CN106130662A - A kind of wireless radio broadband signal detection method and device - Google Patents

Abstract

The invention discloses a kind of wireless radio broadband signal detection method, provide a kind of wireless radio broadband signal detection frequency-division section method for designing and compressed sensing calculation matrix physics realization method, solve in the range of high band, big bandwidth, low noise, high sensitivity and Larger Dynamic aspect technical barriers such as wide-band radio signal detectings.Wireless radio broadband signal detection frequency-division section method for designing make use of sub-band wide-band radio signal restructing algorithm complexity between the openness feature of frequency domain and different sub-band close thinking of trying one's best to construct optimization object function.LDPC check code matrix switch is used to achieve compressed sensing calculation matrix physically, this scheme may be used for communication, radar, observing and controlling and guided missile remote measurement broadband signal and scouts the fields such as integrated, cognitive electronic warfare, the invention belongs to new method, new principle invention category, relate to the field such as signal processing, optimization.

Description

A kind of wireless radio broadband signal detection method and device

Technical field

The present invention relates to radar, electronic reconnaissance signal processing field, particularly relate to a kind of wireless radio broadband signal detection side Method and device.

Background technology

Space-based, ground electronic device works frequency improve constantly, and the high band that radio signal is scouted capture device is detectd Survey ability proposes new demand.Need to solve high band, big bandwidth, low noise, high sensitivity, the broadband signal of Larger Dynamic scope The aspect technical barriers such as reception, storage and signal analysis, for expanding radio signal detecting frequency range, lifting detecting ability establishes skill Art basis.

Radio detection object can be ground, the aerial or radiation source in space, and systems scan frequency band range reaches 40- 110GHz (more than or), instantaneous frequency measurement band is wider than 8GHz, and regular receiver intermediate frequency output bandwidth is limited to the highest sampling of ADC Rate and receiver dynamic sensitivity design difficulty, be difficult to the space having improvement further, such as at four array element electronic reconnaissances In receiver, intermediate frequency output bandwidth is 500MHz, and a sample rate is the high-speed ADC of 5GHz, is configured to 4 Parallel ADC (each Sample rate is 1.25GHz), respectively receiver 4 road IF output signal is sampled, can be just under meeting twice bandwidth sample rate Often work, synchronization can be guaranteed, but wants more high-bandwidth signals of sampling further, does not require nothing more than receiver dynamic range Bigger and sensitivity is higher and it is necessary to use many high-speed ADC parallel samplings, the dynamic range of ADC is proposed more simultaneously Add the requirement of harshness, but, in High-speed Board Design, ADC chip chamber synchronizes also to be difficult to be guaranteed, regular receiver microwave Device performance and index difficulty in a short time has breakthrough.In a word, a series of technical barriers are difficult to make a breakthrough simultaneously.

The most optimal scheme should be that broadband signal directly carries out radio frequency sampling, current software radio system Unite for radiofrequency signal bandwidth narrower, all in the range of ADC sample rate, it is possible to directly sample, and for bandwidth cross over Tens GHz, for instant bandwidth reaches the radio signal of several GHz, traditional direct RF sampling scheme cannot be implemented.

Summary of the invention

In view of this, the invention provides a kind of wireless radio broadband signal detection method and device, saved system cost And power consumption, it is ensured that more ADC number of significant digit, further increase DATA REASONING precision.

There is advantages that

(1) the wireless radio broadband signal detection method of the present invention, uses the mentality of designing of frequency-division section, to radio frequency in each frequency range Signal is sampled after carrying out frequency domain compressed sensing again, largely reduces sample rate, has saved system cost and merit Consumption, it is ensured that more ADC number of significant digit, further increases DATA REASONING precision, carries for signal storage and further analysis Supply more preferable scheme, it is to avoid the strict sequential order stationary problem that multi-disc ADC parallel sampling requires；Make original for arrowband simultaneously The signal reconstruction algorithm of signal can be respectively applied in each sub-band solve,

(2) the frequency domain compressed sensing device that the present invention builds, solves high band, big bandwidth, low noise, high sensitivity With in the range of Larger Dynamic to aspect technical barriers such as wide-band radio signal receptions.

