CN105846835A - Radio multi-bandpass signal receiving method of software - Google Patents

Abstract

This invention provides a radio multi-bandpass signal receiving method of a software. The method comprises the following steps: sampling and separating multi-band radio frequency bandpass signals; performing second-order bandpass sampling on to-be-processed multi-band radio frequency bandpass signals through a first sampling stream and a second sampling stream; setting a sampling frequency so that only two signals are permitted to be aliased in a same one frequency domain after the signals are sampled; introducing a time-delay difference between the first and the second sampling streams to form two paths of sampled signals with a phase difference; designing an anti-aliasing filter according to the phase difference to adjust the phases of two signals so that one signal is zero after being superposed and the other signal is constant; and thus, achieving the separation of two aliased signals. By adopting the method, the signals located at random positions can be received without aliasing by using the fixed sampling frequency; and since the sampling frequency does not need to be changed frequently according to signals located at the different positions in the receiving process, the analog front end is simplified.

Description

A kind of software radio many bandpass signals method of reseptance

Technical field

The present invention relates to Software Radio and receive field, especially a kind of software radio many bandpass signals method of reseptance.

Background technology

Software radio is a kind of based on Modern Communication Theory, with Digital Signal Processing as core, with microelectric technique For the new wireless communication systems structure supported.The core concept of software radio is to make function application software as much as possible Realizing, simplify analog portion as far as possible, therefore, software radio processes requirement and has higher sampling radiofrequency signal Speed and precision, the application of bandpass sample theory can be substantially reduced required radio frequency sampling speed, for real-time place below Reason is laid a good foundation.Bandpass sampling theory is proposed in 1991 by R.G.Vaughan the earliest, in recent years due to AD The development of Sampling techniques, sampling rate and precision improve constantly, and bandpass sampling becomes the strong theory realizing software radio Support.But it is existing Software Radio platform application bandpass sampling theory and few, with existing frequently-used software radio As a example by Universal peripheral (USRP), using zero intermediate frequency sample mode, the hardware designs part of sampler is complex. And simply do the work such as traditional Digital Down Convert at signal process part, do not carry out frequency overlapped-resistable filter Design, it is impossible to the many bandpass signals with aliasing are received, limit the versatility of software radio.

In military and commercial applications, generally require and process the multiple radiofrequency signals on different frequency bands simultaneously, select suitably Sample frequency is the difficult point receiving multi-band signal.Most scholars is when processing multi-band signal, and main consideration exists On the premise of frequency spectrum does not occurs aliasing, select alap sample frequency to alleviate the burden that back-end digital processes.A lot Scholar is also dedicated to find new algorithm to simplify Frequency Select procedures, but, these methods are inevitable for avoiding producing aliasing The selection of meeting limited samples frequency, the most loaded down with trivial details calculating process too increases the difficulty of realization, and sample frequency is the lowest Sampling precision is the highest to the requirement of preposition analog RF band filter, thus simply pursue low sample frequency not It it is optimum method.Development and the enhancing of computer process ability along with hardware such as ADC, it is possible to allow higher adopting Sample frequency.

Summary of the invention

Goal of the invention: for solving above-mentioned technical problem, the present invention proposes a kind of software radio many bandpass signals method of reseptance, Achieve the requirement that many bandpass signals are received by software radio without aliasing.

Technical scheme: for realizing above-mentioned technique effect, the technical scheme that the present invention proposes is: a kind of many bands of software radio Messenger method of reseptance, the method receives the multiple radio frequency band messengers on different frequency bands simultaneously, to multiband bandpass signal First sample, rear separating treatment, it is allowed to have the frequency spectrum of two bandpass signals to overlap after sampling in same frequency range；

The method comprising the steps of:

(1) pending multi-frequency band radio-frequency bandpass signal is carried out second order bandpass sampling by first, second sample streams, arrange Sample frequency so that only allow two signal generation aliasings after sampling in same frequency domain；Between first, second two-way sample streams Introduce delay inequality, form two-way and have dephased sampled signal；

