Background technology
Compared to traditional phased-array radar, MIMO technology has improvement radar angle under conditions of using less physical channel
Spend the advantage of resolving power.Meanwhile from the point of view of hardware low cost, time diversity scheme (TDM) is the one kind for being easiest to realize
Mimo system.Therefore, research extensively is unfolded to TDM-MIMO radars in a large amount of document and Project Realization.
For the radar system using tradition TDM-MIMO schemes, it is considered that each target to be detected is in given time
It is inside to maintain static;Otherwise, movement will introduce the variation phase of one and velocity correlation and is coupling in echo-signal
With on the relevant phase term of angle, and then lead to the drastically decline of radar angle measurement performance.
Array element (the linear array TDM-MIMO thunder that spacing d) is formed with N number of transmitting array element (spacing Pd) is received by P
It reaches, traditional TDM systems respectively emit array element and scheme as shown in Figure 1, i.e., each transmitting array element is used to be pressed according to the spatial order of array element
Fix rule either on or off successively;Wherein pulse-recurrence time is T, modulating bandwidth B, single MIMO realize in include N number of
Pulse, a coherent processing inteval (CPI) are made of M period.
It is assumed that there is K there is different angle in given range cell, the target of friction speed, if echo-signal have been subjected to away from
The process of pulse-compression of descriscent, the echo signal data block in single CPI are as shown in Figure 2;For traditional TDM-MIMO radars r
A range cell, echo-signal is in m-th of period, n-th of pulse (while corresponding n-th of transmitting array element emits signal), pth
A signal representation form for receiving array element is as follows:
Wherein ai, viAnd θiIndicate respectively i-th target through distance to the complex magnitude after " focusing ", radial velocity and
Orientation angles;λ indicates radar emission narrow band signal centre wavelength, εmnpRepresent system additive white Gaussian noise.
The phase change item of related n in above formulaShow space
FrequencyWith Doppler frequencyIt is mutually coupled;It therefore, will necessarily be because when estimating target bearing angle
The movement of target and cause estimation mistake.Existing certain methods are using first in above formula
Doppler-frequency estimation is carried out, coupled phase item when the frequency compensation sky that estimates is passed through.But such method is in target to be estimated
When speed is fuzzy there are odd-multiple, it can not be computed correctly compensation phase, in turn result in decoupling phase error, not simply fail to solve to ask
Topic introduces additional phase term instead, further results in the angle estimation of mistake.
Invention content
The purpose of the present invention is phase coupled problems when moving-target sky for traditional TDM-MIMO radars, propose a kind of base
Phase decoupling and method of estimation when random exomonental time-division MIMO radar sky, to realize separated space frequency and Doppler
The ability of frequency preferably correctly estimates target angle with speed.
The method of the present invention is achieved through the following technical solutions:
Phase decoupling method when one kind is based on random exomonental time-division MIMO radar sky, includes the following steps:
Step 1: designing random emission measurement matrix, control transmitting array element is obtained by the random emission measurement matrix
The sequence of random switching, to will original transmitting array element spatially placement position periodically with the sequence time-division side of impulse ejection
Formula replaces with random transmitting time division way, achievees the effect that decouple moving target angle rates' information;
Step 2: the random emission measurement Matrix Multiplication is obtained perception matrix, profit with the orthogonal basic matrix of Standard Fourier
The reconstruction for carrying out Doppler frequency spectrum along distance dimension with the sparse reconstructing methods of FOCUSS, realizes independently estimating to velocity to moving target
Meter;
Step 3: doing the matched filtering processing of angle dimension along each distance-Doppler unit, the space frequency of the unit is obtained
Rate is composed, and then extracts target angle information, completes the independent estimations to target angle.
Further, the random emission measurement matrix described in step 1 is designed using following methods:
If the calculation matrix of n-th of transmitting array element is Ψn, the dimension of matrix is M × MN, the matrix by from a MN ×
It extracts M rows in the unit matrix of MN dimensions out at random to obtain, the probability drawn per a line is 1/MN, and independently of each other, and M, N are just
Integer.
Beneficial effects of the present invention:
The present invention upsets set time lag characteristic existing for periodic emission between array element using the random emission characteristics of array element, into
And the decoupling to velocity to moving target, angular phase information is completed, then utilize sparse reconfiguration technique and matched filtering technique point
It is other to Doppler frequency spectrum and the independent reconstruction of spatial frequency spectrum, finally, complete estimation to target velocity and angle information;
Due to considering the not overlapping each other property between transmitting array element when calculation matrix designs so that each array element emits not at random
Only meet TDM-MIMO operation principles, also ensures the identical weighting weight of each Virtual array.
Specific implementation mode
It elaborates below in conjunction with the accompanying drawings to the embodiment of the method for the present invention.
As shown in figure 3, a kind of distinguished based on sparse reconfiguration technique and matched filtering technique from random transmitting pulse train
Independent reconstruct target Doppler frequency spectrum and object space frequency spectrum, and then solve phase when the moving-target sky in TDM-MIMO radars
The method of coupling, detailed process are:
Step 1: designing random emission measurement matrix, the suitable of control transmitting array element random switching is obtained by calculation matrix
Original transmitting array element is spatially periodically replaced with random transmitting in placement position by sequence with the sequence time division scheme of impulse ejection
Time division scheme achievees the effect that decouple moving target angle rates' information.
