at 28MHz, and the offset estimation value df calculated by the offset estimation module 122 ={(fM+Fs/2)% Fs-Fs/2}/P={(28+10)%20)-10}/4={18-10}/4=8/4=2 (step 208 ), can correctly reflect the frequency offset dF of 2MH, even if there is multi-path interference. On the other hand, if the frequency offset dF=-2MHz and there is multipath interference, the peak coordinate value fM of the spectrum amplitude |Z(f)| will be located at 12MHz (not shown), and the offset calculated by the offset estimation module 122 Shift estimated value df={(fM+Fs/2)%Fs-Fs/2}/P={(12+10)%20)-10}/4={2-10}/4=-8/4 =-2 (step 208), the frequency offset dF of -2 MHz can still be correctly reflected under multipath interference.

Comparing the two examples in Fig. 4B and Fig. 5B, it can be seen that in the frequency spectrum Z(f) of the high-order signal y ^p (t) with P = 4th power, the frequency offset dF of 2 MHz will cause the peak coordinate value fM to change from Fig. 0MHz in 4B is changed to 8MHz in FIG. 5B, that is, fM=P*dF. Comparing the two examples in Figure 5B and Figure 6B, it can be seen that although the actual frequency offset dF of these two examples is equal to 2MHz, the multipath interference in Figure 6B will cause the peak coordinate value fM to change from 8MHz in Figure 5B to Figure 6 28MHz in. In fact, multipath interference will make the peak coordinate value fM additionally shifted by an integer multiple of the symbol frequency Fs, ie fM=P*dF+L*Fs, where L is an integer.

In a prior art, the offset estimation value df is calculated by dividing the peak coordinate value fM by the exponent P, that is, df=fM/P. For example, in the example of FIG. 5B , the peak coordinate value fM=8 MHz, directly divided by P=4, it can be obtained that the frequency offset is 2 MHz. However, this prior technique cannot be correctly applied to multipath interference; for example, in the example of Fig. 6B, the peak coordinate value fM = 28 MHz, directly divided by P = 4 gives the offset estimate df equal to 28/4 =7MHz, it cannot correctly reflect the real frequency offset dF=2MHz.

In contrast, the offset estimation module 122 in the calculation circuit 114 in the example of the present invention will first remove an integer multiple of the symbol frequency Fs (for example, {(fM-Fs/2)% Fs from the peak coordinate value fM +Fs/2}), and then divide the result by the exponent P to correctly reflect the true frequency offset dF (step 208).

In another example, the offset estimation module 122 may also calculate {fM-Fs*round(fM/Fs)}/d to generate an estimated offset value df, wherein the function round(r) is an integer closest to the variable r; In other words, round(fM/Fs) is to calculate the integer multiple L to subtract the integer multiple Fs*round(fM/Fs) of the symbol frequency from the peak coordinate value fM. For example, in the example of FIG. 6B, the peak coordinate value fM=28, the offset estimated value df can be calculated as {28-20*round(28/20)}/4={28-20*round(1.4 )}/4={28-20}/4=8/4=2. On the other hand, if the actual frequency offset dF=-2MHz and there is multipath interference, the peak coordinate value fM will be located at 12MHz (not shown), and the estimated offset value df can be calculated as {12-20*round(12/ 20)}/4={12-20*round(0.6)}/4={12-20}/4=-8/4=-2.

In another example, the offset estimation module 122 may also use a periodic function and a related inverse function to remove an integer multiple of the symbol period Fs from the peak coordinate value, for example, to generate the offset estimation value df.

In another example, the offset estimation module 122 can also iteratively calculate the offset estimation value df. The offset estimation module 122 can check whether the peak coordinate value fM is in the frequency domain range (-Fs/2, Fs/2), if so, can directly calculate fM/P to generate the offset estimated value df; if the peak coordinate value fM is greater than the range (-Fs/2, Fs/2), you can subtract twice the symbol frequency Fs from the peak coordinate value fM to get the first difference (fM-Fs), and check the first difference Whether the value (fM-Fs) is in the range (-Fs/2, Fs/2); if so, the first difference (fM-Fs) can be divided by the exponent P to generate the offset estimate df, if the first difference The value (fM-Fs) is still greater than the range (-Fs/2, Fs/2), and the second difference ( fM-2*Fs), and check whether the second difference (fM-2*Fs) is within the range (-Fs/2, Fs/2); if so, the second difference (fM-2*Fs) Divided by the exponent P to generate the offset estimate df, if the second difference (fM-2*Fs) is still larger than the range (-Fs/2, Fs/2), it can be obtained from the second difference (fM-2*Fs ) to subtract twice the symbol frequency Fs to obtain the third difference (fM-3*Fs), and so on.

The estimated offset value df generated by the calculation circuit 114 can be used as a basis for compensating the frequency offset. For example, a mixer (mixer, not shown) may be provided in the front-end circuit 112 of FIG. 1 , and this mixer may perform compensatory mixing according to the offset estimation value df to offset the frequency offset dF . And/or, the receiving circuit 110 can also compensate the frequency deviation in the subsequent processing of the baseband received signal y(t) (for example, symbol retrieval).

