CN109270344B - Coherent pulse signal frequency estimation method under pulse loss - Google Patents

Coherent pulse signal frequency estimation method under pulse loss Download PDF

Info

Publication number
CN109270344B
CN109270344B CN201811164794.XA CN201811164794A CN109270344B CN 109270344 B CN109270344 B CN 109270344B CN 201811164794 A CN201811164794 A CN 201811164794A CN 109270344 B CN109270344 B CN 109270344B
Authority
CN
China
Prior art keywords
pulse
frequency estimation
segment
frequency
small
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811164794.XA
Other languages
Chinese (zh)
Other versions
CN109270344A (en
Inventor
姚志均
高锐
杨睛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yangzhou University
Original Assignee
Yangzhou University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yangzhou University filed Critical Yangzhou University
Priority to CN201811164794.XA priority Critical patent/CN109270344B/en
Publication of CN109270344A publication Critical patent/CN109270344A/en
Application granted granted Critical
Publication of CN109270344B publication Critical patent/CN109270344B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
    • G01R23/02—Arrangements for measuring frequency, e.g. pulse repetition rate; Arrangements for measuring period of current or voltage
    • G01R23/12—Arrangements for measuring frequency, e.g. pulse repetition rate; Arrangements for measuring period of current or voltage by converting frequency into phase shift
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0205—Details
    • G01S5/021—Calibration, monitoring or correction

Abstract

The invention provides a coherent pulse signal frequency estimation method under pulse loss, which comprises the steps of firstly carrying out phase ambiguity resolution and single-pulse-level frequency estimation in a pulse range; secondly, performing phase ambiguity resolution and frequency estimation twice by taking two pulses as a group, and obtaining pulse loss position information; thirdly, dividing the coherent pulse string into a plurality of large segments according to the pulse loss position information, subdividing the large segments by taking the effective minimum pulse number in all the large segments as the segment length, carrying out frequency estimation of a small segment level, eliminating small segments with too large errors, and synthesizing the rest small segments to obtain a frequency estimation value; and finally, searching the largest continuous effective small sections, respectively forming observation intervals by the continuous effective small sections in the large sections, carrying out frequency estimation of observation interval levels, and taking an average value as a final frequency estimation value. The invention can still accurately estimate the frequency of the coherent pulse train under the condition of pulse loss.