(2) if the signal after Cai Yang does not occur aliasing, then the signal after sampling is carried out down coversion and is converted into baseband signal, And each signal in multiband bandpass signal is separated one by one；If there have two signals to occur after Cai Yang in same frequency domain to be mixed Folded, then enter step (3)；

(3) for two Design of Signal frequency overlapped-resistable filters that aliasing occurs in same frequency domain:

Two signals of aliasing are occurred to be respectively R after definition sampling₀(f) and R₁F (), between first, second sample streams Delay inequality is Δ T；Two signals frequency spectrum after the first sample streams sampling is R_AF (), samples through the second sample streams After frequency spectrum be R_B(f)；R_B(f) and R_AF () meets relation:

$\left\{\begin{array}{c}{R}_A\left(f\right)=R_{0A}\left(f\right)+R_{1A}\left(f\right)\\ R_B\left(f\right)=R_{0B}\left(f\right)+R_{1B}\left(f\right)\\ R_B\left(f\right)=R_{0A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_0}+R_{1A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_1}\end{array}\right.,(n-1/2)f_s<\left|f\right|<(n+1/2)f_s---\left(1\right)$

In formula, n₀And n₁It is respectively R₀(f) and R₁(f) location index value in frequency field；

The phase contrast 2 π Δ Tf formed after second order bandpass sampling for two-way multiband bandpass signal_sn₀With 2 π Δ Tf_sn₁If Count three filter unit S_A(f)、WithWherein, S_A(f) andUnit meets following condition:

$\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A+}(f-2n_{0}B)\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A-}(f-2n_{0}B)\end{array}---\left(2\right)$

And, $\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=0\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=0\end{array}---\left(3\right)$

S_A(f) andUnit meets following condition:

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A+}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A-}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=0$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=0$

Wherein, B is the bandwidth of multiband bandpass signal, R_0A+(f) and R_0A-F () represents R respectively_A(f) positive and negative frequency domain In frequency spectrum；R_0B+(f) and R_0B-F () represents R respectively_BFrequency spectrum in (f) positive and negative frequency domain；C represents amplitude gain.

(4) S is used_AF () is to R_AF () is filtered, useWithRespectively to R_BF () is filtered； By R_AF () is through S_A(f) filtered signal respectively with R_BF () passes throughFiltered signal and R_BF () passes throughFiltered signal is superimposed, it is achieved the separation of two signals；

(5) after step (4) carries out anti-aliasing process to two signals, under two signals after separating are carried out respectively Frequency-conversion processing, is converted into baseband signal.

Concrete, described multiband bandpass signal is: multiple bandpass signals of frequency spectrum non-overlapping copies before sampling.

Further, in described step (1), the sample frequency of second order bandpass sampling meets following condition:

$\left\{\begin{array}{c}{n}_k\&CenterDot;f_s-f_s/2<f_{lk}<f_{uk}<n_k\&CenterDot;f_s+f_s/2\\ \left|f_{Ncb}-f_{Nca}\right|\&GreaterEqual;\frac{B_b+B_a}{2}\end{array}\right.$

Wherein, n_kFor the location index of the kth signal in multi-frequency band radio-frequency bandpass signal, f_sFor sample frequency, f_lkWith f_ukIt is respectively the minimum and highest frequency of bandpass signal k, B_aAnd B_bIt is respectively the place of the two-way radio frequency band messenger after sampling Reason bandwidth, wherein, B_aAnd B_bAllow to there are two signals in processing bandwidth；f_NcaAnd f_NcbIt is respectively the two-way after sampling Bandpass signal signal center frequency in the first Nyquist zone.

