CN103344945B - Method for inhibiting multi-path interference of direct wave - Google Patents

Abstract

The invention relates to a direct wave and a method for inhibiting multi-path interference of the direct wave, and belongs to the technical field of radar target acquisition. According to the method, sub-bands adapt to RLS plus NLMS filtering by itself. The method comprises the steps that (1) the sub-bands are decomposed, (2) signals of all the sub-bands respectively pass through the RLS plus NLMS filtering, and (3) the sub-bands are integrated. The method improves the consistency of frequency characteristics between passages of the direct wave and an echo, improves the offset performance of a system, largely reduces the calculation amount of a self-adaptation offset algorithm under wideband signals, and has a high convergence rate and good tracking capacity when an antenna carries out scanning work.

Description

The suppressing method of a kind of direct wave and multi-path jamming thereof

Technical field

The present invention relates to the suppressing method of a kind of direct wave and multi-path jamming thereof, the method adopts sub-band adaptive RLS to add the method for NLMS filtering, belong to technical field of radar target acquisition, in particular to when utilizing digital television signal as irradiation source, the method for external radiation source radar system direct wave and MPI suppression thereof.

Background technology

External illuminators-based radar itself does not transmit, but the civilian broadcasting equipment (TV, broadcast etc.) utilizing extensively distribution is irradiation source, the electromagnetic signal reflected by receiving target is gone forward side by side line correlation process, and realize the detection to target and location, its system configuration as shown in Figure 1.Echo antenna is except receiving the signal of target reflection, direct wave and multi-path jamming thereof also will enter from its secondary lobe and main lobe, the power of these undesired signals, generally higher than target echo signal more than power 30dB, if do not suppressed these interference, just cannot detect target.Therefore the key issue realizing external radiation source radar system detection performance is exactly suppress direct wave in echo channel and multi-path jamming thereof.

Adaptive cancellation is a kind of effective ways of AF panel, and in external illuminators-based radar, conventional adaptive cancellation algorithm has normalization minimum mean-square (NLMS) wave filter, recursive least-squares (RLS) wave filter etc.Its characteristic comparing is as shown in table 1, and in table, N is filter order.

Table 1 NLMS and RLS wave filter characteristic comparing

During antenna scanning, direct wave and multi-path jamming fluctuations thereof are very large, and do not meet smooth conditions, now NLMS performance of filter will decline; Simultaneously due to the broader bandwidth (7.56MHz) of digital television signal, the filter order very high (filter order is about 2000) that same clutter distance needs, therefore adopt RLS algorithm operation quantity huge (to square being directly proportional of filter order, as shown in table 1), be difficult to real-time implementation.

In addition, the cancellation performance of external radiation source radar system is directly proportional to the frequency invariance of direct wave, echo two passages, and the better then cancellation performance of consistance is higher.Outside in the narrower situation of emitter Signals bandwidth, as fm broadcast signal bandwidth is less than 200KHz, the frequency characteristic consistance of echo reception antenna and direct wave receiving antenna is better; And outside in emitter Signals broader bandwidth situation, if digital television signal bandwidth is 7.56MHz, the frequency characteristic consistance of echo reception antenna and direct wave receiving antenna is deteriorated, as shown in Fig. 2 (a), (b), have a strong impact on cancellation performance, needed to take corresponding signal transacting measure.

Summary of the invention

The object of the invention is the suppressing method in order to propose a kind of direct wave and multi-path jamming thereof, be applicable to Digital Television external illuminators-based radar, the object of the method is that direct wave passage, echo channel frequency characteristic consistance are deteriorated, and affect the problem of cancellation performance in order under solving broadband signal; And under broadband signal, high-order sef-adapting filter operand is large, the problem of real-time difficult treatment.

The object of the invention is to be achieved through the following technical solutions.