The functional modules 116 to 122 of the computing circuit 114 can be implemented by dedicated hardware circuits, or can be implemented by general-purpose arithmetic logic circuits executing software or firmware. For example, the frequency domain conversion module 118 may be implemented by hardware, and the offset estimation module 122 may be implemented by a general-purpose arithmetic logic circuit executing firmware. Please note that after reading the above description, those skilled in the art should have the ability to implement the technology of the present invention with various feasible technologies in the field (including software, firmware, hardware or a combination thereof), and details will not be repeated here.

In summary, the present invention calculates the high-order signal based on the received signal, and then searches for the peak coordinate value of the spectrum amplitude of the high-order signal, and can overcome the integer times symbol period additionally introduced in the peak coordinate value by multipath interference, The frequency offset between the transmitting end and the receiving end is correctly estimated according to the difference between the peak coordinate value and the integral multiple symbol period.

Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make some modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the present invention The scope of protection should be defined by the claims.

Claims (14)

1. a kind of receiving circuit for estimating frequency shift (FS), comprising：

One front-end circuit, is used to receive the remote signaling that a radiating circuit is transmitted, and produces one to receive according to this Signal；And

One counting circuit, couples the front-end circuit, comprising：

First power computing module, calculates the power of an index of the reception signal to produce a high order signal,

One frequency domain modular converter, frequency domain conversion is carried out to the high order signal to produce a frequency spectrum；

One peak value search module, searches a peak value of the amplitude of the frequency spectrum, and the hair of the peak value is reflected to produce One peak value coordinate values of raw frequency；And

One bias estimation module, the peak value coordinate values is added with an offset to produce one and value, by this With value divided by one first divisor to produce a remainder, the remainder and the offset are subtracted each other to produce a difference, And by the difference divided by one second divisor to produce a bias estimation value, wherein, the bias estimation value reflection should Frequency shift (FS) between the local frequency of radiating circuit and the local frequency of the receiving circuit.

2. receiving circuit as claimed in claim 1, it is characterised in that accorded with comprising multiple in the reception signal Unit, the plurality of symbol has a symbol frequency, and the bias estimation module more should according to the symbol frequency setting First divisor.

3. receiving circuit as claimed in claim 1, it is characterised in that accorded with comprising multiple in the reception signal Unit, the plurality of symbol has a symbol frequency, and the bias estimation module more should according to the symbol frequency setting Offset.

4. receiving circuit as claimed in claim 1, it is characterised in that accorded with comprising multiple in the reception signal Unit, the plurality of symbol has a symbol frequency, and the bias estimation module is more according to the half of the symbol frequency Set the offset.

5. receiving circuit as claimed in claim 1, it is characterised in that the bias estimation module was more according to should Index sets second divisor.

6. receiving circuit as claimed in claim 1, it is characterised in that the bias estimation module more make this Two divisors are equal to the index.

7. receiving circuit as claimed in claim 1, it is characterised in that the radiating circuit is according to four phase keys Move (QPSK, quadrature phase shift keying) and modulate the remote signaling, and the power calculates mould The index is more set as 4 by block.

8. receiving circuit as claimed in claim 1, it is characterised in that the remote signaling includes multiple symbols, Each symbol is by being selected in multiple constellation points first, each constellation point includes a real part and an imaginary part；And the power Computing module can more set the index, and the index be set so that the index power of the plurality of constellation point after Totalling be not equal to zero.

9. a kind of method for estimating frequency shift (FS), is applied to a receiving circuit, comprising：

The remote signaling that one radiating circuit is transmitted is received with the receiving circuit, and produces one to receive letter according to this Number；

The power of an index of the reception signal is calculated to produce a high order signal；

Frequency domain conversion is carried out to the high order signal to produce a frequency spectrum；

A peak value of the amplitude of the frequency spectrum is searched, to produce a peak value coordinate of the occurrence frequency for reflecting the peak value Value；

According to the peak value coordinate values and one first divisor integral multiple between difference produce a difference, the difference be situated between Between a negative lower limit and a positive upper limit, wherein the absolute value of the positive upper limit and the negative lower limit is equal to this and first removes Several half；And

By the difference divided by one second divisor to produce a bias estimation value, wherein, bias estimation value reflection Frequency shift (FS) between the local frequency of the radiating circuit and the local frequency of the receiving circuit.

10. method as claimed in claim 9, it is characterised in that comprising multiple symbols in the reception signal, The plurality of symbol has a symbol frequency, and the method is further included：According to the symbol frequency setting, this first is removed Number.

11. methods as claimed in claim 9, further include：Second divisor is set according to the index.

12. methods as claimed in claim 9, further include：Second divisor is set to be equal to the index.

13. methods as claimed in claim 9, it is characterised in that the radiating circuit is moved according to four phase keys The remote signaling is modulated, and the method is further included：The index is set as 4.

14. methods as claimed in claim 9, it is characterised in that the remote signaling includes multiple symbols, Each symbol is by being selected in multiple constellation points first, each constellation point includes a real part and an imaginary part；And the method Further include：The judgement for being not equal to zero according to the totalling after the index power of the plurality of constellation point is selected this and is referred to Number.