Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a target detection method based on GLRT under a plurality of heterogeneous satellites.
The invention is realized in such a way that a target detection method based on GLRT under a plurality of heterogeneous satellites comprises the following steps: modeling the received signals of the reference channel and the monitoring channel, separating the reference channel signals, and substituting the reference channel signals into the monitoring channel signals; establishing a binary hypothesis according to the received signals of the monitoring channel, and giving probability density functions of the signals under the two hypotheses; estimating the maximum likelihood estimation of the unknown parameters under two assumptions by using a maximum likelihood estimation method according to the two probability density functions; obtaining detection statistics based on the generalized likelihood ratio according to the maximum likelihood estimation of the unknown parameters to obtain a detector; and obtaining the optimal detection threshold of the detector according to the probability distribution of the detection statistics, and judging to detect the target.
Further, the method for detecting the target based on the GLRT under the plurality of heterogeneous satellites comprises the following steps:
separating a plurality of direct wave signals of a reference channel and using the signals as local reference signals;
secondly, establishing a binary hypothesis model and probability density functions of the signals under two hypotheses according to the local reference signals and the signals of the monitoring channel;
step three, estimating the unknown parameters α when the noise variance is known under two assumptions by using a maximum likelihood estimation method
η,c
η,kUnknown parameters α when the noise variance is unknown
η,c
η,k,σ
2And unknown parameters α when the noise variance is unknown and an interfering target is present
η,
c
η,k,σ
2;
Step four, constructing detection statistics based on the generalized likelihood ratio under three conditions;
and step five, setting detection thresholds under three conditions, and comparing and judging with the detection statistics under the three conditions to detect the target.
Further, the signal of the step one channel is expressed as:
where N is 0,1, … N-1, M represents the number of satellites, P is the number of multipath paths under a single satellite signal, N
η,Ω
η,α
ηTime delay, doppler shift, and amplitude of the echo signal, respectively, and when k is 1, c is
η,kIs the amplitude of the direct wave signal, and when k is 2,3, …, P,
c
η,kis the Doppler shift, time delay, amplitude of the multipath signal, K represents the number of the interference targets,
respectively representing the Doppler shift, delay, amplitude, n of the interfering echo signal
s(t) is white Gaussian noise.
Further, in the second step:
the binary hypothesis model is represented as:
H0x [ n ] of]The probability density function of (a) is:
H1x [ n ] of]The probability density function of (a) is:
further, the unknown parameter α when the noise variance is known in the third stepη,cη,kThe maximum likelihood estimate of (c) is as follows:
H
0the maximum likelihood estimate of the unknown parameters of:
wherein
It shows the assumption H
0Amplitude c of kth multipath of the lower η th satellite signal
η,kWhere k is 1, is an estimated value of the amplitude of the direct wave signal corresponding to the satellite;
Rcrepresenting the correlation of multipath signals, RcIs a matrix of P, [ Rc]skIs RcIs represented as:
r
xcrepresenting received signals and direct waves in the monitoring channel and multipath signals in
Correlation of (a) with [ r ]
xc]Is a matrix of P x M, representing [ r
xc]
qsIs represented as:
H1the maximum likelihood estimate of the unknown parameters of:
wherein
It shows the assumption H
1Amplitude c of kth multipath of the lower η th satellite signal
η,kWhere k is 1, is an estimate of the amplitude of the direct wave signal corresponding to the satellite,
indicating the Mth echo signal amplitude α
MIs estimated.
r
ssShowing the autocorrelation between echoes, [ r ]
ss]Is a matrix of M, [ r [ ]
ss]
rηIs represented by r
ssIs represented as:
rscrepresenting the correlation of multipath and echo, [ r ]sc]Is a matrix of M, [ r [ ]sc]qηIs represented by rscIs represented as:
rxseach target echo signal representing a received signal of the monitor channel is represented by (n)r,Ωr) Cross correlation of [ r ]xs]Is a vector of M1, [ rxs]rIs represented by rxsIs represented as:
unknown parameters α when noise variance is unknownη,cη,k,σ2The maximum likelihood estimate of (c) is as follows:
αη,cη,kwith unknown parameters α when the noise variance is knownη,cη,kThe maximum likelihood estimates of (a) are consistent;
suppose H
1Lower, variance of noise
Maximum likelihood estimation of
Comprises the following steps:
H
0variance of noise
Maximum likelihood estimation of
Comprises the following steps:
unknown parameters α when noise variance is unknown and an interfering target is present
η,
c
η,k,σ
2The maximum likelihood estimate of (c) is as follows:
maximum likelihood estimation of unknown parameters is similar in principle to the previous method, assuming H0The following unknown parameters are as follows:
suppose H1The following unknown parameters are as follows:
wherein the content of the first and second substances,
presentation assumption H
iUnknown parameter of
Is estimated by the estimation of (a) a,
it shows the assumption H
iThe estimated value of the amplitude of the next b interference echo;
r
tsrepresenting interfering target echo signals at
And the echo to be detected is at (n)
η,Ω
η) Correlation of [ r ]
ts]Is a vector of M x K, [ r
ts]
mrIs represented by r
tsIs represented as:
Rtcindicating that the interfering target echo is at (n)η,Ωη) And a multipath signal in (n)q,s,Ωq,s) Correlation between [ R ]tc]Is a vector of K P, [ Rtc]smIs RtcIs represented as:
r
xtrepresenting received signals and interfering target echo signals at
Correlation of [ r ]
xt]Is a vector of K1, [ r
xt]
fIs r
xtIs represented as:
Rttphase representing echoes of an interfering targetOff, [ R ]tt]Is a vector of K by K, [ R ]tt]fmIs RttThe elements of (a) are:
the maximum likelihood estimates of the noise variance under the two assumptions are:
further, in the fourth step, the detection statistics based on the generalized likelihood ratio under the three conditions are constructed as follows: detection statistic when noise variance is known:
detection statistic when noise variance is unknown:
wherein ξ', ηKWhich are the detection thresholds for the three cases, respectively.