The suppressing method of a kind of direct wave of the present invention and multi-path jamming thereof, the filtering method that the method adopts RLS algorithm to combine with NLMS algorithm in each subband, concrete steps are:

1) sub-band division

Direct wave, echoed signal are decomposed into L subband respectively by analysis filterbank, and relevant mathematical description is as follows:

(1) formula is mathematic(al) representation direct-path signal being decomposed into L subband; (2) formula is mathematic(al) representation target echo signal being decomposed into L subband; (3) formula is the mathematical description of sub-band division bank of filters:

x _m(n)＝x(n)*h _m(n) (1)

d _m(n)＝d(n)*h _m(n) (2)

h _m(n)＝h(n)e ^-j2πmn/M(3)

M=0 in formula, 1 ..., L-1; The prototype filter that h (n) adopts for sub-band division wave filter;

2) signal of each subband respectively by RLS in conjunction with NLMS filtering

In antenna scanning situation, adopt RLS and NLMS combination method, if the signal of m subband carry out M doubly extract after direct-path signal vector be

X _m(n)＝[x _m(Mn) x _m(M(n-1))…x _m(M(n-M+1))] ^T

Filter coefficient update formula is as follows:

$w(n-1)={\left[\begin{array}{cc}{w}_{\mathrm{RLS}}^T(n-1)& w_{\mathrm{NLMS}}^T(n-1)\end{array}\right]}^T---\left(4\right)$

In (4) formula,

$w_{\mathrm{RLS}}\left(n\right)=w_{\mathrm{RLS}}(n-1)+k\left(n\right)e_m^{*}\left(\mathrm{Mn}\right)---\left(5\right)$

$w_{\mathrm{NLMS}}\left(n\right)=w_{\mathrm{NLMS}}(n-1)+\mathrm{\μ}\frac{e_m^{*}\left(\mathrm{Mn}\right)X_m\left(n\right)}{\mathrm{\δ}+{X_m}^H\left(n\right)X_m\left(n\right)}---\left(6\right)$

In (5), (6) formula,

e _m(Mn)＝d _m(Mn)-w ^H(n-1)X _m(Mn) (7)

$k\left(n\right)=\frac{{\mathrm{\λ}}^{-1}P(n-1)X_m\left(n\right)}{1+{\mathrm{\λ}}^{-1}{X_m}^H\left(n\right)P(n-1)X_m\left(n\right)}---\left(8\right)$

P(n)＝λ ^-1P(n-1)-λ ^-1k(n)X _m ^H(n)P(n-1) (9)

3) subband is comprehensive

The signal cancellation result of each subband is carried out interpolation, then synthesizes final output signal by sub-band synthesis filter, its mathematical description is as shown in (10) formula; (11) formula is the mathematic(al) representation of sub-band synthesis filter:

$y\left(n\right)=\underset{m=0}{\overset{M-1}{\mathrm{\Σ}}}{f}_m\left(n\right)\⊗e_m\left(n\right)---\left(10\right)$

f _m(n)＝f(n)e ^j2πmn/M，m＝0,1,…,L-1 (11)

F in formula _mn () is m synthesis filter coefficients, f=[f (0) ..., f (N-1)] ^tfor the prototype filter of adopted sub-band synthesis filter;

X (n): direct-path signal;

D (n): echoed signal;

H _m(n): m Subband Analysis Filter;

X _m(n): the signal of m the subband that direct-path signal obtains through sub-band division;

X _m(Mn): x _mthe signal that M doubly obtains is taken out under (n);

D _m(n): the signal of m the subband that echoed signal obtains through sub-band division;

D _m(Mn): d _mthe signal that M doubly obtains is taken out under (n);

W (n): RLS+NLMS filter coefficient;

W _rLS(n): RLS filter coefficient;

W _nLMS(n): NLMS filter coefficient;

E _m(Mn): m subband offset result;

E _m(n): e _m(Mn) interpolation M doubly after the result that obtains;

The forgetting factor of λ: RLS wave filter;

Time-varying gain vector in k (n): RLS algorithm;

The intermediate variable of k (n) is calculated in P (n): RLS algorithm

δ: little normal number, prevent division from overflowing, δ gets 0.001 here;

The renewal step-length of μ: NLMS algorithm;

F _m(n): m sub-band synthesis filter;

Y (n): offset rear output signal;

M: lower extraction and interpolation multiple.

Beneficial effect

1) improve the consistance of frequency characteristic between direct wave and echo channel, improve the cancellation performance of system;

2) greatly reduce the operand of adaptive cancellation algorithm under broadband signal, when antenna scanning works, there is speed of convergence and good tracking power faster.