Further, S in described step (3)_A(f)、WithExpression formula be respectively as follows:

$S_A\left(f\right)=\left\{\begin{array}{cc}1/B& \left|f\right|<B\\ 0& otherwise\end{array}\right.$

$S_B^0\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n0}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_0}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.$

$S_B^1\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n_1}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_1}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.$

$\mathrm{\&beta;}=e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s}.$

Beneficial effect: compared with existing Software Radio reception technique, present invention have the advantage that owing to using Second order Band-pass Sampling Technology, it is achieved that receiving without aliasing many bandpass signals in software radio.Use fixing adopting Sample frequency just can realize receiving without aliasing of optional position signal, need not for diverse location signal during receiving Change sample frequency frequently, simplify AFE (analog front end).As long as adjusting frequency overlapped-resistable filter in real time for diverse location signal Parameter just can realize two Signal separator, it is not necessary to change hardware, improve versatility and the motility of software radio.

Accompanying drawing explanation

Fig. 1 is the inside schematic diagram of embodiments of the invention；

Fig. 2 is the functional module structure figure that in the embodiment of the present invention, Software Radio receives device；

Fig. 3 is bandpass sampling signal spectrum figure in the embodiment of the present invention；

Fig. 4 is the hardware platform schematic diagram that in the embodiment of the present invention, Software Radio receives device；

Fig. 5 is the functional module structure figure of second order bandpass sampling module in the embodiment of the present invention；

Fig. 6 is the cut-away view of FPGA in the embodiment of the present invention；

Fig. 7 is the structure chart of GUN Radio module in the embodiment of the present invention.

Detailed description of the invention

Assuming that radio frequency band messenger to be sampled is R (f), it carries a width of B.Use sample frequency f_sFor f_s=2B.Institute Having the signal in frequencies below region to be all defined as the index signal for n, we define n is location index:

(n-1/2)f_s＜ | f | ＜ (n+1/2) f_s

Through bandpass sampling, all indexes are that the signal of n can be mapped in frequency range-B ＜ f ＜ B.Adopt according to band is logical Sample principle, the bandpass signal of optional position can recover in 0 ＜ f ＜ B.But, for being positioned at different index position The many bandpass signals put, may produce overlap, as shown in Figure 3 after bandpass sampling in 0 ＜ f ＜ B.Existing Technology, when processing multi-band signal, mainly considers on the premise of frequency spectrum does not occurs aliasing, selects alap sampling Frequency is to alleviate the burden that back-end digital processes.A lot of scholars are also dedicated to find new algorithm and selected to simplify frequency Journey, but, these methods produce premised on aliasing avoiding, will necessarily the selection of limited samples frequency, the most loaded down with trivial details Calculating process too increases the difficulty of realization, and the lowest sampling precision of sample frequency is to preposition analog RF band filter Requirement the highest.

For solving above-mentioned technical problem, the present invention proposes a kind of software radio many bandpass signals method of reseptance, receives not simultaneously With the multiple radio frequency band messengers on frequency band, multiband bandpass signal is first sampled, rear separating treatment, it is allowed to after sampling The frequency spectrum having two bandpass signals in same frequency range overlaps, it is possible to alleviate the burden of AFE (analog front end), and the method includes step Rapid:

(1) pending multi-frequency band radio-frequency bandpass signal is carried out second order bandpass sampling by first, second sample streams, arrange Sample frequency so that only allow two signal generation aliasings after sampling in same frequency domain；Between first, second two-way sample streams Introduce delay inequality, form two-way and have dephased sampled signal；

(2) if the signal after Cai Yang does not occur aliasing, then the signal after sampling is carried out down coversion and is converted into baseband signal, And each signal in multiband bandpass signal is separated one by one；If there have two signals to occur after Cai Yang in same frequency domain to be mixed Folded, then enter step (3)；

(3) for two Design of Signal frequency overlapped-resistable filters that aliasing occurs in same frequency domain:

Two signals of aliasing are occurred to be respectively R after definition sampling₀(f) and R₁F (), between first, second sample streams Delay inequality is Δ T；Two signals frequency spectrum after the first sample streams sampling is R_AF (), samples through the second sample streams After frequency spectrum be R_B(f)；R_B(f) and R_AF () meets relation:

