CN101420411B - Fast carrier capture method with low signal-noise ratio - Google Patents

Abstract

A method for quickly catching the low signal-to-noise ratio carrier comprises the following steps: (1) multiplying an input signal with a cos branching data generated by a first NCO module; (2) executing half band filtering; (3) executing down sampling to a first accumulated down sampling module, temporarily storing into an exterior SDRAM through an FIFO-IN module, and the data speed fsl and the maximum capture bandwidth after the first accumulated down sampling satisfy the relationship: fsl>2*f, [f1, f2] is the capture range, (4) reading the data in the exterior SDRAM through an FIFO-OUT module, and multiplying data of two branches of sin and cos generated by a second NCO; (5) entering a second down-sampling through the LPF; (6) executing FFT procession; and (7) adding a second NCO localoscillating frequency, repeating the steps (4)-(6) for n times to find out a maximum value in the results of step (6) and obtain the final frequency offset result. The method according to the invention has the advantages of adaptability for low signal-to-noise ratio, wide capture range and quick capture on the base that the modulation degree characteristic of the universal controlling signal is satisfied.

Description

Fast carrier capture method with low signal-noise ratio

Technical field

The present invention relates to a kind of low signal-to-noise ratio high-frequency offset carrier quick capturing method, be mainly used in the fast carrier capture under the standard USB observing and controlling system, solved the problem of fast carrier capture under the lower big offset frequency situation of signal to noise ratio.

Background technology

Unified S frequency range (USB) observing and controlling system is a kind of observing and controlling system of maturation, and it can be finished angle tracking simultaneously, tests the speed, range finding, remote measurement, remote control and function of voice communication.Along with the continuous development of digital communication system, digitlization capturing carrier technology becomes a common technology wherein.For the estimation and the rectification of demodulating process medium frequency side-play amount, existing ripe algorithm, as phase-locked loop (PLL) method, methods such as automatic frequency control (AFC) method and Frequency Estimation.Owing to contain Doppler frequency shift in the intermediate-freuqncy signal, PLL, AFC or some are applicable to that high speed Frequency Estimation method can not finish the task that frequency deviation value is caught fast under low noise, and the condition that also can't adapt to the capture range broad, this opens up new method with regard to needs and solves this problem.On the frequency spectrum of unified USB measurement and control signal, the amplitude of residual carrier is maximum in modulation degree when not being very big, therefore can utilize this specific character to extract the frequency values of carrier wave; Simultaneously the frequency spectrum of measurement and control signal is symmetrically distributed on frequency domain, also can estimate centre frequency according to this symmetrical structure, both frequency values of carrier wave, but the common speed of this method is slower, and the memory space of data is bigger, need take big hardware resource.Article " the method research of USB answering machine fast carrier capture ", Zhang Chunlong etc., 2003 national emulation technology academic meeting paper collection, characteristic according to the frequency symmetry, method with FFT-DFT combination is judged the frequency of carrier wave, the resolution of this method is low, and capture range is narrow, the characteristic that can not satisfy wide capture range simultaneously and catch fast.

Summary of the invention

Technology of the present invention is dealt with problems and is: overcome the deficiencies in the prior art, a kind of fast carrier capture method with low signal-noise ratio is provided, this method has the advantages that to take into account low signal-to-noise ratio, wide capture range, catch fast on the basis of satisfying general measurement and control signal modulation degree characteristic.

Technical solution of the present invention is: fast carrier capture method with low signal-noise ratio comprises the following steps:

(1) coS that produces of an input signal and a NCO (digital controlled oscillator) props up circuit-switched data and multiplies each other, and wherein the local frequency of a NCO is f _NCO1, the frequency values of input signal is f ₀+ Δ f, f ₀Be the centre frequency of input signal, Δ f is the frequency deviation value of input signal;

(2) step (1) multiplied result is carried out semi-band filtering and handle the filtering high fdrequency component;

(3) result of step (2) is sent into the first down-sampled module that adds up and carry out down-sampled processing, the data after the processing are temporarily stored among the outside SDRAM by the FIFO_IN module, and the storage number of data points is greater than N among the SDRAM _FFT* L, wherein $L=\frac{f_{s1}}{f_{s2}}$ Be first down-sampled the counting of adding up, f _S1Be first frequency after the down-sampled data that adds up, f _S2For being the first down-sampled back frequency that adds up, the first data rate f that adds up after the down-sampled processing _S1And should satisfy between the maximum acquisition bandwidth f: f _S12 * f, [f ₁, f ₂] be capture range, f ₁Be negative direction minimum frequency value, f ₂Be the maximum acquisition bandwidth of positive direction, f=f ₂-f ₁