Accompanying drawing explanation

Fig. 1 is the wireless radio broadband signal detection method flow diagram of the present invention.

Fig. 2 is the hardware realization apparatus schematic diagram of the compressed sensing calculation matrix of the present invention.

Detailed description of the invention

Develop simultaneously embodiment below in conjunction with the accompanying drawings, describes the present invention.

The present invention can solve wide-band radio in the range of high band, big bandwidth, low noise, high sensitivity and Larger Dynamic Number detect receipts, sample, store, the aspect technical barrier such as channelizing and signal analysis.

Although radio signal bandwidth spans tens GHz, the most only radar signal, signal of communication also have measurement and control signal With guided missile telemetered signal etc., each signal characteristic, purposes are the most different, determine antenna pattern, signal processing complexity with And it is the most different to cover frequency range, so needing scientific and reasonable frequency-division section design, and each sub-band all shows frequency The openness feature of rate, creates condition for compressed sensing application, and such condition makes ADC lead to without the band meeting twice bandwidth Sampling thheorem, i.e. can be with relatively low sample rate and with high probability reconstruct original frequency domain sparse signal, and the establishment of degree of rarefication needs relevant Priori, relevant with concrete application background, not within the coverage of the present invention.

Assume wireless radio broadband signal at frequency-domain sparse degree K it is known that frequency-domain sparse degree be not only frequency-division section design foundation, Also being the foundation building compressed sensing calculation matrix, the present invention proposes a kind of antenna frequency-division section method for designing and compressed sensing is measured Matrix implementation.Its workflow is as follows: first different according to application background, wireless radio broadband signal is carried out frequency band and draws Point, it being divided into P sub-band, each sub-band is wide to be determined by an object function, the then wide-band radio signal of different frequency From corresponding sub-band antenna element input, after sub-band compressed sensing calculation matrix carries out multiply-add operation, export new Radiofrequency signal, then carry out ADC sampling, reduce sample rate, after digital receiver, finally carry out frequency-domain sparse signal reconstruction, Synchronously complete channelizing process.

Wireless radio broadband signal detection object includes radar signal, signal of communication, guided missile telemetered signal and measurement and control signal Deng, in different overlay areas, these signals are the most different at frequency-domain sparse degree K, and this just requires the spirit of frequency-division section method for designing Live and reach optimum in some sense.

Assuming the wireless radio broadband a width of B of signal detection band, this magnitude is adopted far beyond the ADC of business highest ranking Sample rate, so whole broadband signal cannot directly be sampled by ADC, it is necessary to uses frequency-division section design, it is assumed that bandwidth B is divided into P Individual sub-band, and B₁+B₂+…+B_P=B, gives each sub-band for solution optimum allocation bandwidth B, and the requirement of the present invention is: first First, it is ensured that noise is little, meanwhile, the accuracy of signal reconstruction wants height, and finally, in each sub-band, signal reconstruction complexity is to the greatest extent Measure close, because in wireless wide band signal detecting bandwidth, signal intensity quickly, synchronizes reconstruction signal the most crucial, with conventional Iteration assignment restructing algorithm (Iterative Shrinkage/Thresholding Algorithm, IST) as a example by illustrate, its The design philosophy of its algorithm is similar to therewith, and in each sub-band, IST algorithm computation complexity is O (2M_iN_i)∝O(2K_iN_i), M_iFor i-th sub-band observing matrix row vector dimension, N_iFor i-th sub-band observing matrix column vector dimension, K_iFor i-th Sub-band frequency-domain sparse degree, K_i∝B_i, and B_i=w_iB, w_iWeight is distributed, so meeting O (2K for bandwidth_iN_i)∝O(2w_iBN_i), OrderThen constrained optimization equation is