Accompanying drawing explanation

Fig. 1 is external radiation source radar system configuration schematic diagram;

Fig. 2 is that wideband scenarios lower channel consistance worsens and after sub-band division, the consistent type of passage improves schematic diagram;

Fig. 3 is the schematic diagram of suppressing method of the present invention;

Fig. 4 is RLS+NLMS filtering algorithm structured flowchart;

Fig. 5 is RLS, subband RLS+NLMS algorithm cancellation performance compares;

Fig. 6 is RLS, subband RLS+NLMS algorithm operation quantity compares.

Embodiment

Below in conjunction with accompanying drawing and experiment analysis of data collected the present invention will be further described.

A suppressing method for direct wave and multi-path jamming thereof, the filtering method that the method adopts RLS algorithm to combine with NLMS algorithm in each subband, as shown in Figure 3, concrete steps are:

1) sub-band division

Direct wave, echoed signal are decomposed into L subband respectively by analysis filterbank, and (L can be 8,16,32,64,, concrete condition is chosen as required), due to each subband narrow bandwidth, therefore the frequency characteristic consistance between direct wave passage and echo channel improves, as shown in Fig. 2 (c), (d); Carry out M doubly lower extraction to each subband, the exponent number that offsets like this needed for each subband is just reduced to original 1/M, reduces the operand of adaptive cancellation simultaneously;

Relevant mathematical description is as follows, and (1) formula is mathematic(al) representation direct-path signal being decomposed into L subband; (2) formula is mathematic(al) representation target echo signal being decomposed into L subband; (3) formula is the mathematical description of sub-band division bank of filters

x _m(n)＝x(n)*h _m(n) (1)

d _m(n)＝d(n)*h _m(n) (2)

h _m(n)＝h(n)e ^-j2πmn/M(3)

M=0 in formula, 1 ..., L-1; The prototype filter that h (n) adopts for sub-band division wave filter;

2) signal of each subband respectively by RLS in conjunction with NLMS filtering

In antenna scanning situation, offset and adopt RLS algorithm performance best, although take out under sub-band division, but the operand of RLS algorithm is still very large, at this moment can adopt method that RLS and NLMS combines as shown in Figure 4, due to direct wave and noise jamming nearby stronger, so adopt RLS wave filter to offset effect to reach reasonable, and noise jamming is at a distance relatively weak, the NLMS wave filter that operand can be adopted smaller or variable step NLMS wave filter.Both combined, obtain a kind of new adaptive cancellation wave filter, front portion coefficient adopts the method for RLS to upgrade, and rear portion coefficient adopts the method for NLMS or variable step NLMS to upgrade.While meeting very fast speed of convergence and tracking power reduce operand, if the signal of m subband carry out M doubly extract after direct-path signal vector be

X _m(n)＝[x _m(Mn) x _m(M(n-1))…x _m(M(n-M+1))] ^T

Filter coefficient update formula is as follows:

$w(n-1)={\left[\begin{array}{cc}{w}_{\mathrm{RLS}}^T(n-1)& w_{\mathrm{NLMS}}^T(n-1)\end{array}\right]}^T---\left(4\right)$

In (4) formula,

$w_{\mathrm{RLS}}\left(n\right)=w_{\mathrm{RLS}}(n-1)+k\left(n\right)e_m^{*}\left(\mathrm{Mn}\right)---\left(5\right)$

$w_{\mathrm{NLMS}}\left(n\right)=w_{\mathrm{NLMS}}(n-1)+\mathrm{\μ}\frac{e_m^{*}\left(\mathrm{Mn}\right)X_m\left(n\right)}{\mathrm{\δ}+{X_m}^H\left(n\right)X_m\left(n\right)}---\left(6\right)$

In (5), (6) formula,

e _m(Mn)＝d _m(Mn)-w ^H(n-1)X _m(Mn) (7)

$k\left(n\right)=\frac{{\mathrm{\λ}}^{-1}P(n-1)X_m\left(n\right)}{1+{\mathrm{\λ}}^{-1}{X_m}^H\left(n\right)P(n-1)X_m\left(n\right)}---\left(8\right)$

P(n)＝λ ^-1P(n-1)-λ ^-1k(n)X _m ^H(n)P(n-1) (9)

3) subband is comprehensive

The signal cancellation result of each subband is carried out interpolation, then synthesizes final output signal by sub-band synthesis filter, its mathematical description is as shown in (10) formula.(11) formula is the mathematic(al) representation of sub-band synthesis filter.