Description

Coherent pulse signal frequency estimation method under pulse loss
Technical Field
The invention belongs to the technical field of electronic countermeasure, and particularly relates to a coherent pulse signal frequency estimation method under pulse loss.
Background
Passive positioning is an important means in electronic reconnaissance, and passively receives electromagnetic waves of a radiation source under the condition that the radiation source does not radiate any signal, measures various parameters of the electromagnetic waves, and covertly determines the position and the motion state of the radiation source. The positioning is to estimate the position of the radiation source by using a set of measured parameter values, and the tracking is to estimate not only the position of the radiation source but also parameters such as the speed, the acceleration and the like of the radiation source. The single-station passive positioning and tracking technology is a technology for positioning and tracking a target radiation source by using one observation platform, has the advantages of no need of synchronous work and data transmission among a plurality of observation platforms, simple structure, easy engineering realization and the like, is increasingly attracted by researchers in various countries, and becomes a hotspot of military technology research in various countries.
The earliest single-station passive positioning method is a direction-finding positioning method based on the wave arrival angle of a radiation source, and a Doppler frequency positioning method, an azimuth/Doppler frequency positioning method and the like are developed later due to low positioning accuracy. The positioning accuracy of the doppler frequency based positioning method depends on the frequency or frequency change rate measurement accuracy. In other words, the stability of the positioning and tracking algorithm can be improved by improving the estimation accuracy of the frequency change rate, so that the research on the high-accuracy frequency change rate estimation algorithm under the condition specific to the single-station passive received signal has great significance.
At present, many radar signals are coherent signals, so that the doppler frequency change rate can be estimated by using the coherent characteristics of the signals, and many researchers have studied the problem of high-precision frequency estimation of coherent radar pulse trains. Some have proposed a frequency estimation algorithm based on chi-squared detection for coherent bursts, which differs significantly from the cara-roche limit in performance. Some people perform non-coherent estimation on each single pulse, and then perform search in a certain range to remove frequency ambiguity, wherein the estimation precision and the convergence speed of the method depend on the step length of the search. A multi-stage frequency estimation algorithm suitable for coherent pulse trains is provided, and the estimation performance under the condition of high signal-to-noise ratio is close to the Clarmet-Roche limit. Although the estimation accuracy of the multistage frequency estimation algorithm is indeed high, it does not take into account the loss of pulses, i.e. the algorithm requires that the signal received by the receiver is continuous. It is obvious that in practical applications, there are many reasons that the pulse may be lost, thereby affecting the application of the multi-stage frequency estimation algorithm.
Disclosure of Invention
The invention aims to provide a coherent pulse signal frequency estimation method under the condition of pulse loss, and solves the problem that the existing radar signal frequency estimation accuracy is not high.
The technical solution for realizing the invention is as follows: the method for estimating the frequency of the coherent pulse signal under the condition of pulse loss comprises the following specific steps:
step 1, extracting fuzzy phases of coherent pulses, performing phase ambiguity resolution in a pulse range, performing frequency estimation on each pulse respectively, and averaging to obtain a primary frequency estimation value;
step 2, segmenting the coherent pulse train by taking 2 pulses as a group, then respectively carrying out phase deblurring processing and frequency estimation on each segment, finding out the position LostPosition1 of the lost pulse through error analysis, and recording the residual effective frequency estimation value
Figure BDA0001820852910000021
And 3, deleting the first pulse, segmenting the residual coherent pulse train by taking 2 pulses as a group, then respectively carrying out phase ambiguity resolution and frequency estimation on each segment, finding out the position LostPosition2 of the lost pulse through error analysis, and recording the residual effective frequency estimation value
Figure BDA0001820852910000022
Step 4, dividing the reference pulse train into a plurality of large segments by utilizing the position information LostPosition1 and LostPosition2 of the lost pulse, acquiring the effective minimum pulse number in all the large segments, further segmenting each large segment by using the effective minimum pulse number, then respectively carrying out phase deblurring and frequency estimation on each small segment, eliminating invalid small segments through error analysis, and synthesizing the rest small segments to obtain a frequency estimation value
Figure BDA0001820852910000023
And 5, searching the largest continuous effective small sections, forming an observation interval by the continuous effective small sections in the large sections, performing phase ambiguity resolution and frequency estimation on each observation interval, and averaging to obtain a final frequency estimation value.
Preferably, the specific method for performing phase deblurring processing in a pulse range in step 1, then performing frequency estimation on each pulse respectively, and obtaining a primary frequency estimation value after averaging comprises:
11) calculating the difference value of adjacent phases in the (p + 1) th pulse, wherein the specific formula is as follows:
Δφp,o(n)＝φo(n+Kp)-φo(n-1+Kp),n＝1,…,Ns-1；
φo(N + Kp) is the fuzzy phase, NsThe number of sampling points in a pulse, K is the number of sampling points in a repetition period, and P is the number of pulses in the whole observation time;
12) starting from n equal to 1, if Δ φp,o(n) is greater than or equal to 0, then phip(n)＝φo(n + Kp), n ═ n +1, repeat step 11), φp(n) is the phase of the (n + 1) th sampling signal in the (p + 1) th pulse after the deblurring; otherwise, phio(k+Kp)＝φo(k+Kp)+2π,k＝n,n+1,…,Ns-1，φp(n)＝φo(n + Kp), repeating step 11);