(4) by FIFO_OUT the data among the outside SDRAM are read in, and multiply each other the local frequency f of the 2nd NCO with the sin of the 2nd NCO generation and the data of two branch roads of cos _NCO2Variable, the local frequency f of the 2nd NCO _NCO2Initial value be 0;

(5) step (4) result is passed through LPF (low pass filter) respectively, the filtering high fdrequency component enters the second down-sampled module that adds up then and carries out the down-sampled processing second time, and down-sampled purpose is to improve signal to noise ratio for the second time, the data rate f after for the second time down-sampled _S2Local carrier frequency N with the 2nd NCO _CO2The step value Δ between relation should satisfy: f _S2=2 * Δ;

(6) result of step (5) is carried out N _FFTThe FFT of point handles, wherein N _FFTBe counting that FFT handles; N wherein _FFTValue determined the precision of the frequency deviation value that calculates, N _FFTBig more, precision is high more, but while N _FFTCan not be too big, because N _FFTBig more, the resource that program takies is big more, calculates I simultaneously ²+ Q ²Maximum, and record maximum pairing frequency deviation value f _m, wherein the I road is a real part, the Q road is an imaginary part, m ∈ [1,2 ... .., n];

(7) the 2nd NCO local frequency f in the increase step (4) _NCO2, repeating step (4)～(6) n time, wherein $n=\mathrm{int}[\frac{f_2-f_1}{\mathrm{\Δ}}+1],$ Find out the maximum f among n repeating step (6) result _i, i ∈ [1,2 ... .., n], final frequency deviation value result is:

The principle of the inventive method: Mathematical Modeling is as follows, k wherein ₁, k ₂, k ₃Be constant, ω _cBe the centre frequency of input signal, Δ ω is a frequency offseting value:

If the expression formula of input signal is:

a(t)＝k ₁sin(ω _ct+Δω)　　　　(1)

The signal that the one NCO produces is:

b(t)＝k ₂cos(ω _ct) (2)

The result was respectively after input signal and local carrier multiplied each other:

$a\left(t\right)*\left(t\right)=k_1\sin ({\mathrm{\ω}}_{c}t+\mathrm{\Δ\ω})\×k_2\cos \left({\mathrm{\ω}}_{c}t\right)$

(3)

$=\frac{1}{2}{k}_1{k}_2[\sin (2{\mathrm{\ω}}_{c}t+\mathrm{\Δ\ω})+\sin \left(\mathrm{\Δ\ω}\right)]$

Through behind the half-band filter, the high fdrequency component in the filtered signal is output as:

$s\left(t\right)=\frac{1}{2}{k}_1{k}_2\sin \left(\mathrm{\Δ\ω}\right)---\left(4\right)$

The two-way orthogonal signalling that the 2nd NCO module produces are respectively, wherein

Being the local carrier frequency of the 2nd NCO, is variable, and the each data that produce with the 2nd NCO of signal multiply each other, the local carrier of the 2nd NCO step value Δ that adds up:

c _I(t)＝k ₃cos(φ) (5)

c _Q(t)＝k _3sin(φ) (6)

Signal through first add up down-sampled after, deposit among the outside SDRAM by FIFO_IN, the two-way orthogonal data that data of reading in from FIFO_OUT and the 2nd NCO produce multiplies each other, and obtains:

Wherein, $A=\frac{1}{2}{k}_1{k}_2{k}_3.$

Through LPF with the high fdrequency component filtering, carry out then second add up down-sampled after, be equivalent to pass through a comb filter, improved signal to noise ratio in the image frequency component the filtering band outside, at this moment data are output as:

Give FFT module as I, Q two paths of data respectively with (9), (10), carry out FFT and handle.

The present invention's beneficial effect compared with prior art is:

(1) the present invention has solved the big data quantity storage problem that in the past occurred because of resource-constrained in the FPGA sheet by using the method for outside SDRAM, has improved acquisition speed simultaneously;

(2) the present invention adopts the two-stage down-sampled processing that adds up, and adopts FFT Frequency Estimation method, when reducing storage and counting, has improved signal to noise ratio, can catch accurately under than the low signal-to-noise ratio situation;

(3) the present invention adopts two-stage NCO processing mode, the 2nd NCO local carrier frequency is variable, has increased the frequency acquisition scope, and has solved the problem that needs repeatedly to store data in the frequency sweep facture in the past, only needs storage one secondary data to get final product among the present invention.

