CN1222267A - DSP implementation of cellular base station receiver - Google Patents

Abstract

Demodulating FM signals using digital signal processing extracts a carrier signal from digitized channel signals, multiplies the digital channel signal with this extracted carrier signal, and further filters out the carrier signal to produce the demodulated signal. The DSP technique first down converts a group of channels to baseband which are then processed through an A/D converter to produce a digitized composite signal. A bank of bandpass filters, typically based on FFT processors, applied to the composite signal produce (a group of) digitized channel signal(s). The digitized channel signal is then demodulated by recovering a carrier signal by digitally filtering, for example, using a Hilbert bandpass filter, the channel signal and digitally filtering the product of the carrier signal and the channel signal to recover the modulating voice signals.

Description

The DSP of cellular network base stations receiver implements

Invention field

The present invention relates to demodulation, particularly, relate to the demodulation FM signal of utilization Digital Signal Processing (DSP) the signal that in the Cellular Networks radio telephone system, sends.

The description of related art

For the conventional method of FM demodulation generally based on analogue system.Though simulation FM demodulation techniques are not expensive, their general very flexible dissimilar FM demodulation that has been adapted to for dissimilar signal (such as the Cellular Networks environment).As in other field, the utilization of Digital Signal Processing (DSP) has produced the greater flexibility of utilization FM signal (such as in the Cellular Networks phone).

In the Cellular Networks telephone system, say for given base station to a certain extent and relate to arbitrarily, because can distribute to arbitrary base station to any channel combination channel allocation.When utilization simulation FM demodulation techniques handle to receive or the demodulation of people office signal on, this has caused some problems, because typical FM demodulator (generally being the phase-locked loop (PLL) under analog form) is only in response to some frequency range.When existence needed a plurality of channel of demodulation, this problem can take place.Because giving arbitrary given base station those channel allocation can be arbitrarily, can processing channel or any feasible combination of frequency so must become base station design.In addition, held the variation of channel demands according to the multiple research of Cellular Networks, the channel of distributing to given base station can change along with the time.Utilization standard FM technology, people can be used in tunable one group of standard analog PLL in the whole FM frequency spectrum.This is very wide for this scope, and the enforcement of this analog PLL is become very expensive and complicated.In addition, the analog circuit that is used for the FM demodulation comprises phase place or frequency-locked loop and multiple discriminator circuit.Generally, when the essential several channel of demodulation (such as in cellular network communication), simulation FM demodulation is not flexible and changeable, supposes that the FM signal in the Cellular Networks system can be " standard cellular net " (AMPS) or arrowband AMPS (NAMPS).

Summary of the invention

The objective of the invention is to, extract carrier signal and preferably utilize digital AM demodulation and multiplication and integration to extract voice signal, provide digital FM to separate the method and apparatus that withers by preferably utilizing the digital Hibert conversion that recovers for subcarrier.

People can use DSP to implement, rather than the simulation enforcement of withering is separated in utilization to FM.Only need device is carried out minimum variation because be suitable for unlike signal and different frequency, such as, some constant and the subprogram in dsp software changed, so the DSP of the FM demodulation of utilization AM demodulation techniques implements to provide flexibility.The DSP of FM demodulation implements to have strengthened with simulating the flexibility that the FM demodulation techniques can't obtain.For example, digital FM separates to wither to provide and easily changes parameter adapting to the freedom of various frequencies, and can easily adapt to subdistrict position to hold TDMA, NAMPS, GSM and AMPS technology.

According to one embodiment of present invention, the method for utilization Digital Signal Processing (DSP) demodulation FM signal comprises the step of selecting the frequency band of channel by the utilization band pass filter in analog domain.For example, then, the channel group from Cellular Networks or other wireless network that receives in the base station is displaced downwardly to IF.From IF, select one group of adjacent channel from the frequency band of channel, selecting, and be displaced downwardly to base band, sample to move into numeric field at the twice place of the frequency band upper limit.Utilization DSP implements, at the base band place, and every group of channel of separate processes.For example, sampling is at the channel of base band group, and with the digital channel signal that generation has homophase and quadrature component, wherein in-phase component has carrier component, and quadrature component has modulation signal (voice) component.Each digitized sampling channel of separate processes.The logical algorithm of utilization narrow-band, such as, Hilbert conversion band leads to algorithm, and the filtering figure channel signal is to extract carrier signal.Then, the mixed signal that formation comprises modulation signal is multiply by in carrier signal and digital channel signal digitlization mutually.At last, leach carrier wave and high-frequency signal only to produce modulation signal.In preferred embodiment of the present invention, Hilbert bandpass filtering algorithm has the passband less than 600Hz, with the carrier component at the central frequency place that is extracted in filter.