13) respectively carrying out frequency estimation on each pulse, and averaging to obtain a primary frequency estimation value, wherein the formula for respectively carrying out frequency estimation on each pulse is as follows:
Figure BDA0001820852910000031
the primary frequency estimate obtained is:
Figure BDA0001820852910000032
preferably, the specific method for segmenting the coherent pulse train by using 2 pulses as a group, then performing phase deblurring processing and frequency estimation on each segment, and then finding out the position LostPosition1 of the missing pulse through error analysis comprises the following steps:
21) segmenting coherent pulse train by taking 2 pulses as a group, and setting the total M1And (3) if the number of pulses in the whole observation time is an odd number, the last pulse is not considered, the phase of two pulse signals in each segment is deblurred by using the frequency value estimated in the step (1), and the phase deblurring formula of the 2 nd pulse in the m-th segment signal is as follows:
Figure BDA0001820852910000033
where round (x) is the nearest integer to x, φm,1(n) and phim,2(n) phases of the (n + 1) th sampling signals of the 1 st pulse and the 2 nd pulse in the m-th section, respectively,
Figure BDA0001820852910000034
for a first frequency estimate, Δ ═ 1/fs，fsFor the sampling rate, k is taken for the first pulse signal in the m-th segmentm,1When the phase of each pulse signal in the mth data segment is 0, the phase deblurring expression is as follows:
φm(n+K(j-1))＝φm,j(n)+2km,jπ,j＝1,2
22) and carrying out frequency estimation on each section of data to obtain a frequency estimation value, wherein the expression is as follows:
Figure BDA0001820852910000035
23) calculating M1The average value and the standard deviation of the frequency estimation values of the data of the segment, when the difference between the frequency estimation value and the average value of the data of the segment exceeds the set multiple of the standard deviation, the pulse loss exists in the data of the segment, the serial number of the data of the segment is recorded into the position LostPosition1 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure BDA0001820852910000041
Preferably, the specific method of deleting the first pulse in step 3, segmenting the remaining coherent pulse train by using 2 pulses as a group, then performing phase deblurring processing and frequency estimation on each segment, and finding out the position LostPosition2 of the missing pulse through error analysis includes:
31) deleting the first pulse, and then segmenting the rest coherent pulse train in the step 2 by taking 2 pulses as a group, wherein the total M is set2Segment, if P is an even number, the last pulse is disregarded,repeating the steps 21) and 22) to obtain the frequency estimation value in each section of data
Figure BDA0001820852910000042
32) Calculate this M2The average value and the standard variance of the frequency estimation values of the data segments are determined, when the difference between the frequency estimation value and the average value of a certain data segment exceeds a certain multiple of the standard variance, the data segment is considered to have pulse loss, the serial number of the data segment is recorded into the position LostPosition2 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure BDA0001820852910000043
Preferably, in step 4, the position information of the missing pulse, LostPosition1 and LostPosition2, is used to divide the reference pulse train into a plurality of large segments, obtain the effective minimum number of pulses in all the large segments, further segment each large segment by using the effective minimum number of pulses, then perform phase deblurring and frequency estimation on each small segment, remove the ineffective small segments, and synthesize the remaining small segments to obtain a frequency estimation value, and the specific method is as follows:
41) the effective frequency estimates obtained in steps 2 and 3 are
Figure BDA0001820852910000044
And
Figure BDA0001820852910000045
combined together and their average calculated and recorded as
Figure BDA0001820852910000046
42) Combining the pulse loss position information LostPosition1 and LostPosition2 obtained in steps 2 and 3 to obtain more complete pulse loss position information LostNo, the expression of which is:
LostNo＝sort[LostPosition1*2-1,LostPosition2]
43) segmenting the whole coherent pulse train by using the position information in the LostNo, calculating the number of pulses in each segment, and then taking the minimum value with the number of the pulses larger than 2 and recording the minimum value as Q;
44) for the section with the number of pulses larger than Q, taking Q pulses as a group, subdividing to obtain smaller sections, and performing phase ambiguity resolution and frequency estimation on each small section, wherein for the Q pulse of the mth small section in a certain large section, the phase ambiguity resolution formula is as follows:
Figure BDA0001820852910000051
for the first pulse signal of the small segment, k is takenm,1If it is 0, the phase deblurring expression of each pulse signal in the small segment data is:
φm(n+K(q-1))＝φm,q(n)+2km,qπ,q＝1,…,Q
the expression of the frequency estimate for the small segment data is:
Figure BDA0001820852910000052
meanwhile, recording the information that each small segment belongs to a certain large segment, and recording the information as Flag;
45) calculating the frequency estimation value estimated from all small segments
Figure BDA0001820852910000053
When the difference between the frequency estimation value of a certain small segment data and the mean value exceeds the set multiple of the standard variance, the frequency estimated by the small segment data is considered to be unreliable, and the value corresponding to the small segment in the Flag is set to be invalid;
46) calculating the average value of the frequency estimated values obtained by estimating all effective small segments and recording the average value as
Figure BDA0001820852910000054
Preferably, step 5 searches for the largest continuous effective segments, and forms the effective segments in the largest continuous effective segments into an observation interval, and then performs phase ambiguity resolution and frequency estimation on each observation interval, and the average to obtain the final frequency estimation value specifically includes:
51) counting the number of the continuous effective small segments in each large segment, finding out the large segment with the maximum number of the continuous effective small segments, and if the number T of the large segments meets the condition, respectively combining the continuous effective small segments in the large segments into an observation interval, namely T observation intervals;
52) performing pulse phase deblurring processing in an observation interval, wherein if Q' pulses are shared in the observation interval, the phase deblurring formula of the Q pulse in the t observation interval is as follows:
Figure BDA0001820852910000055
for the first pulse signal of the observation interval, k is takent,1If the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is 0, the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is as follows:
φt(n+K(q-1))＝φt,q(n)+2kt,qπ,q＝1,…,Q'
and carrying out frequency estimation on each observation interval, wherein the frequency estimation expression of a single observation interval is as follows:
Figure BDA0001820852910000061
and synthesizing the frequency estimation values of the T observation intervals to obtain a final frequency estimation value, wherein the expression is as follows:
Figure BDA0001820852910000062
compared with the prior art, the invention has the following remarkable advantages: the invention can still accurately estimate the frequency of the coherent pulse train under the condition of pulse loss.
The present invention is described in further detail below with reference to the attached drawings.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a schematic phase diagram of sampling points in a pulse without intra-pulse phase deblurring.
Fig. 3 is a schematic phase diagram of each sampling point in a pulse after the intra-pulse phase deblurring processing.
Fig. 4 is a graph illustrating frequency values estimated from each pulse.
FIG. 5 is a schematic phase diagram of sampling points in a segment without phase deblurring processing in two pulse ranges.
Fig. 6 is a schematic phase diagram of each sampling point after phase deblurring processing in two pulse ranges is performed on a certain segment.
FIG. 7 is a schematic diagram of frequency values estimated by each segment in step 2 without deletion.
Fig. 8 is a schematic diagram of frequency values obtained by estimating each segment in step 2 and after deleting an invalid value.
FIG. 9 is a schematic diagram of frequency values estimated by each segment in step 3 without deletion.
Fig. 10 is a schematic diagram of frequency values obtained by estimating each segment in step 3 and after deleting an invalid value.
FIG. 11 is a schematic diagram of frequency values estimated by each small segment in step 4.
Detailed Description
The method for estimating the frequency of the coherent pulse signal under the condition of pulse loss comprises the following specific steps:
step 1, extracting fuzzy phases of coherent pulses, performing phase ambiguity resolution in a pulse range, performing frequency estimation on each pulse respectively, and averaging to obtain a primary frequency estimation value, wherein the specific method comprises the following steps:
11) calculating the difference value of adjacent phases in the (p + 1) th pulse, wherein the specific formula is as follows:
Δφp,o(n)＝φo(n+Kp)-φo(n-1+Kp),n＝1,…,Ns-1；
φo(N + Kp) is the fuzzy phase, NsThe number of sampling points in a pulse, K is the number of sampling points in a repetition period, and P is the number of pulses in the whole observation time;
12) starting from n equal to 1, if Δ φp,o(n) is greater than or equal to 0, then phip(n)＝φo(n + Kp), n ═ n +1, repeat step 11), φp(n) is the phase of the (n + 1) th sampling signal in the (p + 1) th pulse after the deblurring; otherwise, phio(k+Kp)＝φo(k+Kp)+2π,k＝n,n+1,…,Ns-1，φp(n)＝φo(n + Kp), repeating step 11);
13) respectively carrying out frequency estimation on each pulse, and averaging to obtain a primary frequency estimation value, wherein the formula for respectively carrying out frequency estimation on each pulse is as follows:
Figure BDA0001820852910000071
wherein, Delta is 1/fs，fsIs the sampling rate.
The primary frequency estimate obtained is:
Figure BDA0001820852910000072
step 2, segmenting the coherent pulse train by taking 2 pulses as a group, then respectively carrying out phase deblurring processing and frequency estimation on each segment, finding out the position LostPosition1 of the lost pulse through error analysis, and recording the residual effective frequency estimation value
Figure BDA0001820852910000073
The specific method comprises the following steps:
21) segmenting coherent pulse train by taking 2 pulses as a group, and setting the total M1And (3) if the number of pulses in the whole observation time is odd, the last pulse is not considered, the phase of the two pulse signals in each segment is deblurred by using the frequency value estimated in the step (1), and the mth segment of signals is subjected to the second stage of pulse signalThe phase deblurring formula for the 2 pulses is:
Figure BDA0001820852910000081
where round (x) is the nearest integer to x, φm,1(n) and phim,2(n) phases of the (n + 1) th sampling signals of the 1 st pulse and the 2 nd pulse in the m-th section, respectively,
Figure BDA0001820852910000082
for a first frequency estimate, Δ ═ 1/fsFs is the sampling rate, and k is taken for the first pulse signal in the mth segmentm,1When the phase of each pulse signal in the mth data segment is 0, the phase deblurring expression is as follows:
φm(n+K(j-1))＝φm,j(n)+2km,jπ,j＝1,2
22) and carrying out frequency estimation on each section of data to obtain a frequency estimation value, wherein the expression is as follows:
Figure BDA0001820852910000083
23) calculating M1The average value and the standard deviation of the frequency estimation values of the data of the segment, when the difference between the frequency estimation value and the average value of the data of the segment exceeds the set multiple of the standard deviation, the pulse loss exists in the data of the segment, the serial number of the data of the segment is recorded into the position LostPosition1 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure BDA0001820852910000084
And 3, deleting the first pulse, segmenting the residual coherent pulse train by taking 2 pulses as a group, then respectively carrying out phase ambiguity resolution and frequency estimation on each segment, finding out the position LostPosition2 of the lost pulse through error analysis, and recording the residual effective frequency estimation value
Figure BDA0001820852910000085
The specific method comprises the following steps:
31) deleting the first pulse, and then segmenting the rest coherent pulse train in the step 2 by taking 2 pulses as a group, wherein the total M is set2Segment, if P is even, the last pulse is not considered, repeat steps 21), 22), obtain the frequency estimate in each segment of data