$y\left(n\right)=\underset{m=0}{\overset{M-1}{\mathrm{\Σ}}}{f}_m\left(n\right)\⊗e_m\left(n\right)---\left(10\right)$

f _m(n)＝f(n)e ^j2πmn/M，m＝0,1,…,L-1 (11)

F in formula _mn () is m synthesis filter coefficients, f=[f (0) ..., f (N-1)] ^tfor the prototype filter of adopted sub-band synthesis filter;

X (n): direct-path signal;

D (n): echoed signal;

H _m(n): m Subband Analysis Filter;

X _m(n): the signal of m the subband that direct-path signal obtains through sub-band division;

X _m(Mn): x _mthe signal that M doubly obtains is taken out under (n);

D _m(n): the signal of m the subband that echoed signal obtains through sub-band division;

D _m(Mn): d _mthe signal that M doubly obtains is taken out under (n);

W (n): RLS+NLMS filter coefficient;

W _rLS(n): RLS filter coefficient;

W _nLMS(n): NLMS filter coefficient;

E _m(Mn): m subband offset result;

E _m(n): e _m(Mn) interpolation M doubly after the result that obtains;

The forgetting factor of λ: RLS wave filter;

Time-varying gain vector in k (n): RLS algorithm;

The intermediate variable of k (n) is calculated in P (n): RLS algorithm

δ: little normal number, prevent division from overflowing, δ gets 0.001 here;

The renewal step-length of μ: NLMS algorithm;

F _m(n): m sub-band synthesis filter;

Y (n): offset rear output signal;

M: lower extraction and interpolation multiple.

Embodiment

Adopt digital television signal, its radiation source radar system configuration schematic diagram as shown in Figure 1, cell site is no-cooperative radiate (external sort algorithm), there are two receiving antennas in receiving station, a direct wave antenna direction radiation source receives its signal launched, the signal of another echo reception antenna direction searching sector receiving target reflection.Echo reception antenna is while reception moving target reflection echo, its secondary lobe also receives the strong direct-wave jamming directly launched from external sort algorithm, also has being transmitted the multi-path jamming caused through scatterings such as multiple high-lager building, mountain ranges by external sort algorithm of entering from main lobe in addition.

Adopt digital television signal, bandwidth 7.56MHz, baseband data samples rate 9MHz:

1) sub-band division

The sub band number that sub-band division adopts is L=128, sub-filter exponent number is 1024, then after decomposing, the bandwidth of each subband signal is 1/128 of original signal, then carries out lower extraction to each subband signal, lower extracting multiple is M=128, and data transfer rate becomes 1/128 of raw data.

2) RLS+NLMS filtering is done to each subband

Offset process respectively to direct wave, each corresponding subband of echoed signal, after lower extraction, each subband RLS filter order is 2 rank, and NLMS filter order 16 rank, substantially reduce operand.

3) subband is comprehensive

Each subband outputs signal first interpolation 128 times reduction original data rate, then obtains finally exporting offseting result by sub-band synthesis filter.

As shown in Figure 5, adopt RLS algorithm, it is 35.0818dB that system offsets gain (offset front input signal energy/offset rear output signal energy), and adopting subband RLS+NLMS filtering algorithm to offset gain is 39.0306dB, improves about 4dB than RLS algorithm; And subband RLS+NLMS algorithm reduces about 9 times than RLS algorithm operation quantity when filter order is 1024 as can be seen from Figure 6, when filter order is 2048, reduce about 37 times.Therefore adopt subband RLS+NLMS algorithm to not only increase and offset gain, and greatly can reduce operand.

Method of the present invention improves the consistance of frequency characteristic between direct wave and echo channel, improves the cancellation performance of system, as shown in Figure 5; And the operand of adaptive cancellation algorithm under greatly reducing broadband signal, as shown in Figure 6, when antenna scanning works, there is speed of convergence and good tracking power faster.