In conjunction with the accompanying drawing that the multifrequency nature of embodiments of the invention is shown with the method that exemplifies, by following detailed, other characteristic of the present invention and advantage will be apparent.

Description of drawings

With reference to accompanying drawing, will describe embodiments of the invention in detail, wherein same numeral is done corresponding expression.

Fig. 1 is the block diagram that phase place and warbled equivalence (equivalence) are shown;

Fig. 2 is the block diagram that is used for the Armstrong modulator of arrowband angle modulated (phase modulated);

Fig. 3 A-3C illustrates the conversion of channel group to number format;

Fig. 4 is the band pass filter group that is used for channelizing;

Fig. 5 is each channel frequency domain representation;

Fig. 6 is the block diagram that the principle of DSP base arrowband angle modulator (DSP based narrow band angle modulator) is shown;

Fig. 7 is that the frequency domain of channel signal is represented; With

Fig. 8 is the flow chart that the utilization Digital Signal Processing is separated the FM signal that withers.

The detailed description of preferred embodiment

For convenience of explanation, shown in the drawings according to embodiments of the invention, the method for the demodulation FM signal of utilization DSP.Particularly, preferred embodiment of the present invention concentrates on the DSP enforcement of AMPL/NAMPS/TDMA.In addition, as illustrative example, following description concentrates on uses arrowband FM to overlapping the demodulation of the single voice channel on the arrowband.

In the frequency band of Cellular Networks telephone system between 824MHz and 894MHz of the U.S., operate.Send (from the mobile radio station to the base station) for " oppositely ", be distributed in the lower part between 824MHz and the 849Mhz.Send (from the base station to the mobile radio station) for " forward direction ", be distributed in the high part between 869MHz and the 894MHz.Generally, normal channel spacing is 30KHz,, frequency band is subdivided into " piece " that is, and wherein central frequency is 30KHz at interval, and normal bandwidth is 30KHz (either side at central authorities or narrow band frequency is 15KHz).Each channel is used carrier wave by utilization angle modulated, particularly frequency modulation (FM), and wherein maximum frequency deviation is 12KHz.Because the voice channel bandwidth normally is 3KHz, so by using the Carlson rule, voice channel occupies the bandwidth of 2* (3+12)=30Khz.Said frequencies distributes permission to go up to 832 pieces, generally is counted as two " bands ", that is, A and B, each band has the capacity of 416 pieces.

Consider that with frequency division multiplexing (FDM) method segmentation frequency spectrum, wherein each piece is called channel.Each channel is represented a telephone conversation.Each mobile unit is assigned and is used for the particular carrier wave frequency, generally is to be lower than the carrier frequency 45MHz. that uses in the forward direction that is undertaken by base station reverse sends.This scheme is commonly referred to as " AMPS " (AMPSAdvanced Mobile Phone Service).

The basic principle of Cellular Networks phone is that arbitrary geographic area can be subdivided into " sub-district ", and wherein adjacent cell uses nonoverlapping frequency chunks.As a result, each sub-district and base station can only be with the subclass of these 416 channels.This has limited the quantity of while user (simultaneous subscriber) in by the zone of base station services.In order to increase the number of users of service, other operator scheme of executed.General Principle is that each channel of utilization provides service to the user more than.The prerequisite of NAMPS is that each 30KHz piece is subdivided into 3 sub-pieces (bandwidth of each sub-piece is 10KHz), and very narrow frequency band FM is used for each voice channel.This method makes and can be increased to 3 times by the number of users that each base station is served simultaneously.

The second method that makes the while number of users increase by 3 times is again utilized TDMA (time division multiple access) technology.TDMA supposes the digitized speech channel, and it is encoded, thereby about 8kbits/sec is used for voice channel.The 30KHz bandwidth of using in the Cellular Networks phone can be supported modulated digital bit stream.When frame structure is added in this digital bit stream, can be by with different " time slot " distributes to different users and hold a plurality of users in piece.Nominally each slot transmission of all channel is used for a user's signal.By keeping the FDM layering of 30KHz piece, TDMA can be present in the identical sub-district simultaneously with analog cellular net (such as AMPS).So, arbitrary 30KHz piece is distributed to numeral (TDMA) or simulation transmission.