$\left\{\begin{array}{c}{R}_A\left(f\right)=R_{0A}\left(f\right)+R_{1A}\left(f\right)\\ R_B\left(f\right)=R_{0B}\left(f\right)+R_{1B}\left(f\right)\\ R_B\left(f\right)=R_{0A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_0}+R_{1A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_1}\end{array}\right.,(n-1/2)f_s<\left|f\right|<(n+1/2)f_s---\left(1\right)$

In formula, n₀And n₁It is respectively R₀(f) and R₁(f) location index value in frequency field；

The phase contrast 2 π Δ Tf formed after second order bandpass sampling for two-way multiband bandpass signal_sn₀With 2 π Δ Tf_sn₁If Count three filter unit S_A(f)、WithWherein, S_A(f) andMeet following condition:

$\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A+}(f-2n_{0}B)\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A-}(f-2n_{0}B)\end{array}---\left(2\right)$

And, $\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=0\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=0\end{array}---\left(3\right)$

S_A(f) andMeet following condition:

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A+}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A-}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=0$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=0$

Wherein, B is the bandwidth of multiband bandpass signal, R_0A+(f) and R_0A-F () represents R respectively_A(f) positive and negative frequency domain In frequency spectrum；R_0B+(f) and R_0B-F () represents R respectively_BFrequency spectrum in (f) positive and negative frequency domain；C represents amplitude gain.

(4) S is used_AF () is to R_AF () is filtered, useWithRespectively to R_BF () is filtered； By R_AF () is through S_A(f) filtered signal respectively with R_BF () passes throughFiltered signal and R_BF () passes throughFiltered signal is superimposed, it is achieved the separation of two signals；

(5) after step (4) carries out anti-aliasing process to two signals, under two signals after separating are carried out respectively Frequency-conversion processing, is converted into baseband signal.

In order to technical scheme fully, is clearly described, below in conjunction with specific embodiments and the drawings, the present invention is made Further description.

Embodiment: for realizing technique scheme, it is many that the present embodiment designs a software radio being used for realizing the program Bandpass signal receives device, and the structured flowchart of this device is as in figure 2 it is shown, include: RF front-end module U1, second order bandpass sampling Module U2, frequency overlapped-resistable filter U3, GUN Radio module U4 and clock generator U5；This device implements above-mentioned technical side The inside schematic diagram of case is as shown in Figure 1.Apply two ADC (being respectively defined as ADC A and ADC B) and clock generator U5 forms second order bandpass sampling system, and clock generator provides time delays for ADC B, and sampled signal is sent into FPGA and entered Row IF process, after realizing frequency overlapped-resistable filter, Digital Down Convert and down coversion by FPGA in the present embodiment, signal is also String conversion.Signal after conversion is sent into GNU Radio via USB interface and is further processed, such as demodulation etc..Wherein, GNU Radio obtains USB interface information by custom block and carries out serioparallel exchange and isolate each road signal.Anti-aliasing The parameter of wave filter and Digital Down Convert can be designed by the custom block in GNU Radio by spi bus.

(1) RF front-end module U1

RF front-end module U1 receives the multi-frequency band radio-frequency bandpass signal that external equipment sends, and by multi-frequency band radio-frequency bandpass signal It is sent to second order bandpass sampling module U2 sample.

(2) second order bandpass sampling module U2

Second order bandpass sampling module U2 includes two ADC (being respectively defined as ADC A and ADC B) and a clock generator U5, two ADC produce two-way sampling channel respectively, are defined as passage A and passage B.Two ADC are arranged when operation There are delay inequality Δ T, delay inequality Δ T can introduce difference in the sampled signal of two-way ADC, show as prolonging of phase angle in a frequency domain Late.