Step 101 sets the calculation matrix of n-th of transmitting array element as Ψn, the dimension of matrix is M × MN, which is to pass through
It extracts M rows out at random from the unit matrix of MN × MN dimension to obtain, the probability drawn per a line is 1/MN, and mutually solely
It is vertical.
The problem can actually be equivalent to a combinatorial problem, i.e., forA possible composite sequence, the 1st transmitting array element
One-shot measurement matrix realize it is practical to choose a kind of combination from these possible combination equal probabilities, gather in serial number
As the M row serial numbers that will be extracted from unit matrix;To ensure that one and only one transmitting array element exists in same one-pulse time
Work, when designing the calculation matrix of the 2nd transmitting array element, the realization of calculation matrix is from remaining MN-M rows in unit matrix
Choose M rows again, i.e., it is equiprobable fromIt is chosen in kind combination a kind of;And so on, the calculation matrix of n-th array element is realized
The matrix of rear remaining last M rows composition is selected as in unit matrix by each transmitting array element.
Step 102, according to designed calculation matrix Ψ in step 101n, by following mapping relations by calculation matrix
Be converted to random emission control orders of the control unit to each transmitting array element.For n-th of array element, its survey known to step 101
Moment matrix ΨnIn a total of M 1, and often go one and only one.Therefore, calculation matrix and emit array element mapping relations be,
If calculation matrix ΨnI-th row, jth column element Ψi,j=1 indicates j-th of pulse in MN pulse in total, n-th of array element
Emit signal;Remaining time, the array element did not emitted signal.
To sum up, control unit can be quickly obtained the sequence that each array element emits at random by calculation matrix.Due to measuring
The not overlapping each other property between transmitting array element is considered when matrix design so that transmitting not only conforms with TDM-MIMO works to each array element at random
Make principle, also ensures the identical weighting weight of each Virtual array.
Step 2: being multiplied by the orthogonal basic matrix of Standard Fourier using calculation matrix obtains perception matrix, and then utilize
" FOCUSS " sparse reconfiguration technique carries out Doppler frequency spectrum along distance dimension and rebuilds, and realizes and independently estimates to velocity information after separation
Meter.
Step 201, to via lower mixing, amplifying, filtering, the base band echo-signal after sampling carries out process of pulse-compression
(or other distances " focusing " are handled), and it is retaken.After rearrangement battle array is received with p-th via n-th of transmitting array element transmitting
The range cell echo-signal that member receives is ynp=[x1np,x2np…xMnp]T。
Step 202, the sparse characteristic based on a range cell target, can be to echo-signal ynpIt is written as expression shape
Formula:
ynp=Ψn·Φ·an+n
Wherein the Φ matrixes of dimension MN × MN are to contain MN possible Doppler by what normal orthogonal Fourier's base formed
The steering vector of frequency;The vector an of dimension MN × 1 is to represent the corresponding target complex magnitude on be possible to frequency location
Sparse spike;N is incoherent white Gaussian noise signal, mean power δ.In general, claiming ΨnΦ is perception matrix, by
In ΨnΦ, which meets, constrains equidistant restrictive condition (RIP), and usually in a range cell, the number of moving-target is much small
In MN possible doppler positions.Therefore, the theory based on sparse reconstruct is it is found that an1 norm of following convex relaxation can be passed through
Optimization linear equation solves and (completes reconstruction process using classical " FOCUSS " sparse recovery algorithms in the present invention):
min(||an||1),||Ψn·Φ·an-ynp||≤δ
It should be noted that a rebuildnThe not fuzzy range that tests the speed of its Doppler is N times of traditional TDM schemes, and
Object to be measured speed beyond it is not fuzzy test the speed range when, this method is still effective.
Step 3: for each distance-Doppler unit for having extracted target range, velocity information, using traditional
The spatial frequency spectrum that processing can be obtained the unit is focused to the space phase information of each Virtual array with filtering method,
Namely realization is to the independent estimations of target angle information.
With reference to the field trial example of specific Ka wave bands TDM-MIMO radar systems, to the actual effect of the present invention into
Row explanation.
Radar system transmitting signal is chirp sequences signal in this example, and modulating bandwidth 280MHz, center carries
Frequency is 24GHz, pulse repetition period 100us;Linear array composition includes 4 transmitting array elements and 3 reception array elements, wherein sending out
It is 1.5 λ to penetrate array element spacing, and reception array element spacing is 0.5 λ;Include 64 repetition periods in one coherent processing time.Target is
The people of two static trihedral angle reflectors and a movement, wherein the people's speed moved is 2.54m/s.
According to the explanation of step 1, we devise the random transmitting time division scheme (primary realization therein) such as Fig. 4, phase
Than with traditional time division scheme, array element shooting sequence no longer be in cyclically-varying, i.e., for the sky of linear coupling when phase reach solution
The purpose of coupling sum.It is essentially the same with target actual positions by the spectrum peak point after reconstruct, illustrate that this method effectively solves
It has determined due to coupled problem when target moves the sky generated in traditional TDM-MIMO radars in turn, and the unambiguous model that tests the speed
It encloses for ± 9.75m/s, is 4 times of traditional time division scheme.
In conclusion the foregoing is merely a prefered embodiment of the invention, it is not intended to limit the scope of the present invention.It is all
Within the spirit and principles in the present invention, any modification, equivalent replacement, improvement and so on should be included in the protection of the present invention
Within the scope of.