As shown in Figure 1, the integrator 26 with using phase-modulator 27 and before phase modulated signal 24 being quadratured can obtain frequency modulation(FM) 28.Integration is equal to the filter that the response of 6dB/ frequency multiplication is provided, and it is positioned at the opposite of preemphasis filter.So,, can eliminate preemphasis circuit by utilization phase-modulator 27.

(AM) is different for amplitude modulation(PAM) with being considered to " linearity ", and angle modulated itself is non-linear.This means, in AM, keep the bandwidth of modulation signal.By frequency modulating signal is come conversion with the amount that equals carrier frequency, obtain composite signal.In angle modulated, bandwidth is several times of modulation signal.Yet for the low system index that withers, the method that can be similar to the amplitude modulation(PAM) that is used for DSP is treated phase modulated.

Above-mentioned relation between PM and AM is based on Armstrong modulator (as shown in Figure 2).Notice that (double-side band compression carrier wave, combination DSB-SC) is added to carrier component in the quadrature component of above-mentioned AM signal combination gained (compound) signal can be regarded as the AM signal.In traditional DSB-AM signal (non-compression carrier wave), additional carrier component is an in-phase component, wherein generates the DSB-SC signal with latent carrier wave (implied carrier).So, if DSB-SC signal and the addition of homophase carrier component, the DSB-AM signal of generation rule so.If additional carrier component is a quadrature component, the signal of gained is equal to arrowband PM so.Top description is represented with following mathematic(al) representation:

If | m (t) |＜＜1 (modulation signal is little), composite signal PM signal so, vpm (t) can be expressed as follows:

vpm(t)=Acos(2πf _ct+m(t))

=Acos(2πf _ct)-Am(t)sin(2πf _ct)

For the legacy demodulator that the utilization simulation is implemented, can directly locate to carry out this processing in carrier frequency or in intermediate frequency (IF).Above-mentioned processing can comprise restriction (hard or soft) so that the arbitrary AM of adding component minimum.Utilization PLL (phase-locked loop) method or frequency discriminator method are carried out demodulation.When at the base band place, during digital form, it is suitable not having a kind of processing (amplitude limit, PLL, frequency discriminator) in these processors to channel signal.For example, amplitude limit generates harmonic wave, in simulation is implemented, and can the above-mentioned harmonic wave of elimination.In DSP implemented, amplitude limit had fuzzy,, introduced the effect of band endoparasitism component (inband spurious component) by aliasing that is.So in all cases, the imitation simulation process is unsuitable.

Because the DSP technology can be used for the situation of AM, so, can design digital demodulator by adopting the phase antilogical of above-mentioned Armstrong modulator.This is the important principles of preferred embodiment of the present invention.All handle the enforcement that all is suitable in dsp processor, and can implement in hardware, firmware or software.

For example, referring now to Fig. 3 A-3C, the processing that channel group is converted to digital signal is discussed.Analog front circuit selects one group of preferable n FM channel (shown in Fig. 4 A) to add 2 boundary belts as 14 30KHz channels, and converting base band under this channel group.Be used in preliminary election band pass filter 42 in the whole frequency band (it [F place reduces interference), the FM modulation signal of a bandpass filtering n channel.Then, by multiply by a fosc41, channel group is converted to IF from the output of band pass filter 42.Then, from the output of multiplier 43, by narrow-band pass filter (sharp bandpass filter) 44, it finishes the selective channel group.Band pass filter 44 has the frequency response of Fig. 3 B that applies in whole frequency bands of the channel 30 of Fig. 3 A.The selection of channel 1 to 14 (wherein, be preferably in the protection channel 0 and 15 sharply roll-offing of 6dB/ frequency multiplication arranged) is reduced to disturb.Second multiplier 45 of Fig. 3 C is multiplying each other from narrow-band pass filter 44 and the 2nd fosc49, so that channel group is converted to base band.Before 47 conversions of analog to digital (A/D) converter, carry out filtering by 46 pairs of outputs of anti-aliasing low pass filter from second multiplier 26.Output from low pass filter 46 has 16 channels altogether, and from frequency band 0 to frequency band 15, it is wide that each channel is approximately 30KHz.As the Nyquist principle was desired, A/D converter 47 had frequency range (for example, (for example, sampling rate 960KHz) of twice 480KHz) that is at least in the selected channel group at base band place.For example, as n=16 (for DSP, selected channel group number), are 960KHz for the nominal sampling rate of A/D converter 47.Unshowned in Fig. 3 C is automatic gain control (AGC) function of carrying out before the A/D conversion, and the resolution maximum that is provided that is provided by the operation of the A/D in gamut is provided for it.