The sample frequency of second order bandpass sampling module U2 should follow below equation:

$\left\{\begin{array}{c}{n}_k\&CenterDot;f_s-f_s/2<f_{lk}<f_{uk}<n_k\&CenterDot;f_s+f_s/2\\ \left|f_{Ncb}-f_{Nca}\right|\&GreaterEqual;\frac{B_b+B_a}{2}\end{array}\right.$

Wherein, n_k·f_s-f_s/ 2 ＜ f_lk＜ f_uk＜ n_k·f_s+f_sIn/2, n_kFor in multi-frequency band radio-frequency bandpass signal The location index of k signal, f_sFor sample frequency, f_lkAnd f_ukIt is respectively the minimum and highest frequency of bandpass signal k, this After formula limited samples, signalling mirror is as aliasing；

Also should meet simultaneously for many bandpass signals:This formula can be used for avoiding sampling There is two or more signal overlap in rear signal.Wherein, B_aAnd B_bIt is respectively the process of the two-way radio frequency band messenger after sampling Bandwidth；f_NcaAnd f_NcbIt is respectively the signal center frequency in the first Nyquist zone of the two-way bandpass signal after sampling, The present invention allows B_aOr B_bTwo aliasing signals of interior existence.

If the signal after Cai Yang does not occur aliasing, then the signal after sampling is carried out down coversion and be converted into baseband signal, and will be many Each signal in frequency range bandpass signal separates one by one；If having two signal generation aliasings after Cai Yang in same frequency domain, then Carry out analysis below process:

Signal after two-way is sampled by we is respectively defined as R_A(f) and R_B(f).Both meets:

$\left\{\begin{array}{c}{R}_A\left(f\right)=R_{0A}\left(f\right)+R_{1A}\left(f\right)\\ R_B\left(f\right)=R_{0B}\left(f\right)+R_{1B}\left(f\right)\\ R_B\left(f\right)=R_{0A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_0}+R_{1A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_1}\end{array}\right.,(n-1/2)f_s<\left|f\right|<(n+1/2)f_s$

IfThen

The spectrogram of second order bandpass sampling two passage is respectively such as Fig. 3 (a), shown in (b), it can be seen that in passage A two Signal spectrum does not has phase angle difference, and introduces the principle of sampling delay, two signal R due to ADC B in passage B₀(f) and R₁F the frequency spectrum of () creates phase angle difference the most respectivelyWithThe phase angle difference that we utilize two passages different designs anti-aliasing filter Ripple device S_A(f)、WithAdjust two signal phases so that S_A(f) andOutput signal superposition after Signal R₀F () is zero, and another signal R₁F () is constant, it is achieved signal R₁The separation of (f)；S_A(f) andDefeated Go out signal R after Signal averaging₁F () is zero, it is achieved signal R₀The separation of (f).

(3) frequency overlapped-resistable filter

The two ways of digital signals that Second Order Sampling module obtains realizes eliminating through the frequency overlapped-resistable filter of design in FPGA and mixes Folded process.Frequency overlapped-resistable filter design principle is as follows:

By frequency overlapped-resistable filter S_A(f) and S_BF passage A and passage B is designed and be respectively applied to ().Then, the letter of recovery Number frequency spectrum becomes:

R (f)=B [S_A(f)·R_B(f)+S_B(f)·R_B(f)] (4)

Being composed by the positive frequency spectrum and sub-frequency splitting signal, formula (1) can be transformed to:

R (f)=B [S_A(f)·(R_A+(f)+R_A-(f))+S_B(f)·(R_B+(f)+R_B-(f))] (5)

For every paths, all exist from n₀And n₁The signal of frequency location, therefore, formula (1) can enter one Step is decomposed into:

$\begin{array}{c}R\left(f\right)=B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;\left(R_{0A+}\right(f)+R_{1A+}(f)+R_{0A-}(f)+R_{1A-}(f\left)\right)\\ +S_B\left(f\right)\&CenterDot;\left(R_{0B+}\right(f)+R_{1B+}(f)+R_{0B-}(f)+R_{1B-}(f\left)\right)\&rsqb;\end{array}---\left(6\right)$