(4) the present invention is mainly used in the capturing carrier under the standard USB observing and controlling system, has solved the problem of fast carrier capture under than low signal-to-noise ratio, big offset frequency situation.This method is through actual test, prove its have capture time short, catch advantages such as precision height, capture range be wide.Its capture range can reach-300KHz～+ 300KHz, capture time is less than 1S, C/N ₀The minimum 35dB that reaches.

Description of drawings

Fig. 1 is the flow chart of fast carrier capture method with low signal-noise ratio of the present invention.

Fig. 2 is the operation principle block diagram of two fifo modules of the present invention.

Embodiment

With integrated test high speed remote sensing data receiving and processing equipment and many systems detection and control terminal is that example illustrates specific implementation process of the present invention, as shown in Figure 1,

(1) in this method, the hardware system processing clock is 110MHz, and capture range [300,300] KHz ,-300KHz are negative direction minimum frequency value, and 300KHz is the positive direction maximum frequency values, and the local frequency step value Δ of the 2nd NCO is 2.5KHz;

(2) frequency values is f ₀The Cos that the input signal of+Δ f and a NCO module produce props up circuit-switched data and multiplies each other, and wherein the local frequency of a NCO is f _NCO1Be 39.6975MHz, f ₀=40MHz is the centre frequency of input signal, and Δ f=300KHz is the frequency deviation value of input signal;

(3) step (2) multiplied result is carried out semi-band filtering and handle the filtering high fdrequency component;

(4) result of step (3) is sent into the first down-sampled module that adds up and carry out down-sampled processing, the first data rate f that adds up after the down-sampled processing _S1And should satisfy between the maximum frequency deviation value f: f _S12 * f, so f=f ₂-f ₁=600KHz considers that down-sampled the counting that add up should be chosen for positive integer, therefore gets f _S1=1.718750MHz, down-sampled the counting of promptly adding up for the first time is the 110MHz/1.71875MHz=64 point;

(5) data after step (4) processing are temporarily stored among the outside SDRAM by the FIFO_IN module, the storage number of data points is greater than N _FFT* L, wherein $L=\mathrm{int}[\frac{f_{s1}}{f_{s2}}+0.5]=344$ Be down-sampled the counting of adding up for the second time, in actual applications, the storage number of data points is 360000 points, and the storage needed time of data is about 0.21s;

(6) by the FIFO_OUT module data among the outside SDRAM are read in, and multiply each other the local frequency f of the 2nd NCO with the sin of the 2nd NCO generation and the data of two branch roads of cos _NCO2Variable, the local frequency f of the 2nd NCO _NCO2Initial value be 0;

(7) step (6) result is passed through LPF respectively, the filtering high fdrequency component enters the second down-sampled module that adds up then and carries out the down-sampled processing second time, and down-sampled purpose is to improve signal to noise ratio for the second time, the data rate f after for the second time down-sampled _S2Local carrier frequency f with the 2nd NCO _NCO2The step value Δ between the pass be: f _S2=2 * Δ=5KHz, second down-sampled count into $L=\mathrm{int}[\frac{f_{s1}}{f_{s2}}+0.5]=344,$ This moment, data clock was reduced to 320KHz by 110M Hz, and data rate is reduced to 5KHz by 1.71875MHz, and clock sampling 1024 point data of 320KHz need the time to be about 3.2ms;

(8) result of step (7) is carried out N _FFT=1024 FFT handles, wherein N _FFTBe counting that FFT handles; N wherein _FFTValue determined the precision of the frequency deviation value that calculates, N _FFTBig more, precision is high more, but while N _FFTCan not be too big, because N _FFTBig more, the resource that program takies is big more, and in sum, choosing counting of FFT in the practical application is 1024 points, and the resolution behind the FFT can reach 4.88Hz;

(9) I, Q two paths of data after the FFT processing are calculated I ²+ Q ²Maximum, and record maximum pairing frequency deviation value f _m, wherein the I road is a real part, the Q road is an imaginary part, m ∈ [1,2 ... .., n], it is 29.976us that each FFT handles the needed time;

(10) the 2nd NCO local frequency f in the increase step (6) _NCO2, repeating step (6)～(9) n time, wherein $n=\mathrm{int}[\frac{f_2-f_1}{\mathrm{\Δ}}+1]=241,$ Find out the maximum f among 241 repeating steps (9) result _i, i ∈ [1,2 ... .., 241], final frequency deviation value result is:

Δ f=(f ₀-f _NCO1)-(i * Δ+f _i)=299.99997KHz, with the actual value deviation be 0.03Hz;

(11) needed capture time is: 0.21s+241* (29.976us+3.2ms)=0.988s, promptly capture time is less than 1s.