Above-described specific descriptions; the foregoing is only specific embodiments of the invention, the protection domain be not intended to limit the present invention, within the spirit and principles in the present invention all; any amendment of making, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (1)

1. a suppressing method for direct wave and multi-path jamming thereof, is characterized in that: the filtering method that the method adopts RLS algorithm to combine with NLMS algorithm in each subband, and concrete steps are:

1) sub-band division

Direct wave, echoed signal are decomposed into L subband respectively by analysis filterbank, and relevant mathematical description is as follows:

(1) formula is mathematic(al) representation direct-path signal being decomposed into L subband; (2) formula is mathematic(al) representation target echo signal being decomposed into L subband; (3) formula is the mathematical description of sub-band division bank of filters:

x _m(n)＝x(n)*h _m(n) (1)

d _m(n)＝d(n)*h _m(n) (2)

h _m(n)＝h(n)e ^-j2πmn/M(3)

M=0 in formula, 1 ..., L-1; The prototype filter that h (n) adopts for sub-band division wave filter;

2) signal of each subband respectively by RLS in conjunction with NLMS filtering

In antenna scanning situation, adopt RLS and NLMS combination method, if the signal of m subband carry out M doubly extract after direct-path signal vector be

X _m(n)＝[x _m(Mn) x _m(M(n-1)) … x _m(M(n-M+1))] ^T

Filter coefficient update formula is as follows:

$w(n-1)={\left[\begin{array}{cc}{w}_{\mathrm{RLS}}^T(n-1)& w_{\mathrm{NLMS}}^T(n-1)\end{array}\right]}^T$

In (4) formula,

$w_{\mathrm{RLS}}\left(n\right)=w_{\mathrm{RLS}}(n-1)+k\left(n\right)e_m^{*}\left(\mathrm{Mn}\right)---\left(5\right)$

$w_{\mathrm{NLMS}}\left(n\right)=w_{\mathrm{NLMS}}(n-1)+\mathrm{\μ}\frac{e_m^{*}\left(\mathrm{Mn}\right)X_m\left(n\right)}{\mathrm{\δ}+{X_m}^H\left(n\right)X_m\left(n\right)}---\left(6\right)$

In (5), (6) formula,

e _m(Mn)＝d _m(Mn)-w ^H(n-1)X _m(Mn) (7)

$k\left(n\right)=\frac{{\mathrm{\λ}}^{-1}P(n-1)X_m\left(n\right)}{1+{\mathrm{\λ}}^{-1}{X_m}^H\left(n\right)P(n-1)X_m\left(n\right)}---\left(8\right)$

P(n)＝λ ^-1P(n-1)-λ ^-1k(n)X _m ^H(n)P(n-1) (9)

3) subband is comprehensive

The signal cancellation result of each subband is carried out interpolation, then synthesizes final output signal by sub-band synthesis filter, its mathematical description is as shown in (10) formula; (11) formula is the mathematic(al) representation of sub-band synthesis filter:

$y\left(n\right)=\underset{m=0}{\overset{M-1}{\mathrm{\Σ}}}{f}_m\left(n\right)\⊗e_m\left(n\right)---\left(10\right)$

f _m(n)＝f(n)e ^j2πmn/M，m＝0,1,…,L-1 (11)

F in formula _mn () is m synthesis filter coefficients, f=[f (0) ..., f (N-1)] ^tfor the prototype filter of adopted sub-band synthesis filter;

X (n): direct-path signal;

D (n): echoed signal;

H _m(n): m Subband Analysis Filter;

X _m(n): the signal of m the subband that direct-path signal obtains through sub-band division;

X _m(Mn): x _mthe signal that M doubly obtains is taken out under (n);

D _m(n): the signal of m the subband that echoed signal obtains through sub-band division;

D _m(Mn): d _mthe signal that M doubly obtains is taken out under (n);

W (n): RLS+NLMS filter coefficient;

W _rLS(n): RLS filter coefficient;

W _nLMS(n): NLMS filter coefficient;

E _m(Mn): m subband offset result;

E _m(n): e _m(Mn) interpolation M doubly after the result that obtains;

The forgetting factor of λ: RLS wave filter;

Time-varying gain vector in k (n): RLS algorithm;

The intermediate variable of k (n) is calculated in P (n): RLS algorithm

δ: little normal number, prevent division from overflowing, δ gets 0.001 here;

The renewal step-length of μ: NLMS algorithm;

F _m(n): m sub-band synthesis filter;

Y (n): offset rear output signal;

M: lower extraction and interpolation multiple.