Receive two paths of signals R simultaneously₀(f) and R₁F the receptor of () includes that two adjustable radio-frequency filters are for selecting two Bandpass signal, is input to sampler, two paths of signals R after superposition₀(f) and R₁F () is sampled simultaneously, three can weigh Newly configured wave filter S_A(f)、It is used for eliminating signal aliasing, and output R arranged side by side₀(f) and R₁(f)。

Wherein, wave filter S_A(f) andIt is designed for recovering R₀(f) and eliminate R₁(f).Therefore, wave filter S_A(f) WithShould meet:

$\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=C\&le;R_{0A+}(f-2n_{0}B)\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A-}(f-2n_{0}B)\end{array}---\left(7\right)$

With

$\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=0\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=0\end{array}---\left(8\right)$

We select S_AF () is simplest form, it may be assumed that

$S_A\left(f\right)=\left\{\begin{array}{cc}1/B& \left|f\right|<B\\ 0& otherwise\end{array}\right.---\left(9\right)$

Therefore, can be obtained by solution formula (7) and (8)Such as following formula:

$S_B^0\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n_1}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_1}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.---\left(10\right)$

In order to receive signal R₁(f) and eliminate R₀F (), needs to design another wave filterUtilize formula (7), (8) (9) the same method shown in can obtainAs shown in formula (11):

$S_B^1\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n_0}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_0}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.---\left(11\right)$

Formula (9), (10), (11) are wave filter expression formula in frequency domain, permissible by formula (9), (10), (11) Obtain their impulse response: show that filter impulse response is:

$S_A\left(t\right)={\&Integral;}_{f_l}^{f_h}{S}_A\left(f\right)e^{j2\mathrm{\&pi;}ft}df$

$S_B^0\left(t\right)={\&Integral;}_{f_l}^{f_h}{S}_B^0\left(f\right)e^{j2\mathrm{\&pi;}ft}df$

$S_B^1\left(t\right)={\&Integral;}_{f_l}^{f_h}{S}_B^1\left(f\right)e^{j2\mathrm{\&pi;}ft}df$

Wherein, f_l、f_hIt is respectively low-limit frequency and the highest frequency of frequency overlapped-resistable filter.By to above-mentioned two formulas with f_s=2B samples, and obtains sampled value and can be used as filter parameter.If we determined that low-limit frequency and the highest The numerical value of frequency is the widest signal bandwidth, when signal location index changes, still can use same S_A.But For S_B, when location index changes, parameter need to be adjusted.The coefficient that can use FIR Compiler divides Analysis instrument produces a series of filter coefficient, and controls to choose different filter coefficients by external parameter. Filter coefficient can be adjusted in real time by GNU Radio by FPGA running by spi bus.

When designing frequency overlapped-resistable filter, by adding group delay compensation, choose bigger time delays and make anti-aliasing effect More excellent.Apparatus of the present invention use the wave filter design wave filter of 121tap to make it that filtration result of aliasing is higher than 30d B. The present invention uses the way of amplitude compensation to compensate the signal output amplitude being in diverse location simultaneously, so that output increases Benefit is maintained at about 1.Solve the impact anti-aliasing performance produced because of the difference of two position input signal differences.

(4) based on FPGA design frequency overlapped-resistable filter

The present embodiment utilizes FPGA to realize frequency overlapped-resistable filter, eliminates the aliasing of two paths of signals, and is converted by down coversion For baseband signal, Fig. 6 is FPGA cut-away view.The present embodiment devises 4 railway digital down coversion passages, therefore The reception simultaneously to 4 road bandpass signals can be realized, it is allowed to wherein two-way bandpass signal occurs aliasing after sampling.Every road The I of signal and Q signal pass through and go here and there to convert and be transferred to GNU Radio via USB controller.Meanwhile, anti-aliasing filter The parameter of ripple device and Digital Down Convert also can be by USB controller through spi bus control.