Figure BDA0001820852910000086
32) Calculate this M2The average value and the standard variance of the frequency estimation values of the data segments are determined, when the difference between the frequency estimation value and the average value of a certain data segment exceeds a certain multiple of the standard variance, the data segment is considered to have pulse loss, the serial number of the data segment is recorded into the position LostPosition2 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure BDA0001820852910000087
Step 4, dividing the reference pulse train into a plurality of large segments by utilizing the position information LostPosition1 and LostPosition2 of the lost pulse, acquiring the effective minimum pulse number in all the large segments, further segmenting each large segment by using the effective minimum pulse number, then respectively carrying out phase deblurring and frequency estimation on each small segment, eliminating invalid small segments through error analysis, and synthesizing the rest small segments to obtain a frequency estimation value
Figure BDA0001820852910000091
The specific method comprises the following steps:
41) the effective frequency estimates obtained in steps 2 and 3 are
Figure BDA0001820852910000092
And
Figure BDA0001820852910000093
combined together and their average calculated and recorded as
Figure BDA0001820852910000094
42) Combining the pulse loss position information LostPosition1 and LostPosition2 obtained in steps 2 and 3 to obtain more complete pulse loss position information LostNo, the expression of which is:
LostNo＝sort[LostPosition1*2-1,LostPosition2]
43) segmenting the whole coherent pulse train by using the position information in the LostNo, calculating the number of pulses in each segment, and then taking the minimum value with the number of the pulses larger than 2 and recording the minimum value as Q;
44) for the section with the number of pulses larger than Q, taking Q pulses as a group, subdividing to obtain smaller sections, and performing phase ambiguity resolution and frequency estimation on each small section, wherein for the Q pulse of the mth small section in a certain large section, the phase ambiguity resolution formula is as follows:
Figure BDA0001820852910000095
for the first pulse signal of the small segment, k is takenm,1If it is 0, the phase deblurring expression of each pulse signal in the small segment data is:
φm(n+K(q-1))＝φm,q(n)+2km,qπ,q＝1,…,Q
the expression of the frequency estimate for the small segment data is:
Figure BDA0001820852910000096
meanwhile, recording the information that each small segment belongs to a certain large segment, and recording the information as Flag;
45) calculating the frequency estimation value estimated from all small segments
Figure BDA0001820852910000097
When the difference between the frequency estimation value of a small segment and the mean value exceeds the standard valueWhen the quasi-variance is set to be multiple, the frequency estimated by the small segment data is considered to be unreliable, and the value corresponding to the small segment in the Flag is set to be invalid;
46) calculating the average value of the frequency estimated values obtained by estimating all effective small segments and recording the average value as
Figure BDA0001820852910000101
Step 5, searching the largest continuous effective small sections, respectively forming the effective small sections in the large sections into an observation interval, then respectively performing phase ambiguity resolution and frequency estimation on each observation interval, and averaging to obtain a final frequency estimation value, wherein the method specifically comprises the following steps:
51) counting the number of the continuous effective small segments in each large segment, finding out the large segment with the maximum number of the continuous effective small segments, and if the number T of the large segments meets the condition, respectively combining the continuous effective small segments in the large segments into an observation interval, namely T observation intervals;
52) performing pulse phase deblurring processing in an observation interval, wherein if Q' pulses are shared in the observation interval, the phase deblurring formula of the Q pulse in the t observation interval is as follows:
Figure BDA0001820852910000102
for the first pulse signal of the observation interval, k is takent,1If the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is 0, the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is as follows:
φt(n+K(q-1))＝φt,q(n)+2kt,qπ,q＝1,…,Q'
and carrying out frequency estimation on each observation interval, wherein the frequency estimation expression of a single observation interval is as follows:
Figure BDA0001820852910000103
and synthesizing the frequency estimation values of the T observation intervals to obtain a final frequency estimation value, wherein the expression is as follows:
Figure BDA0001820852910000104
example 1
The flow of this example is shown in FIG. 1. Setting the sampling frequency as 100MHz, the observation time as 0.012 seconds, the radar signal carrier frequency as 21.56789MHz, the pulse repetition frequency as 20KHz, the pulse width as 0.25 microseconds, the signal-to-noise ratio as 20, and the pulse loss rate as 0.1, the number of pulses P transmitted in the whole pulse observation time is determined0The number of actually received pulses is 240, the number of actually received pulses is 216, the number of sampling points in each pulse is 25, and the number of sampling points in each pulse repetition period is 5000. The method for estimating the frequency of the coherent pulse signal under the condition of pulse loss comprises the following specific steps:
step 1, extracting fuzzy phases of coherent pulses, performing phase ambiguity resolution in a pulse range, performing frequency estimation on each pulse respectively, and averaging to obtain a primary frequency estimation value.
11) The blurred phases of the coherent pulses are extracted. The phase values of points within a pulse before they are phase deblurred are shown in figure 2.
12) And (5) performing intra-pulse phase deblurring processing. The phase values of the points within a pulse after phase deblurring are shown in figure 3.
13) And (4) estimating the frequency. The estimated frequency values of the pulses, their mean value of 21.5674264MHz and the error of 463.6Hz are shown in FIG. 4.