The object function of equation (1) is a convex optimization problem of Constrained Non-smooth surface, conventional convex relaxed algorithm such as gradient Project sparse restructing algorithm to solve.Utilize Lagrangian multiplier method, by inequality constraints l₁-norm minimum problem becomes nothing Constrained optimization problems, order

Then unconstrained optimization equation is

$\underset{(X,w,\mathrm{\λ},\mathrm{\α})}{\min }L(X,w,\mathrm{\λ},\mathrm{\α})---\left(2\right)$

λ in formula (2)_i＞ 0 is regularization parameter, for balanced signal degree of rarefication and signal reconstruction precision.X represents sampling letter Number, w represents the complexity difference weight between sub-band, and λ represents the accuracy of signal reconstruction, is respectively intended to determine frequency-region signal position Put, size and each sub-band bandwidth.Different users is for signal reconstruction complexity difference between noise sensitivity, sub-band And the requirement of signal reconstruction accuracy is different, the bandwidth of the sub-band the most finally determined is the most different.

Problem as described above, the invention discloses a kind of wireless radio broadband signal detection method, as it is shown in figure 1, I-th sub-band covers f_{I, L}-f_{I, H}Bandwidth, it is assumed that each sub-band receives system and is made up of n antenna element.In certain moment T, the signal of n the antenna element being input to i-th sub-band is expressed as x_{I, 1}(t), x_{I, 2}(t) ..., x_{I, n}(t), respectively Enter respective compressed sensing calculation matrix.In Fig. 1, calculation matrix uses LDPC check code matrix to realize, LDPC check code matrix Meet: 1) RIP iso-distance constraint condition；2) base sparse with Fourier transformation is irrelevant；3) having the most openness, rear end is translated Code complexity is low；4) LDPC code is binary system deterministic encoding structure, and analog element is built simply, and " 0 " state is electronic circuit Closed mode, one state is electronic circuit conducting state.It will be noted from fig. 1 that due in each sub-band, radio Broadband signal is sparse in frequency domain, through the compressed sensing calculation matrix built by LDPC check code matrix and adder After, complete the output of analog radio frequency compressed signal, carry out ADC sampling (now sample rate is the most decreased), it is achieved that radio Broadband signal radio frequency in each sub-band is directly sampled, and signal reconstruction achieves in each sub-band at the channelizing of signal Reason, becomes baseband signal and is further processed.

In FIG, i-th sub-band bandwidth is B_i=f_{I, H}-f_{I, L}, ADC sample rate band to be met under normal circumstances Logical sampling thheorem, i.e. R_i≥2B_i, R_iFor sample rate.Assume that the sampling time is T_i, not through the ADC sampled point of compression sampling Number is N_i=R_iT_i, the degree of rarefication of signal is K_i, then calculation matrix row vector must is fulfilled for M_i≥cK_ilog(N_i/K_i)≈3K_i, this The LDPC check code calculation matrix that sample builds isCalculation matrix design includes that matrix dimension size sets Two aspects of Element generation in meter and matrix, signal length to be allowed, number of samples and primary signal degree of rarefication three meet spy Determining relation, just can make that compression perceptual system is most economical again can be with high probability Accurate Reconstruction sparse signal.

Assume that the radio frequency analog signal of jth antenna element input is x in i-th sub-band_{I, j}(t), i=1 ..., P； J=1 ..., n, then, after LDPC calculation matrix and adder, the radio frequency analog signal of output is y_{I, j}(t)=Ax_{I, j}(t), After ADC sampling, obtain at T_iM in time interval_iIndividual sample, the ADC sample rate after compression sampling is M_i/T_i, and do not have ADC sample rate through compression sampling is compared, and the ADC sample rate after compression sampling is original M_i/N_iTimes, for radio Broadband signal detecting provides feasible scheme.