In signal analysis, we often carry out resolution of vectors signal, are namely that frequency is identical, peak value by signal decomposition Amplitude is identical but two component I of phase 90 and Q component.By stating signal with vector, can intactly retouch State the amplitude of signal, frequency and phase place.DDC (Digital Down Convert) mainly completes the digital mixing to sampled signal, resists and mix Folded filtering, down-sampled (extraction) filtering etc., to obtain the baseband signal output of low sampling rate.In Fig. 6, four DDC are permissible Realize the down-converted to 4 tunnel sampled signals, process Hou Mei road signal and I branch road (component in the same direction) and Q can be divided into prop up Road (quadrature component).

FPGA module U3 by the data stream handled well by USB transmission to GNU Radio module U4, due to from DDC also continuously transmits the data of pc end by USB and have passed through parallel/serial propagation and transformation, GNU Radio module U4 Custom block possess data separating ability.

(5) GUNRadio module U4

Fig. 7 is GNU Radio signal receiving module designed in the present embodiment, and this module can be sampled by setting Signal number (1,2 or 4) selects the number of the signal received, and realizes Signal separator by serioparallel exchange.At GNU Signal can be further processed, such as signal demodulation etc. by other modules in software kit in Radio.

Described RF front-end module U1 can use Ettus company radio frequency daughter board DBSRX2, and reception band limits is 500M -2.35GHz.ADC in second order bandpass sampling module U2 uses TI company ADS54RF63 series, and clock is sent out Raw device uses FPGA to use AD9513 series, it is possible to provide maximum clock signal 800MHz, and mean square deviation shake is less than 0.3ps, adjustable delay 2250ps.Usb interface controller uses CYPRESS company CY7C68013A-100AXC FX2 controller.GNU Radio can install GNU Radio-3.4.2 version under Linux system.Hardware platform schematic diagram As shown in Figure 4.

The above is only the preferred embodiment of the present invention, it should be pointed out that: for those skilled in the art For, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications are also Should be regarded as protection scope of the present invention.

Claims (4)

1. software radio many bandpass signals method of reseptance, it is characterised in that the method receives on different frequency bands simultaneously Multiple radio frequency band messengers, first sample to multiband bandpass signal, rear separating treatment, it is allowed in same frequency range after sampling The frequency spectrum inside having two bandpass signals overlaps；

The method comprising the steps of:

(1) pending multi-frequency band radio-frequency bandpass signal is carried out second order bandpass sampling by first, second sample streams, arrange Sample frequency so that only allow two signal generation aliasings after sampling in same frequency domain；Between first, second two-way sample streams Introduce delay inequality, form two-way and have dephased sampled signal；

(2) if the signal after Cai Yang does not occur aliasing, then the signal after sampling is carried out down coversion and is converted into baseband signal, And each signal in multiband bandpass signal is separated one by one；If there have two signals to occur after Cai Yang in same frequency domain to be mixed Folded, then enter step (3)；

(3) for two Design of Signal frequency overlapped-resistable filters that aliasing occurs in same frequency domain:

Two signals of aliasing are occurred to be respectively R after definition sampling₀(f) and R₁F (), between first, second sample streams Delay inequality is Δ T；Two signals frequency spectrum after the first sample streams sampling is R_AF (), samples through the second sample streams After frequency spectrum be R_B(f)；R_B(f) and R_AF () meets relation:

$\left\{\begin{array}{c}{R}_A\left(f\right)=R_{0A}\left(f\right)+R_{1A}\left(f\right)\\ R_B\left(f\right)=R_{\mathrm{0B}}\left(f\right)+R_{1B}\left(f\right)\\ R_B\left(f\right)=R_{0A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_0}+R_{1A}\left(f\right)\&CenterDot;e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s{n}_1}\end{array}\right.,(n-1/2)f_S<\left|f\right|<(n+1/2)f_S---\left(1\right)$