And 2, segmenting the coherent pulse train by taking 2 pulses as a group, not considering the residual pulses, then respectively performing phase ambiguity resolution and frequency estimation on each segment, and finding out the position LostPosition1 of the lost pulse through error analysis.
21) And (5) performing two-pulse internal phase deblurring processing. The phase values of the points in the two pulses of a certain segment before the points are not subjected to the phase deblurring processing are shown in fig. 5, the phase values of the points in the two pulses of a certain segment after the points are subjected to the phase deblurring processing are shown in fig. 6, and the right graph of fig. 6 is a partial enlargement of the left graph.
22) And (4) estimating the frequency. The frequency values obtained from the segment estimates are shown in fig. 7.
23) Calculating the average value and standard deviation of the frequency values obtained by each segmental estimation, finding out the pulse loss position, recording the pulse loss position in LostPosition1, deleting the corresponding frequency estimation value, and recording the residual effective frequency estimation value as
Figure BDA0001820852910000111
Table 1 shows the segment numbers corresponding to the found pulse losses, and the effective frequency estimation is shown in fig. 8.
TABLE 1
Number of segments 18 40 48 62 69 77 106
And 3, deleting the first pulse, segmenting the coherent pulse train by taking 2 pulses as a group, not considering the rest pulses, performing phase ambiguity resolution and frequency estimation on each segment, and finding out the position LostPosition2 of the lost pulse through error analysis.
31) And (4) estimating the frequency. The frequency values obtained from the segment estimates are shown in fig. 9.
32) Calculating the average value and standard deviation of the frequency values obtained by each segmental estimation, finding out the pulse loss position, recording the pulse loss position in LostPosition2, deleting the corresponding frequency estimation value, and recording the residual effective frequency estimation value as
Figure BDA0001820852910000124
Table 2 shows the segment numbers corresponding to the found pulse losses, and the effective frequency estimation is shown in fig. 10.
TABLE 2
Number of segments 6 11 12 16 24 27 34 44 45 59 92 97 101
And 4, dividing the reference pulse string into a plurality of large sections by using the position information LostPosition1 and LostPosition2 of the lost pulse, acquiring the effective minimum pulse number in all the large sections, further segmenting each large section by using the effective minimum pulse number, then respectively carrying out phase deblurring and frequency estimation on each small section, eliminating the small sections with too large errors, and synthesizing the residual small sections to obtain a frequency estimation value.
41) The effective frequency estimates obtained in steps 2 and 3 are
Figure BDA0001820852910000121
And
Figure BDA0001820852910000122
combined together and their average calculated and recorded as
Figure BDA0001820852910000123
The frequency value obtained by the second estimation is 21.5678472MHz, and the error is 42.8 Hz.
42) The pulse missing position information LostPosition1 and LostPosition2 obtained in steps 2 and 3 are combined to obtain more complete pulse missing position information LostPosition, see table 3.
TABLE 3
Number of pulses 12 22 24 32 35 48 54 68 79 88
Number of pulses 90 95 118 123 137 153 184 194 202 211
43) And (3) segmenting the whole coherent pulse train by using the position information in the LostNo, calculating the number of pulses in each segment, and then taking the minimum value with the number of pulses larger than 2 to be recorded as Q. Table 4 gives the number of pulses in each segment from which Q can be found to be 3.
TABLE 4
Number of segments 1 2 3 4 5 6 7 8 9 10
Number of pulses 11 8 0 6 1 11 4 12 9 7
Number of segments 11 12 13 14 15 16 17 18 19 20
Number of pulses 0 3 21 3 12 14 29 8 6 7
44) For the segments with the number of pulses larger than Q, Q pulses are taken as a group and subdivided to obtain smaller segments, then the phase deblurring and frequency estimation are carried out on each small segment by a method similar to the step 2, and meanwhile, the information of each small segment belonging to a certain large segment is recorded and recorded as Flag. The frequency values estimated from the segments are shown in fig. 11, and table 5 gives information of which large segment each small segment belongs to.
TABLE 5
First group 1 1 1 2 2 4 4 6 6 6 7
Second group 8 8 8 8 9 9 9 10 10 12 13
Third group 13 13 13 13 13 13 14 15 15 15 15
Fourth group 16 16 16 16 17 17 17 17 17 17 17
Fifth group 17 17 18 18 19 19 20 20
45) Calculating the frequency estimation value estimated from all small segments
Figure BDA0001820852910000131
When the difference between the frequency estimation value of a small segment data and the mean value exceeds a certain multiple of the standard variance, the frequency estimated by the small segment data is considered to be unreliable, and the value corresponding to the small segment in the Flag is set to be invalid. It is calculated that the frequency value estimated from the 5 th small segment data is not reliable, as can be seen from fig. 11.
46) Calculating the average value of the frequency estimated values obtained by estimating all effective small segments and recording the average value as
Figure BDA0001820852910000132
The frequency value obtained by the third estimation is 21.5678881MHz, and the error is 1.9 Hz.
And 5, searching the largest effective small sections, forming the effective small sections in the large sections into an observation interval, performing phase ambiguity resolution and frequency estimation on each interval, and performing comprehensive analysis to obtain a final frequency estimation value.
51) Counting the number of the continuous effective small segments in each large segment, finding out the large segment with the maximum number of the continuous effective small segments, and if the number T of the large segments meets the condition, respectively combining the effective small segments in the large segments into an observation interval, namely T observation intervals. As can be seen from table 4, the number of consecutive valid small segments is 9, and only the 17 th large segment satisfies the condition.
52) And deblurring the pulse phase in the observation interval.
53) Frequency estimation was performed for each observation interval in a similar manner to step 2. Since there is only one observation interval, there is only one frequency estimate, and this estimate is the final frequency estimate, which is 21.5678889MHz with an error of 1 Hz.