The hardware side that it is compressed sensing calculation matrix with LDPC check code calculation matrix and adder that the present invention also constructs Case, as in figure 2 it is shown, correspond respectively to LDPC matrix, adder and ADC in Fig. 1.LDPC binary measure matrix is by attenuator (" 0 " state is attenuator maximum attenuation gear, and one state is that attenuator turns on gear entirely, gear when the most not decaying) or Person's phase shifter constitutes (zero phase phase shifter composition " 1 ", i.e. e^j0=1, two back-to-back positives and anti-phase phase shifter constitute " 0 ", i.e. e^j0+e^jπ=0).

In fig. 2, radio frequency analog signal x_{I, j}T () is input to M parallel_i×N_iIndividual it is made up of attenuator (or phase shifter) In LDPC calculation matrix, attenuator (or phase shifter) serves as the effect of electronics mask, it is achieved multiplication function, through each N_iDimension It is to be realized by first adder of Fig. 2 that radio frequency analog signal after electronics mask " focuses on " function, and this process is mathematically Achieve inner product operation between a vector, radio frequency analog signal x_{I, j}T () is through M_iIndividual N_iRadio frequency analog after dimension electronics mask Signal achieves overall " focusing " of compressed sensing calculation matrix radiofrequency signal by second adder of rightmost in Fig. 2, this One process creates new radiofrequency signal y_{I, j}(t)=Ax_{I, j}(t).The principle of the most whole process is to make radio frequency analog Signal x_{I, j}T () sparse component even dispersion on frequency domain is in a complete dictionary A, namely by comprised for signal letter Breath all remains, and in follow-up signal reconstruction, can intactly recover signal all sparse components in sparse territory. Y is become after ADC samples_{I, j}[k], k=1 ..., M_i, decrease sample rate, thus utilize storage and analyze further.

In sum, these are only presently preferred embodiments of the present invention, be not intended to limit protection scope of the present invention. All within the spirit and principles in the present invention, any modification, equivalent substitution and improvement etc. made, should be included in the present invention's Within protection domain.

Claims (4)

1. a wireless radio broadband signal detection method, it is characterised in that comprise the steps:

Step 1, the bandwidth of wireless radio broadband signal reception system received are divided into p sub-band；Wherein, each son frequency Band bandwidth determines according to the actual requirements；

Step 2, make the i-th sub-band in p sub-band receives radiofrequency signal way n that receives of system_iRepresent, its Middle i=1,2 ..., p；For i-th sub-band Zhong Ge road radiofrequency signal, carry out frequency domain compressed sensing process respectively, obtain N after frequency compression_iRoad radiofrequency signal；Wherein, employing has row vector is M_i, column vector is N_iFrequency domain compressed sensing survey Moment matrix carries out frequency compression, M_i≈3K_i, K_iFor target in i-th sub-band in degree of rarefication；N_iFor receiving system requirements The number of samples within the sampling time；

N after step 3, the frequency compression obtained for step 2_iRoad radiofrequency signal is sampled respectively；Then receiver pair is used n_iRoad radiofrequency signal is received；

Step 4, the n that step 3 is received_iRoad radiofrequency signal carries out signal reconstruction respectively, and signal reconstruction has obtained signal at frequency domain Sparse distribution spectrogram, it is achieved the detecting of wireless radio broadband signal.

2. a kind of wireless radio broadband signal detection method as claimed in claim 1, it is characterised in that the width of described sub-band Can be uniform, it is also possible to be heterogeneous.

3. a kind of wireless radio broadband signal detection method as claimed in claim 1, it is characterised in that described signal reconstruction uses The sparse restructing algorithm of gradient projection solves.

4. as claimed in claim 1 a kind of wireless radio broadband signal detection method, it is characterised in that each sub-band bandwidth according to The requirement of signal reconstruction complexity difference and signal reconstruction accuracy between noise-sensitive degree, sub-band is determined by user.