As shown in Figure 2, to be the method that adopts time-sharing multiplex become a big Capacity FIFO Memory with the SDRAM emulation of FPGA outside for FIFO_IN, FIFO_OUT module functions.The 1Kx16bit FIFO that the input and output port of FIFO has FPGA BIock RAM to constitute respectively.FIFO_IN is responsible for the data of the outside input of buffer memory, and gives inner sdram controller with data passes.FIFO_OUT is responsible for the data that the buffer memory sdram controller provides, and these data can be read away by external logic from output port.Sdram controller is responsible for monitoring state and control and the read-write SDRAM of two FIFO.When among the FIFO_IN data being arranged, sdram controller is about to these data and reads and write among the SDRAM, has clearance spaces to be in preserving new data and FIFO_OUT among the SDRAM, and FIFO_OUT is read and write to sdram controller with the data among the SDRAM.Like this, under the time-sharing work of sdram controller, it is a total FIFO that the input and output port of FIFO is modeled into.As long as input and output data transfer rate summation is no more than the interface data rate of SDRAM, can keep this FIFO steady operation.

The frequency deviation value that table 1 obtains for this method Practical Calculation and the table of comparisons of ideal value.

Frequency deviation value (MHz)	C/N0(dB)	Measured value (KHz)	With theoretical value deviation (Hz)
				39.88407	35	186.572269	2.269
39.99031	34	292.812512	2.512
				39.99031	36	292.812512	2.512
40.30	35	602.49997	0.03
				40.0218	35	324.301744	1.744
39.70	35	2.499977	0.023
				39.70	36	2.499977	0.023
40.00	35	302.50485	4.85

In sum, it is as follows to reach a conclusion: method for capturing carrier of the present invention can correctly be caught when C/N0 〉=35dB.

The part that the present invention does not elaborate belongs to techniques well known.

Claims (2)

1. fast carrier capture method with low signal-noise ratio is characterized in that step is as follows:

(1) cos that produces of an input signal and a NCO props up circuit-switched data and multiplies each other, and wherein the local frequency of a NCO is f _NCO1, the frequency values of input signal is f ₀+ Δ f, f ₀Be the centre frequency of input signal, Δ f is the frequency deviation value of input signal;

(2) step (1) multiplied result is carried out semi-band filtering and handle the filtering high fdrequency component;

(3) result of step (2) is sent into the first down-sampled module that adds up and carry out down-sampled processing, the data after the processing are temporarily stored among the outside SDRAM by the FIFO_IN module, the first data rate f that adds up after the down-sampled processing _S1And should satisfy between the maximum acquisition bandwidth f: f _S1＞2 * f, [f ₁, f ₂] be capture range, f ₁Be negative direction minimum frequency value, f ₂Be the positive direction maximum frequency values, maximum acquisition bandwidth f=f ₂-f ₁

(4) by the FIFO_OUT module data among the outside SDRAM are read in, and multiply each other the local frequency f of the 2nd NCO with the sin of the 2nd NCO generation and the data of two branch roads of cos _NCO2Variable, the local frequency f of the 2nd NCO _NCO2Initial value be 0;

(5) step (4) result is passed through LPF respectively, the filtering high fdrequency component enters the second down-sampled module that adds up then and carries out the down-sampled processing second time, the data rate f after for the second time down-sampled _S2Local carrier frequency f with the 2nd NCO _NCO2The step value Δ between relation should satisfy: f _S2=2 * Δ;

(6) result of step (5) is carried out N _FFTThe FFT of point handles, wherein N _FFTBe counting that FFT handles;

(7) the 2nd NCO local frequency f in the increase step (4) _NCO2, repeating step (4)～(6) n time, wherein

Find out the maximum fi among n repeating step (6) result, i ∈ [1,2 ... .., n], final frequency deviation value result is:

2. fast carrier capture method with low signal-noise ratio according to claim 1 is characterized in that: the FIFO_IN module in the described step (3) is temporarily stored into number of data points among the outside SDRAM greater than N _FFT* L, wherein

Be down-sampled the counting of adding up for the second time, f _S1Be the first down-sampled back data rate that adds up, f _S2It is the second down-sampled back data rate that adds up.

Images