In formula, n₀And n₁It is respectively R₀(f) and R₁(f) location index value in frequency field；

The phase contrast 2 π Δ Tf formed after second order bandpass sampling for two-way multiband bandpass signal_sn₀With 2 π Δ Tf_sn₁If Count three filter unit S_A(f)、WithWherein, S_A(f) andMeet following condition:

$\begin{array}{c}B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A+}(f-2n_{0}B)\\ B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^0\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{0A-}(f-2n_{0}B)\end{array}---\left(2\right)$

And,

S_A(f) andMeet following condition:

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B+}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A+}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{1A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{1B-}\left(f\right)\&rsqb;=C\&CenterDot;R_{1A-}(f-2n_{1}B)$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A+}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B+}\left(f\right)\&rsqb;=0$

$B\&CenterDot;\&lsqb;S_A\left(f\right)\&CenterDot;R_{0A-}\left(f\right)+S_B^1\left(f\right)\&CenterDot;R_{0B-}\left(f\right)\&rsqb;=0$

Wherein, B is the bandwidth of multiband bandpass signal, R_0A+(f) and R_0A-F () represents R respectively_A(f) positive and negative frequency domain In frequency spectrum；R_0B+(f) and R_0B-F () represents R respectively_BFrequency spectrum in (f) positive and negative frequency domain；C represents amplitude gain.

(4) S is used_AF () is to R_AF () is filtered, useWithRespectively to R_BF () is filtered； By R_AF () is through S_A(f) filtered signal respectively with R_BF () passes throughFiltered signal and R_BF () passes throughFiltered signal is superimposed, it is achieved the separation of two signals；

(5) after step (4) carries out anti-aliasing process to two signals, under two signals after separating are carried out respectively Frequency-conversion processing, is converted into baseband signal.

A kind of software radio many bandpass signals method of reseptance the most according to claim 1, it is characterised in that described many Frequency range bandpass signal is: multiple bandpass signals of frequency spectrum non-overlapping copies before sampling.

A kind of software radio many bandpass signals method of reseptance the most according to claim 2, it is characterised in that described step Suddenly in (1), the sample frequency of second order bandpass sampling meets following condition:

$\left\{\begin{array}{c}{n}_k\&CenterDot;f_s-f_s/2<f_{lk}<f_{uk}<n_k\&CenterDot;f_s+f_s/2\\ |f_{Ncb}-f_{Nca}|\&GreaterEqual;\frac{B_b+B_a}{2}\end{array}\right.$

Wherein, n_kFor the location index of the kth signal in multi-frequency band radio-frequency bandpass signal, f_sFor sample frequency, f_lkWith f_ukIt is respectively the minimum and highest frequency of bandpass signal k, B_aAnd B_bIt is respectively the place of the two-way radio frequency band messenger after sampling Reason bandwidth, B_aAnd B_bAllow to there are two signals in processing bandwidth；f_NcaAnd f_NcbIt is respectively the two-way band communication after sampling Signal center frequency number in the first Nyquist zone.

A kind of software radio many bandpass signals method of reseptance the most according to claim 2, it is characterised in that described step Suddenly S in (3)_A(f)、WithExpression formula be respectively as follows:

$S_A\left(f\right)=\left\{\begin{array}{cc}1/B& \left|f\right|<B\\ 0& otherwise\end{array}\right.$

$S_B^0\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n0}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_0}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.$

$S_B^1\left(f\right)=\left\{\begin{array}{cc}\frac{-{\mathrm{\&beta;}}^{-n_1}}{B}& -B<f<0\\ \frac{-{\mathrm{\&beta;}}^{n_1}}{B}& 0<f<B\\ 0& otherwise\end{array}\right.$

$\mathrm{\&beta;}=e^{-j2{\mathrm{\&pi;\&Delta;Tf}}_s}.$