Further, in the fifth step:
the decision threshold ψ of the detector when the noise variance is known is:
the detector decision threshold when the noise variance is unknown ψ is:
the decision threshold ψ of the detector when the noise variance is unknown and an interference target is present is:
wherein the content of the first and second substances,
is distributed in the central chi-square
Probability of right tail, P
faIs the false alarm probability.
The invention has the advantages and positive effects that: the parameter data is subjected to 2000 Monte Carlo experimental simulations to obtain the detection performance shown in FIG. 2, which proves that the invention can effectively realize the detection of the moving target under a plurality of heterogeneous satellite radiation sources.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The following detailed description of the principles of the invention is provided in connection with the accompanying drawings.
As shown in fig. 1, a method for detecting a target based on GLRT under multiple heterogeneous satellites according to an embodiment of the present invention includes the following steps:
s101: modeling the received signals of the reference channel and the monitoring channel, separating the reference channel signals, and substituting the reference channel signals into the monitoring channel signals;
s102: establishing a binary hypothesis according to the received signals of the monitoring channel, and giving probability density functions of the signals under the two hypotheses;
s103: estimating the maximum likelihood estimation of the unknown parameters under two assumptions by using a maximum likelihood estimation method according to the two probability density functions;
s104: obtaining detection statistics based on the generalized likelihood ratio according to the maximum likelihood estimation of the unknown parameters to obtain a detector;
s105: and obtaining the optimal detection threshold of the detector according to the probability distribution of the detection statistics, and judging to detect the target.
The target detection method based on GLRT under a plurality of heterogeneous satellites provided by the embodiment of the invention specifically comprises the following steps:
separating a plurality of direct wave signals of a reference channel and using the signals as local reference signals;
secondly, establishing a binary hypothesis model and probability density functions of the signals under two hypotheses according to the local reference signals and the signals of the monitoring channel;
step three, estimating the unknown parameters α when the noise variance is known under two assumptions by using a maximum likelihood estimation method
η,c
η,kUnknown parameters α when the noise variance is unknown
η,c
η,k,σ
2And unknown parameters α when the noise variance is unknown and an interfering target is present
η,
c
η,k,σ
2;
Step four, constructing detection statistics based on the generalized likelihood ratio under three conditions;
and step five, setting detection thresholds under three conditions, and comparing and judging with the detection statistics under the three conditions, thereby detecting the target.
In the first step, a plurality of direct wave signals of a reference channel are separated and used as local reference signals to be processed according to the following steps:
the expression for the reference channel signal z (t) is:
wherein lηIs the amplitude, n, of the direct wave signal of the reference channelr(t) is the noise of the reference channel, and M is the number of satellite radiation sources.
Because the reference channel receives direct wave signals of a plurality of different satellites, the direct wave signals are different in frequency, a band-pass filter can be designed to separate the direct wave signals, and after the direct wave signals are separated by the band-pass filter, the signals in the reference channel can be expressed as follows:
yη(t)=bηsη(t)+nη(t)0≤t<T η=1,2…M;
wherein, bηIs the amplitude, n, of the direct wave signal of the reference channelη(t) is the noise in the single direct wave signal after separation.
The signal of the monitoring channel is represented as:
where N is 0,1, … N-1, M represents the number of satellites, P is the number of multipath paths under a single satellite signal, N
η,Ω
η,α
ηTime delay, doppler shift, and amplitude of the echo signal, respectively, and when k is 1, c is
η,kIs the amplitude of the direct wave signal, and when k is 2,3, …, P,
c
η,kis the Doppler shift, time delay, amplitude of the multipath signal, K represents the number of the interference targets,
respectively representing the Doppler shift, delay, amplitude, n of the interfering echo signal
s(t) is white Gaussian noise.
In the second step, a binary hypothesis model is established according to the local reference signal and the signal of the monitoring channel, and the probability density function of the signal under two hypotheses is carried out as follows:
the binary hypothesis model is represented as:
suppose H0X [ n ] of]The probability density function of (a) is:
suppose H1X [ n ] of]The probability density function of (a) is:
in step three, the maximum likelihood estimation method is used to estimate the unknown parameters α under two assumptions when the noise variance is known
η,c
η,kUnknown parameters α when the noise variance is unknown
η,c
η,k,σ
2And unknown parameters α when the noise variance is unknown and an interfering target is present
η,
c
η,k,σ
2The method comprises the following steps:
number of unknown parameters α when the variance of the noise is knownη,cη,kThe maximum likelihood estimate of (c) is as follows:
H
0the maximum likelihood estimate of the unknown parameters of:
wherein
It shows the assumption H
0Amplitude c of kth multipath of the lower η th satellite signal
η,kWhere k is 1, is an estimate of the amplitude of the direct wave signal corresponding to the satellite.
RcRepresenting the correlation of multipath signals, RcIs a matrix of P, [ Rc]skIs RcIs represented as:
r
xcrepresenting received signals and direct waves in the monitoring channel and multipath signals in
Correlation of (a) with [ r ]
xc]Is a matrix of P x M, representing [ r
xc]
qsIs represented as:
H1the maximum likelihood estimate of the unknown parameters of:
wherein
It shows the assumption H
1Amplitude c of kth multipath of the lower η th satellite signal
η,kWhere k is 1, is an estimate of the amplitude of the direct wave signal corresponding to the satellite,
indicating the Mth echo signal amplitude α
MIs estimatedAnd (6) counting.
r
ssShowing the autocorrelation between echoes, [ r ]
ss]Is a matrix of M, [ r [ ]
ss]
rηIs represented by r
ssIs represented as:
rscrepresenting the correlation of multipath and echo, [ r ]sc]Is a matrix of M, [ r [ ]sc]qηIs represented by rscIs represented as:
rxseach target echo signal representing a received signal of the monitor channel is represented by (n)r,Ωr) Cross correlation of [ r ]xs]Is a vector of M1, [ rxs]rIs represented by rxsIs represented as:
unknown parameters α when noise variance is unknownη,cη,k,σ2The maximum likelihood estimate of (c) is as follows:
αη,cη,kwith unknown parameters α when the noise variance is knownη,cη,kThe maximum likelihood estimates of (a) are consistent.
Suppose H
1Lower, variance of noise
Maximum likelihood estimation of
Comprises the following steps:
H
0variance of noise
Maximum likelihood estimation of
Comprises the following steps:
unknown parameters α when noise variance is unknown and an interfering target is present
η,
c
η,k,σ
2The maximum likelihood estimate of (c) is as follows:
maximum likelihood estimation of unknown parameters is similar in principle to the previous method, assuming H0The following unknown parameters are as follows:
suppose H1The following unknown parameters are as follows:
wherein the content of the first and second substances,
presentation assumption H
iUnknown parameter of
Is estimated by the estimation of (a) a,
it shows the assumption H
iThe estimate of the amplitude of the lower b-th disturbing echo.
r
tsRepresenting interfering target echo signals at
And the echo to be detected is at (n)
η,Ω
η) Correlation of [ r ]
ts]Is a vector of M x K, [ r
ts]
mrIs represented by r
tsIs represented as:
Rtcindicating that the interfering target echo is at (n)η,Ωη) And a multipath signal in (n)q,s,Ωq,s) Correlation between [ R ]tc]Is a vector of K P, [ Rtc]smIs RtcIs represented as:
r
xtrepresenting received signals and interfering target echo signals at
Correlation of [ r ]
xt]Is a vector of K1, [ r
xt]
fIs r
xtIs represented as:
Rttrepresenting the correlation of interfering target echoes, [ R ]tt]Is a vector of K by K, [ R ]tt]fmIs RttThe elements of (a) are:
the maximum likelihood estimates of the noise variance under the two assumptions are:
in the fourth step, the detection statistics based on the generalized likelihood ratio is constructed under three conditions as follows:
detection statistic when noise variance is known:
detection statistic when noise variance is unknown:
wherein ξ', ηKWhich are the detection thresholds for the three cases, respectively.
In the fifth step, the detection thresholds under the three conditions are set and compared with the detection statistics under the three conditions for judgment, so that the target detection is carried out according to the following steps: the decision threshold ψ of the detector when the noise variance is known is:
the detector decision threshold when the noise variance is unknown ψ is:
the decision threshold ψ of the detector when the noise variance is unknown and an interference target is present is:
wherein the content of the first and second substances,
is distributed in the central chi-square
Probability of right tail, P
faIs the false alarm probability.
The application effect of the present invention will be described in detail with reference to the simulation.
Simulation experiment: and performing simulation verification on the echo detection performance based on the generalized likelihood ratio under the three conditions, and performing simulation experiments by using three satellite signals including GPS, DVB-S and inmarsat. The carrier frequencies of the three signals are respectively: f. ofG=1.57GHz,fD=12.38GHz,fiAssuming that the time delays of three echo signals are 1 μ s,2 μ s and 3 μ s respectively, the doppler shifts are 100Hz,150Hz and 200Hz respectively, and the intensities of three direct signal waves are: 130.1dBw-111.83 dBw-120.61 dBw, the difference between the power of the direct wave and the corresponding echo is 40dB, and the number of sampling points is 105And performing simulation by using matlab. The above parameter data was subjected to 2000 monte carlo experimental simulations to obtain the detection performance of fig. 2.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.