In case the selected channel group moves to base band, can separate and each channel of handling concurrently wherein.Fig. 4 illustrates the band pass filter group that is used for generally based on the channelizing of 32 DFT (discrete Fourier transform).For convenience of explanation, the band pass filter central frequency is 30nKHz, wherein n=0 to 15.In addition, on every side of each band pass filter, the filter transition band is 15KHz.Band pass filter separates channel, wherein, at down-sampling (undersampling) afterwards, can the individual channel that channel is regarded as under the 120KHz sampling rate be flowed.Recognize that channel sample speed (120KHz) is the twice of Nyquist sampling rate for example.In the preferred embodiment of band pass filter group, by latent down-sampling (implied undersampling), with the known technology of the personnel that are familiar with this technical field (as by Kishan Shenoi (Prentice Hall, nineteen ninety-five) like that (as the reference data and be incorporated herein) described in the 7th chapter of the Digital Signal Processing in telecommunications), reorientation channel spectrum to be processed (as shown in Figure 5).The effect of frequency translation and down-sampling is the carrier frequency of channel is fixed on 30KHz (nominal), and it is 1/4th of a sample frequency.The carrier frequency at (the perhaps IF frequency place that is being used to handle) should be the integer approximate number of sampling rate (for example, 1/2,1/3,1/4, or the like) at the base band place, thereby makes processing simpler.By the utilization sampling rate (approximately be 2 times) higher than minimum sampling rate, reduce the influence of aliasing, and the complexity of the filter that reduces in bank of filters, to use.

For each channel, the 30KHz frequency spectrum that comprises signal is to clean relatively, because the 15KHz boundary belt is arranged on the both sides of signal band, worsens the signal band to stop interfering components.In preferred embodiment of the present invention, can implement the band pass filter group by utilization DFT (for example, FFT structure).The output signal that DFT handles provides the homophase and the quadrature component (that is sophisticated signal) of each channel.Under the situation of TDMA, sophisticated signal is useful, because simplify digital demodulation in the time can obtaining homophase and quadrature component simultaneously.As shown in Figure 6, the digital channel signal in time domain 122 is X _PMAnd sampling rate f=120KHz (nT).Digital channel signal 122 is the composite signals that obtain from the band pass filter group after the A/D of Fig. 3 C converter 47.X _PM(nT) frequency domain representation is X _PM(f), and with mathematic(al) representation be expressed as follows respectively:

X _PM(nT)=Acos(2πf _onT)-Am(nT)sin(2πf _ont)

X _PM(f)=A/2[δ(f-f _o)+δ(f+f _o)]-A/(2j)[M(f-f _o)-M(f+f _o)]

Fig. 7 illustrates X _PM(f) frequency domain representation.By+30KHz and- pulse 142 and 143 at 30KHz place represents carrier components 141 (that is delta function). Signal component 146 and 147 is quadrature components (shown in " j " of top expression formula) of composite signal.Signal component 146 and 147 is with the AM signal form, that is, and and double-side band compression carrier wave (DSB-SC) AM.Because voice signal do not have DC component, so will provide carrier component 143 in that the narrow-band of about 30KHz of digital channel signal is logical.

General Hilbert conversion of using in DSP can be used for filtering to obtain the 30KHz carrier wave.Hilbert transformed filter 124 is the digital filters with following frequency response:

For | f|＜fs/2, G (f)=-jH (f) sgn (f)

Wherein, sgn (f)=1 f＞0

0 f=0

-1 f＜0

And H (f) is to be the response in a narrow margin at center with 30KHz.For example, have constraint factor to show that in the center in FIR (finite impulse response (FIR)) filter of negative symmetry (odd function), it is pure imaginary number that frequency response guarantees, it is corresponding with 90 degree phase shifts.Particularly, strange length, odd symmetry filter (the N length of filter is odd-integral number) will have the phase shift of 90 degree, and (N-1)/2 smooth delay (flat delay) of sampling.Because N is an odd number, so smooth delay will be the integer sampling.In addition, if select N, thereby (N-1)/2 be 4 multiple, the benefit that can obtain to double so.Because the carrier wave mark deserves to be called 30KHz, carrier cycle is that sampling rate is 4 samplings of 120KHz so.So, if selected smooth delay is 4 multiple, so will be very accurate at the phase modulation of the output of Hilbert transformed filter 124.