Claims (1)

1. The method for estimating the frequency of the coherent pulse signal under the condition of pulse loss is characterized by comprising the following specific steps of:
step 1, extracting fuzzy phases of coherent pulses, performing phase ambiguity resolution in a pulse range, performing frequency estimation on each pulse respectively, and averaging to obtain a primary frequency estimation value, wherein the specific method comprises the following steps:
11) calculating the difference value of adjacent phases in the (p + 1) th pulse, wherein the specific formula is as follows:
Δφp,o(n)＝φo(n+Kp)-φo(n-1+Kp),n＝1,…,Ns-1；
φo(N + Kp) is the fuzzy phase, NsThe number of sampling points in a pulse, K is the number of sampling points in a repetition period, and P is the number of pulses in the whole observation time;
12) starting from n equal to 1, if Δ φp,o(n) is greater than or equal to 0, then phip(n)＝φo(n + Kp), n ═ n +1, repeat step 11), φp(n) is the phase of the (n + 1) th sampling signal in the (p + 1) th pulse after the deblurring; otherwise, phio(k+Kp)＝φo(k+Kp)+2π,k＝n,n+1,…,Ns-1，φp(n)＝φo(n + Kp), repeating step 11);
13) respectively carrying out frequency estimation on each pulse, and averaging to obtain a primary frequency estimation value, wherein the formula for respectively carrying out frequency estimation on each pulse is as follows:
Figure FDA0002746002070000011
wherein [ Delta ] is 1/fs，fsIs the sampling rate;
the primary frequency estimate obtained is:
Figure FDA0002746002070000012
step 2, segmenting coherent pulse trains by taking 2 pulses as a group, and then respectively segmenting each pulse trainPhase deblurring processing and frequency estimation are carried out on the sections, then the position LostPosition1 of the lost pulse is found out through error analysis, and the residual effective frequency estimation value is recorded
Figure FDA0002746002070000013
The specific method comprises the following steps:
21) segmenting coherent pulse train by taking 2 pulses as a group, and setting the total M1And (3) if the number of pulses in the whole observation time is an odd number, the last pulse is not considered, the phase of two pulse signals in each segment is deblurred by using the frequency value estimated in the step (1), and the phase deblurring formula of the 2 nd pulse in the m-th segment signal is as follows:
Figure FDA0002746002070000021
where round (x) is the nearest integer to x, φm,1(n) and phim,2(n) phases of the (n + 1) th sampling signals of the 1 st pulse and the 2 nd pulse in the m-th section, respectively,
Figure FDA0002746002070000022
for a first frequency estimate, Δ ═ 1/fs，fsFor the sampling rate, k is taken for the first pulse signal in the m-th segmentm,1When the phase of each pulse signal in the mth data segment is 0, the phase deblurring expression is as follows:
φm(n+K(j-1))-φm,j(n)+2km,jπ,j＝1,2
22) and carrying out frequency estimation on each section of data to obtain a frequency estimation value, wherein the expression is as follows:
Figure FDA0002746002070000023
23) calculating M1Mean and standard deviation of frequency estimation of data of a segment, when the difference between the frequency estimation of a data segment and the meanWhen the standard deviation exceeds the set multiple of the standard deviation, the pulse loss exists in the data, the serial number of the data is recorded in the position LostPosition1 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure FDA0002746002070000024
And 3, deleting the first pulse, segmenting the residual coherent pulse train by taking 2 pulses as a group, then respectively carrying out phase ambiguity resolution and frequency estimation on each segment, finding out the position LostPosition2 of the lost pulse through error analysis, and recording the residual effective frequency estimation value
Figure FDA0002746002070000025
The specific method comprises the following steps:
31) deleting the first pulse, and then segmenting the rest coherent pulse train in the step 2 by taking 2 pulses as a group, wherein the total M is set2Segment, if P is even, the last pulse is not considered, repeat steps 21), 22), obtain the frequency estimate in each segment of data
Figure FDA0002746002070000026
32) Calculate this M2The average value and the standard variance of the frequency estimation values of the data segments are determined, when the difference between the frequency estimation value and the average value of a certain data segment exceeds a certain multiple of the standard variance, the data segment is considered to have pulse loss, the serial number of the data segment is recorded into the position LostPosition2 of the lost pulse, the corresponding frequency estimation value is deleted, and the residual effective frequency estimation value is recorded as
Figure FDA0002746002070000027
Step 4, dividing the reference pulse train into a plurality of large segments by utilizing the position information LostPosition1 and LostPosition2 of the lost pulse, acquiring the effective minimum pulse number in all the large segments, and further dividing each large segment by the effective minimum pulse numberAnd respectively carrying out phase ambiguity resolution and frequency estimation on each small section, eliminating invalid small sections through error analysis, and synthesizing the remaining small sections to obtain a frequency estimation value
Figure FDA0002746002070000031
The specific method comprises the following steps:
41) the effective frequency estimates obtained in steps 2 and 3 are
Figure FDA0002746002070000032
And
Figure FDA0002746002070000033
combined together and their average calculated and recorded as
Figure FDA0002746002070000034
42) Combining the pulse loss position information LostPosition1 and LostPosition2 obtained in steps 2 and 3 to obtain more complete pulse loss position information LostNo, the expression of which is:
LostNo＝sort[LostPosition1*2-1,LostPosition2]
43) segmenting the whole coherent pulse train by using the position information in the LostNo, calculating the number of pulses in each segment, and then taking the minimum value with the number of the pulses larger than 2 and recording the minimum value as Q;
44) for the section with the number of pulses larger than Q, taking Q pulses as a group, subdividing to obtain smaller sections, and performing phase ambiguity resolution and frequency estimation on each small section, wherein for the Q pulse of the mth small section in a certain large section, the phase ambiguity resolution formula is as follows:
Figure FDA0002746002070000035
for the first pulse signal of the small segment, k is takenm,1If it is 0, the phase deblurring expression of each pulse signal in the small segment data is:
φm(n+K(q-1))＝φm,q(n)+2km,qπ,q＝1,…,Q
the expression of the frequency estimate for the small segment data is:
Figure FDA0002746002070000036
meanwhile, recording the information that each small segment belongs to a certain large segment, and recording the information as Flag;
45) calculating the frequency estimation value estimated from all small segments
Figure FDA0002746002070000037
When the difference between the frequency estimation value of a certain small segment data and the mean value exceeds the set multiple of the standard variance, the frequency estimated by the small segment data is considered to be unreliable, and the value corresponding to the small segment in the Flag is set to be invalid;
46) calculating the average value of the frequency estimated values obtained by estimating all effective small segments and recording the average value as
Figure FDA0002746002070000041
Step 5, searching the largest continuous effective small sections, respectively forming an observation interval by the continuous effective small sections in the large sections, then respectively performing phase ambiguity resolution and frequency estimation on each observation interval, and averaging to obtain a final frequency estimation value, wherein the method specifically comprises the following steps:
51) counting the number of the continuous effective small segments in each large segment, finding out the large segment with the maximum number of the continuous effective small segments, and if the number T of the large segments meets the condition, respectively combining the continuous effective small segments in the large segments into an observation interval, namely T observation intervals;
52) performing pulse phase deblurring processing in an observation interval, wherein if Q' pulses are shared in the observation interval, the phase deblurring formula of the Q pulse in the t observation interval is as follows:
Figure FDA0002746002070000042
for the first pulse signal of the observation interval, k is takent,1If the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is 0, the phase ambiguity resolution expression of each pulse signal in the data in the observation interval is as follows:
φt(n+K(q-1))＝φt,q(n)+2kt,qπ,q＝1,…,Q'
and carrying out frequency estimation on each observation interval, wherein the frequency estimation expression of a single observation interval is as follows:
Figure FDA0002746002070000043
and synthesizing the frequency estimation values of the T observation intervals to obtain a final frequency estimation value, wherein the expression is as follows:
Figure FDA0002746002070000044
CN201811164794.XA 2018-10-07 2018-10-07 Coherent pulse signal frequency estimation method under pulse loss Active CN109270344B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811164794.XA CN109270344B (en) 2018-10-07 2018-10-07 Coherent pulse signal frequency estimation method under pulse loss