The bandwidth of Hilber transformed filter 124 (perhaps implementing the processor of Hilbert transformed filter) is made as on every side of carrier signal greater than approximately 300Hz is wide.This is because make do not have audio signal in the carrier wave of 300Hz by the audio signal filtering in transmitter.So whole bandwidth of Hilbert transformed filter 124 are preferably less than about 600Hz.Output from Hilber transformed filter 124 is the digital signal that the exalted carrier signal is used to receive.Then, further carrier signal is separated from voice signal multiplying each other with the carrier signal of digital signal quadrature and digital signal.

Processing section 126 is implemented corresponding with 128 with the numeral of the synchronous demodulator that is used for AM among Fig. 6.Generally, the difference (as opposite with public envelope monitoring) of separating in withering at synchronous AM is generation and is used to produce AM composite signal X _PMThe local carrier y (nt) of carrier phase (nt) and Frequency Synchronization.Yet according to preferred embodiment of the present invention, by design, " local oscillator " that guarantee to be used to generate carrier wave y (nt) has correct frequency and correct phase place in fact.As mentioned above, detect the in-phase component 143 (Fig. 7) of digital channel signal 122, obtain carrier signal by using digital Hilber conversion band pass filter.The output 129m (nt) of demodulator 126 will be proportional (on the amplitude) digital representation at the modulation signal m at transmitter place (t), and wherein, the amplitude of the modulation index of using in modulator (not shown) is in fact less than 1.126 digital channel signals that recover of software DSP base arrowband angle demodulator:

X _PM(nT)=Acos(2πf _onT)-Am(nT)sin(2πf _ont)

Multiply by carrier wave:

y(nT)=Acos(2πf _onT+π/2)=Asin(2πf _onT)

Draw demodulator output;

w(nT)=A/2cos(4πf _onT)-A ²/2m(nT)[1-cos(4πf _onT)]

Only just can obtain to recover proportional through the demodulation modulation signal by filter 128 low-pass filtering:

x ₁(nT)=A ²/2m(nT)

Therefore, the understanding for y (nT) 125 latent carrier waves is very useful.Though, not shown in the accompanying drawings, also this signal can be used to control AGC (automatic gain control) function, to use A/D converter 47 sampled signals and/or every channel AGC function.Particularly, can amplify (perhaps decay) channel signal, thereby signal y (nT) 125 has known regulation amplitude.

The embodiment of above-mentioned DSP is a channel of selecting from n channel at IF level place.Down-sampling in the digital band-pass filter group has guaranteed that all n channel is all identical, wherein carrier component under 30KHz, in the book of Kishan Shenoi the 7th chapter described.Therefore, can implement identical step, with all n of demodulation channel.This allows a DSP or a plurality of identical software modules of DSP group utilization, carries out base station functions.

Utilization flow process Figure 200 as shown in Figure 8 can conclude above-mentioned DSP FM demodulation techniques.At first, in whole frequency band, the FM signal that bandpass filtering 201 receives.From selecting 202 channel group, and move down into IF204 from the FM signal, preferably move on to base band 206 then.Low-pass filtering 208 now is positioned at the selected channel group of base band, is higher than any high-frequency signal of the frequency band upper limit with removal.Then, under its about doubled frequency of the frequency band upper limit 210, sampling selected channel group.A channel in digitized filtered (channelizing) the 212 selected channel groups is to carry out the digital channel signal of digitized processing to it, to extract modulation signal.In this, can be by Attached Processor 214 these processing of parallel execution of each channel.By to channel signal 216 combine digital arrowband bandpass filterings (such as, utilization Hilber conversion filtering algorithm), extract the carrier signal that is used for digital channel signal.Then, exalted carrier signal and digital channel signal 218 digitlizations are multiplied each other, to recover modulation signal.The digital low-pass filtering algorithm leaches the displacement carrier signal, thereby only recovers modulation signal 220.If a plurality of DSP are not used in parallel processing (step 214), for the additional channel in group 222, single DSP can repeat this processing so.The bandpass filtering algorithm that is used for channelizing walks abreast all n is provided a channel, thereby can extract modulation signal to walk abreast by other channel of digitized processing.

Though top description relates to specific embodiment of the present invention, it should be understood that and to carry out numerous variations and do not depart from its design.For example, can be at IF, rather than under base band frequency, carry out described demodulation method.Appending claims attempts to cover all changes that drop in scope of the present invention and the design.

Therefore, consider that in all respects, the embodiment that is disclosed is illustrative rather than definitive thereof here, by appended claims, rather than foregoing description, limit scope of the present invention, thereby comprise intention and the interior all changes of scope that drop on claims here.

Claims (15)

1. one kind by with one of IF that the frequency modulating signal modulated carrier signal produced and base band frequency down, the method of digital demodulation FM signal, described FM signal has homophase and quadrature component, wherein said in-phase component has described carrier signal, and described quadrature component has described modulation signal, it is characterized in that described method comprises the following steps:

At suitable sampling rate down-sampling FM signal, to produce digital channel signal;

The described digital channel signal of digitized filtered extracts described carrier signal with the logical algorithm of utilization narrow-band;

The mixed signal that formation comprises described modulation signal and moves on to the described carrier signal of different frequency is multiply by in described carrier signal and described digital channel signal digitlization mutually; With

The described mixed signal of digitized filtered is to recover described modulation signal.

2. the method for claim 1 is characterized in that, the logical algorithm of described narrow-band is the logical algorithm of Hilbert band.

3. the method for claim 1 is characterized in that, the logical algorithm of described narrow-band has the logical scope of band of about 300Hz in every side of described carrier signal.

4. method as claimed in claim 2 is characterized in that, it is the passband at center that described Hilbert band pass filter has with about 30KHz, selects described sampling rate, thereby described base band FM carrier wave is at the 30KHz place.

5. method as claimed in claim 2 is characterized in that, described Hilbert band pass filter is strange length and odd symmetry, the FIR filter.

6. method as claimed in claim 2 is characterized in that, the logical algorithm of described Hilbert band comprises the phase shift of about 90 degree and smooth delay (N-1)/2, and wherein N is an odd-integral number.

7. method as claimed in claim 6 is characterized in that, the FM carrier frequency of locating in one of IF and base band is an integer with the ratio of described sampling rate.

8. method as claimed in claim 6 is characterized in that, (N-1)/2 is multiples of 4.

9. the method for claim 1, it is held to levy and is, by leaching high-frequency signal, carries out filtering to described mixed signal to produce described modulation signal.

10. the method for claim 1 is characterized in that, the described sampling rate of the described FM signal that is used to sample is about twice of the described upper frequency limit of described FM signal.

11. the method for claim 1 is characterized in that, also comprises for other digital channel signal in described FM signal executed in parallel step as claimed in claim 1.

12. a digital demodulation has the method for the composite signal of a plurality of channels, by producing each channel with the frequency modulating signal modulated carrier signal, the FM signal has homophase and quadrature component, wherein, described in-phase component has carrier component, and quadrature component has described modulation signal, it is characterized in that, described method comprises the following steps:

A plurality of channels of described FM signal are moved to base band;

Under the speed of twice of bandwidth that is described FM band signal at least, the described FM band signal of sampling is to produce digital channel signal;

Digital band pass filtering is to produce a plurality of digital channel signals;

Then for each digital channel;

The described digital channel signal of digitized filtered extracts described carrier signal with utilization Hilbert bandpass filtering algorithm, and it is the passband at center that wherein said Hilbert bandpass filtering algorithm has with about 30KHz;

Described carrier component be multiply by the formation mixed signal mutually with described digital channel signal digitlization, and it comprises described modulation signal and moves on to the described carrier signal of different frequency; With

The described mixed signal of digitized filtered is to recover described modulation signal.

13. the FM signal demodulating equipment of a utilization Digital Signal Processing (DSP) is characterized in that described device comprises:

Be used for the described FM signal of digitized sampling and comprise the device of the digital channel signal of homophase and quadrature component with generation, wherein said in-phase component has carrier component and described quadrature component has signal component;

Be used to use the logical described digital channel signal of algorithm digitized filtered of narrow-band to extract first device of described carrier signal;

Be used for the device that formation comprises the modulation signal and the mixed signal of the described carrier signal that moves on to different frequency is multiply by in described carrier signal and described digital channel signal digitlization mutually; With

Be used for the described mixed signal of digitized filtered to recover second device of described modulation signal.

14. device as claimed in claim 13 is characterized in that, the described first digitized filtered device is carried out the logical algorithm of Hilbert band.

15. method as claimed in claim 7 is characterized in that, described delay is described FM carrier wave and the integral multiple of the ratio of described sampling rate.

Images