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811164794.XA CN109270344B (en) 2018-10-07 2018-10-07 Coherent pulse signal frequency estimation method under pulse loss

Publications (2)

Publication Number Publication Date
CN109270344A CN109270344A (en) 2019-01-25
CN109270344B true CN109270344B (en) 2021-01-08

Family

ID=65195636

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811164794.XA Active CN109270344B (en) 2018-10-07 2018-10-07 Coherent pulse signal frequency estimation method under pulse loss

Country Status (1)

Country Link
CN (1) CN109270344B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112162152B (en) * 2020-08-31 2024-01-26 南京亿杰明信息技术有限公司 Sine wave coherent pulse train signal frequency estimation method based on phase straight line fitting

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08146022A (en) * 1994-11-21 1996-06-07 Meidensha Corp Pulse-counting circuit
JP2004012378A (en) * 2002-06-10 2004-01-15 Furuno Electric Co Ltd Frequency estimating apparatus and apparatus for receiving signal for positioning
CN101080646A (en) * 2004-12-18 2007-11-28 莱卡地球系统公开股份有限公司 Method for electronic measurement
CN103487669A (en) * 2013-08-16 2014-01-01 西安电子科技大学 Phase noise measurement method based on phase characteristic processing between any frequency signals
CN104483544A (en) * 2014-12-01 2015-04-01 陕西海泰电子有限责任公司 High-accuracy frequency/cycle measuring method of single-channel counter
CN105738696A (en) * 2016-04-18 2016-07-06 天津大学 Frequency estimation method and device for all-phase time-shift phase difference
CN106886007A (en) * 2017-02-24 2017-06-23 广州比逊电子科技有限公司 Unmanned plane localization method and system
KR20170084955A (en) * 2016-01-13 2017-07-21 목포대학교산학협력단 Failure diagnosing method and apparatus of wind turbine
CN107561357A (en) * 2017-07-05 2018-01-09 中国电子科技集团公司第三十八研究所 A kind of high-precision instantaneous frequency measurement method and apparatus based on channelizing

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08146022A (en) * 1994-11-21 1996-06-07 Meidensha Corp Pulse-counting circuit
JP2004012378A (en) * 2002-06-10 2004-01-15 Furuno Electric Co Ltd Frequency estimating apparatus and apparatus for receiving signal for positioning
CN101080646A (en) * 2004-12-18 2007-11-28 莱卡地球系统公开股份有限公司 Method for electronic measurement
CN103487669A (en) * 2013-08-16 2014-01-01 西安电子科技大学 Phase noise measurement method based on phase characteristic processing between any frequency signals
CN104483544A (en) * 2014-12-01 2015-04-01 陕西海泰电子有限责任公司 High-accuracy frequency/cycle measuring method of single-channel counter
KR20170084955A (en) * 2016-01-13 2017-07-21 목포대학교산학협력단 Failure diagnosing method and apparatus of wind turbine
CN105738696A (en) * 2016-04-18 2016-07-06 天津大学 Frequency estimation method and device for all-phase time-shift phase difference
CN106886007A (en) * 2017-02-24 2017-06-23 广州比逊电子科技有限公司 Unmanned plane localization method and system
CN107561357A (en) * 2017-07-05 2018-01-09 中国电子科技集团公司第三十八研究所 A kind of high-precision instantaneous frequency measurement method and apparatus based on channelizing

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
frequency estimation of uncooperative coherent pulse radars;Jing Gai et.al;《MILCOM 2007-IEEE Military Communication Conference》;20071231;正文第1-7页 *
一种新的脉冲重复频率估计方法;李杨寰 等;《电子信息对抗技术》;20070331;第22卷(第2期);第18-22页 *

Also Published As

Publication number Publication date
CN109270344A (en) 2019-01-25

Similar Documents

Publication Publication Date Title
Su et al. High-speed multi-target detection with narrowband radar
CN109164438B (en) LFM coherent pulse train-based joint estimation method for arrival time and arrival frequency
CN107356922B (en) Multi-frequency radar target tracking method for ambiguity resolution by spread-table-aided method
CN101592733A (en) A kind of parallel real-time imaging processing method for inverse synthetic aperture radar
CN108809354B (en) Long-time accumulation method for high dynamic Doppler capture
CN109031261B (en) Time difference estimation method and device
CN115828074B (en) Combined estimation method for target positioning parameters of radiation source
CN111623703A (en) Novel Kalman filtering-based Beidou deformation monitoring real-time processing method
CN114924229B (en) Time delay-Doppler domain underwater target differential positioning method with asynchronous clocks
CN107976676B (en) Airborne synthetic aperture radar moving target imaging method
CN109975842A (en) A kind of blind catching method of Big Dipper satellite signal high-precision based on wavelet transformation
Zheng et al. Radar detection and motion parameters estimation of maneuvering target based on the extended keystone transform (July 2018)
CN109270344B (en) Coherent pulse signal frequency estimation method under pulse loss
CN105548985B (en) Maneuvering target tracking method based on RAV-Jerk models
CN108196238B (en) Clutter map detection method based on adaptive matched filtering under Gaussian background
CN115877350B (en) Method and device for estimating time-varying target angle of radar with sum and difference beam system
CN112114296B (en) Parameter estimation method and system for unmanned aerial vehicle collaborative TDOA/FDOA composite positioning
CN113534115A (en) Active and passive combined processing sonar target high-precision line spectrum extraction method
CN113030945B (en) Phased array radar target tracking method based on linear sequential filtering
CN106569182A (en) Phase coding signal carrier frequency estimation method based on minimum entropy
CN103185879A (en) Method for detecting single-channel synthetic aperture radar moving target with high radial velocity target
CN115221768A (en) Improved unscented particle filter direct tracking method based on multi-nested array
Li et al. Doppler keystone transform for SAR imaging of moving targets
CN111983579B (en) Method for eliminating radar speed ambiguity by using pulse repetition time of difference
CN111190173B (en) Phased array radar target tracking method based on predicted value measurement conversion

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant