JP2003133922A - Pulse rain detecting device and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To overcome a problem such that, when the pulse train from a signal source included in a received pulse train is short and fluctuation is included in its PRI, the peak level at the position of the PRI is reduced, noise increases, and consequently the pulse train is hardly detected. SOLUTION: This device and method is configured to compute a moving sum Dk (t) of complex trigonometric function within a moving window [t-ντk , t+ντk ], which moves with time, based on arrival times tm of receiving pulse which construct a pulse interval tm -tn detected by simple PRI filters 121 to 12k .

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pulse train detection device and a pulse train detection method for detecting a plurality of pulse trains oscillated from a plurality of signal sources.

[0002]

2. Description of the Related Art FIG. 9 is a block diagram showing a conventional pulse train detection apparatus disclosed in, for example, Japanese Patent Publication No. 62-26603, in which 1 is a plurality of pulse trains oscillated from a plurality of signal sources. Arrival time t ₁ , t ₂ , ...
.., t _N are input terminals, 2 ₁ to 2 _K refer to the arrival times t ₁ , t ₂ , ..., T _N of the received pulse and refer to all pulse pairs (t _n , t _m ). pulse interval _t m -t
_n is calculated, and the calculated pulse interval t _m −
It is a simple PRI filter that detects a pulse interval t _m -t _n belonging to a subdivision section (hereinafter, referred to as PRI bin) of a preset pulse repetition interval (hereinafter, referred to as PRI) among t _n . However, m = 2, ..., N,
n = 1, ..., N-1.

3 _{1 to} 3 _K are simple PRI filters 2 ₁ to 2
Complex trigonometric function calculator that performs an operation of a complex trigonometric function based on the pulse interval is detected by the _K t _m -t _n arrival of a received pulse constituting time t _m, 4 ₁ ~4 _K is a complex trigonometric function calculator 3 a complex adder for obtaining the sum of ₁ to 3 _K calculation result.

Next, the operation will be described. The conventional pulse train detection device detects the pulse train for each signal source based on the difference in PRI, and thus takes the sum of complex trigonometric functions over the entire observation time. First, the basic principle of the pulse train detection device will be described. The basic principle corresponds to a method called PRI conversion, which is shown in "New method for estimating pulse repetition interval", National Convention of Information Systems Division, Institute of Electronics, Information and Communication Engineers, 1983.

First, the arrival time of the received pulse is t ₁ ,
As t ₂ , ..., T _N , the input pulse train is expressed by the sum of delta functions as shown below.

[Equation 1] Here, the PRI conversion is the input pulse train g of the equation (1).
(T) is converted by the following equation.

[Equation 2]

The function D (τ) after PRI conversion is a function on the PRI axis (variable τ), and the equation (2) is as follows for the input pulse train g (t) of the equation (1). Take shape.

[Equation 3] Here, the expression without the term of the complex trigonometric function on the right side of Expression (3) is the autocorrelation function of the input pulse train g (t). In the autocorrelation function of the input pulse train, the PRI of the included signal source
Although a sharp peak appears at the position of PRI, a similar peak appears at the position of integer multiple of PRI (subharmonic), which makes PRI detection difficult, while PRI conversion suppresses subharmonic. Then, the PRI can be detected from the position of the peak.

By the way, the PRI conversion of the equation (3) is an expression of continuous time, but when it is realized as an actual device, it is handled in discrete time. In order to discretize the PRI conversion, it is necessary to subdivide the PRI axis and divide it into a plurality of PRI bins. The width of the PRI bin is Δτ, the kth PR
When the center PRI of the I bin is τ _k = (k−1 / 2) Δτ, the value of the discrete PRI transformation corresponding to the kth PRI bin is defined by the following equation.

[Equation 4] The pulse train detection device of FIG. 9 obtains the value of this discrete PRI conversion.

The operation of the pulse train detecting device shown in FIG. 9 will be specifically described below. First, the arrival time t of each received pulse of a plurality of pulse trains that the input terminal 1 oscillates from a plurality of signal sources.
₁ , t ₂ , ..., T _N are input, the simple PRI filters 2 ₁ to 2 _K receive the arrival times t ₁ , t ₂ ,
, T _N , all pulse pairs (t _n ,
calculating the pulse interval _t m -t _n of t _m). Of the pulse interval _t m -t _n that the calculated, to detect the pulse interval _t m -t _n belonging to a preset PRI bottle.

For example, the k-th simple PRI filter 2_k
Is all pulse pairs (t_n, T _m) Pulse interval t
_m-T_nPulse matching the kth PRI bin in
Interval t_m-T_nArrival time t of the received pulses that make up_mof
Pass only. That is, τ_k−Δτ / 2 <t_m-T_n≤
τ_kArrival time t of the received pulse satisfying + Δτ / 2_mof
Pass only.

The complex trigonometric function calculators 3 _{1 to} 3 _K are simple P
Arrival time t _{m of the} reception pulse from the RI filters 2 _{1 to} 2 _K
When receiving, the complex trigonometric function exp (2πit _m / τ _k ) is calculated based on the arrival time t _m . The complex adders 4 ₁ to 4 _K calculate the sum of the calculation results of the complex trigonometric function calculators 3 _{1 to} 3 _K as shown below.

[Equation 5] Then, each of the complex adders 4 ₁ to 4 _K outputs a pulse train detection signal when the sum of the calculation results is larger than a predetermined threshold value.

Here, FIGS. 10 and 11 are explanatory views showing the detection results of the conventional pulse train detection device. However, it is assumed that the input pulse train is a composite pulse train in which pulse trains from three signal sources (A, B, C in the figure) as shown in FIG. 5 are overlapped. As shown in FIG. 5, A and B of the signal sources can receive the pulse only for a part of the entire observation time, and in particular, B is a very short time. Even for such an input pulse train, if there is no fluctuation in the PRI, as shown in FIG. 10, the graph of the absolute value of the discrete PRI conversion corresponds to three signal sources and the positions of those PRIs are sharp. Since the peak appears, the pulse train can be easily detected. On the other hand, when the PRI fluctuates by about 10%, as shown in FIG. 11, the peak level at the PRI position decreases and noise increases, making it difficult to detect the pulse train.

[0012]

Since the conventional pulse train detection apparatus is configured as described above, when the pulse train of the signal source included in the received pulse train is short and the PRI includes fluctuations, There is a problem that the peak level at the PRI position decreases and noise increases, making it difficult to detect the pulse train.

The present invention has been made to solve the above problems, and accurately detects a pulse train even when the pulse train of the signal source included in the received pulse train is short and the PRI includes fluctuations. An object of the present invention is to obtain a pulse train detection device and a pulse train detection method that can perform the above.

[0014]

SUMMARY OF THE INVENTION A pulse train detection apparatus according to the present invention uses a complex trigonometric function within a moving window that moves with time, based on the arrival time of received pulses that form the pulse intervals detected by the detection means. A moving sum calculating means for calculating the moving sum is provided.

In the pulse train detecting device according to the present invention, the moving sum calculating means uses a moving window having a width proportional to the pulse repetition interval.

In the pulse train detecting apparatus according to the present invention, when the detecting means detects the pulse interval belonging to the sub-interval of the pulse repetition interval, the width of the sub-interval is set according to the expected fluctuation width. Is.

In the pulse train detecting apparatus according to the present invention, the comparing means counts the number of pulses in the moving window and sets the threshold value according to the number of pulses.

In the pulse train detecting apparatus according to the present invention, the detecting means detects the pulse interval belonging to the subdivided section of the pulse repetition interval by using the PRI filter.

In the pulse train detecting method according to the present invention, the moving sum of the complex trigonometric function is calculated within the moving window moving with time based on the arrival time of the received pulse which constitutes the pulse interval belonging to the sub-interval of the pulse repetition interval. It is something that is done.

The pulse train detecting method according to the present invention uses a moving window having a width proportional to the pulse repetition interval.

In the pulse train detecting method according to the present invention, when the pulse interval belonging to the subdivided section of the pulse repetition interval is detected, the width of the subdivided section is set according to the expected fluctuation width.

In the pulse train detecting method according to the present invention, the number of pulses in the moving window is counted, and the threshold value is set according to the number of pulses.

The pulse train detection method according to the present invention is a PR
The I filter is used to detect a pulse interval belonging to a subdivided section of the pulse repetition interval.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described below. Embodiment 1. 1 is a block diagram showing a pulse train detection device according to a first embodiment of the present invention. In FIG.
Is an input terminal for inputting the arrival times t ₁ , t ₂ , ..., T _N of each received pulse of a plurality of pulse trains oscillated from a plurality of signal sources, and 12 _{1 to} 12 _K are arrival times t _{1 of the} received pulses. , _t 2, · · ·, with reference to _{t n,} all pulse pair _(t n, t _m) and calculates the pulse interval _t m -t _n of, among the pulse interval _t m -t _n that the calculated ,
It is a simple PRI filter (pulse interval calculation means, detection means) that detects a pulse interval t _m -t _n belonging to a PRI bin of a preset PRI. However, m = 2, ...
., N, n = 1, ..., N-1.

13 _{1 to} 13 _K are simple PRI filters 12
_1-12 complex trigonometric function calculator that performs an operation of a complex trigonometric function based on the arrival time _{t m} of the received pulses constituting the detected pulse interval _t m -t _n by _K, moves with 14 1 to 14 _K time Is a moving window type complex adder that calculates the moving sum of the complex trigonometric function within the moving window. Note that the complex trigonometric function calculators 13 _{1 to} 13 _K and the moving window type complex adder 14 ₁
The moving sum calculation means is composed of ˜14 _K. 15 ₁
˜15 _K is a threshold processing unit (comparing means) that outputs a pulse train detection signal when the absolute value of the moving sum calculated by the moving window type complex adders 14 _{1 to} 14 _K is larger than a predetermined threshold.
16 _{1 to} 16 _K are output terminals. FIG. 2 is a flowchart showing a pulse train detection method according to the first embodiment of the present invention.

Next, the operation will be described. First, when the input terminals 11 receive the arrival times t ₁ , t ₂ , ..., T _N of the received pulses of a plurality of pulse trains oscillated from a plurality of signal sources, the simple PRI filters 12 _{1 to} 12 _K The pulse intervals t of all the pulse pairs (t _n , t _m ) are referred to with reference to the arrival times t ₁ , t ₂ , ..., T _N of the received pulses.
calculating the _m -t _n (step ST1). Of the pulse interval _t m -t _n that the calculated, to detect the pulse interval _t m -t _n belonging to a preset PRI bins (step ST2).

For example, the k-th simple PRI filter 12
_kIs all pulse pairs (t_n, T_m) Pulse interval
t_m-T_nPal that matches the kth PRI bin in
Interval t_m-T_nArrival time t of the received pulses that make up_m
Only pass through. That is, τ _k−Δτ / 2 <t_m-T_n
≤τ_kArrival time t of the received pulse satisfying + Δτ / 2 _m
Only pass through.

The complex trigonometric function calculator ₁₃ 1 to 13 _K are simple PRI the filter ₁₂ 1 receive the arrival time _{t m} of the received pulses from to 12 _K, a complex trigonometric function exp (2πit _m based on the arrival time _{t m} / Τ _k ) is calculated (step ST3). Moving window type complex adder 14 _{1 to} 14
As shown below, _K calculates the moving sum D _k (t) of the complex trigonometric function within the moving window [t−ντ _k , t + ντ _k ] moving with time (step ST4). Where ν
Is a constant of proportionality.

[Equation 6]

Threshold processing unit 15₁~ 15_KIs a moving window type
Prime adder 14₁~ 14_KIs moving sum D _kCalculate (t)
And the absolute value of its moving sum | D_k(T) |
Threshold A_kAnd the absolute value | D_k(T) |
Threshold A_kOutput pulse train detection signal when larger
(Step ST5). Where threshold A_kAs a run
| D when processing the pulse train arriving at the dam_k(T)
Γ times the average level of | (square root of root mean square)
It The value of γ is about 3 according to the 3-σ rule.
Set.

The square root of the root mean square value of | D _k (t) | in the case of processing a pulse train that arrives at random is given by the following equation, where T is the width of the moving window and b is the width of the PRI bin. Can be approximated as

[Equation 7] In this case, since T = ντ _k and b = Δτ, the threshold value A _k
Is as follows.

[Equation 8]

As is clear from the above, the first embodiment
According to the simple PRI filter 12 ₁~ 12_KDetected by
Issued pulse interval t_m-T_nOf the received pulses that make up
Arrival time t_mIn a moving window that moves over time based on
Is configured to calculate the moving sum of complex trigonometric functions
, The pulse train of the signal source included in the received pulse train is short.
Therefore, even if the PRI contains fluctuations,
This has the effect of being able to detect a pulse train. Also,
Make the moving window width proportional to the pulse repetition interval.
Therefore, it is equivalent for various pulse repetition intervals.
This has the effect of enabling detection with quality.

Embodiment 2. In the first embodiment, the PRs in which the simple PRI filters 12 _{1 to} 12 _K are set in advance are used.
When detecting the pulse intervals t _m -t _n belonging to the I-bin, it is shown that all the PRI bins have the same width, but the pulse intervals t _m -t _n are detected. You may make it set according to the assumed fluctuation width.

That is, when there is a large fluctuation in the PRI of the signal source (for example, fluctuation of about 10%), the pulse of the same signal source branches into a plurality of PRI bins and the peak value decreases. If there is such a large fluctuation,
As shown in FIG. 3, the width of the PRI bin may be widened according to the expected fluctuation width. Even in this case, since the center PRI of each PRI bin is set so as to satisfy the resolution requirement, the PRI bins will overlap.

Assuming that the expected fluctuation is proportional to PRI, the width of the _k -th PRI bin is b _k = 2ε.
Let τ _k (ε is a constant representing the magnitude of fluctuation). In this case, when processing a pulse train that arrives randomly |
The square root of the root mean square value of D _k (t) | can be approximated as follows.

[Equation 9] Therefore, the threshold A _k is as follows.

[Equation 10] The threshold value processing units 15 ₁ to 15 _K use the above threshold value A _k , so that the pulse train can be accurately detected even when there is a large fluctuation in the PRI of the signal source.

Embodiment 3. 4 is a block diagram showing a pulse train detecting device according to a third embodiment of the present invention. In the figure, the same reference numerals as those in FIG. 17 _{1 to} 17 _K are moving window type adders (comparing means) that count the number of pulses in the moving window and set the threshold value A _k according to the number of pulses.

Next, the operation will be described. In the above-described first and second embodiments, the threshold value processing units 15 ₁ to 15 _{K use the} preset threshold value A _k to obtain the absolute value of the moving sum | D _k (t).
Although the processing for comparing with | has been shown, the moving window type adders 17 _{1 to} 17 _K count the number of pulses in the moving window and set the threshold value A _k according to the number of pulses. You may

That is, the moving window type adders 17 _{1 to} 17 _K are
The number N _k (t) of pulses in the moving window is counted as follows.

[Equation 11] In this case, the square root of the root mean square value of | D _k (t) | when a pulse train that arrives at random is processed can be approximated as follows.

[Equation 12] Therefore, the threshold A _k is as follows.

[Equation 13] The threshold value processing units 15 ₁ to 15 _K use the above-mentioned threshold value A _k , so that the pulse train can be detected with higher accuracy.

FIG. 5 is an explanatory view showing an input pulse train, and the input pulse train is composed of three signal sources (A, B, and
FIG. 7C is a composite pulse train of overlapping pulse trains from C). Both signal sources have 10% fluctuation in PRI, and
It is assumed that 0% pulse omission occurs, and A and B of the signal sources are composed of pulse trains that are shorter than the length of the entire pulse train. When the above-mentioned input pulse train is applied to the pulse detector of the third embodiment, the moving window type complex adder 1
The absolute value | D _k (t) | of the outputs of 4 _{1 to} 14 _K is as shown in FIG. Further, the threshold value A _k set by the moving window type adders 17 _{1 to} 17 _K is as shown in FIG. 7. Then, the threshold value processing units 15 ₁ to 15 _K compare the absolute value | D _k (t) | with the threshold value A _k, and the detection signals output from the output terminals 16 _{1 to} 16 _K are as shown in FIG. 8.

Fourth Embodiment In the above-described _first to third embodiments, the arrival times t _m of the reception pulses that are allowed to pass are determined on the condition that the simple PRI filters 12 _{1 to} 12 _K have the pulse widths corresponding to the two pulses. As described above, the conditions are further tightened so that the arrival time t _m of the reception pulse is passed only when three pulses or more pulses have a pulse width corresponding to equal intervals. You may That is, the PRI filter may be used to detect the pulse interval t _m -t _n belonging to the PRI bin of the preset PRI. Thereby, there is an effect that the pulse train can be detected with higher accuracy.

Embodiment 5. In the above-described _first to fourth embodiments, the simple PRI filters 12 _{1 to} 12 _K or the PRI filters pass the arrival time t _m of the reception pulse, and the complex trigonometric function calculators 13 _{1 to} 13 _K are operated by the complex trigonometric function exp (2.
It has been shown that the calculation of πit _m / τ _k )
The present invention is not limited to this. For example, the simple PRI filters 12 _{1 to} 12 _K or the PRI filters allow the arrival time t _n of the reception pulse to pass, and the complex trigonometric function calculators 13 ₁ to 13 ₁
3 _K may perform the calculation of the complex trigonometric function exp (2πit _n / τ _k ).

Further, the simple PRI filters 12 _{1 to} 12 _K
Alternatively, the PRI filter may set the arrival time t _m , t _{n of the} received pulse.
It was allowed to pass, performs a calculation of the complex trigonometric function calculator ₁₃ 1 to 13 _K the complex trigonometric function _{exp (2πit m / (t m} -t n)) or _{exp (2πit n / (t m} -t n)) You may do it. Furthermore, the complex conjugate of the complex trigonometric function may be calculated.

[0042]

As described above, according to the present invention, the moving sum of the complex trigonometric functions is moved within the moving window that moves with time, based on the arrival time of the received pulse forming the pulse interval detected by the detecting means. Since it is configured to include the moving sum calculation means for calculating, the pulse train can be accurately detected even when the pulse train of the signal source included in the reception pulse train is short and fluctuations are included in the pulse repetition interval. effective.

According to the present invention, since the moving sum calculating means is configured to use the moving window having the width proportional to the pulse repetition interval, it is possible to detect the same quality for various pulse repetition intervals. There is an effect.

According to the present invention, when the detecting means detects the pulse interval belonging to the subdivided section of the pulse repetition interval,
Since the width of the subdivided section is set according to the expected fluctuation width, there is an effect that the detection accuracy of the pulse train can be improved.

According to the present invention, the comparison means is configured to count the number of pulses in the moving window and set the threshold value according to the number of pulses, so that the accuracy of detecting the pulse train can be improved. There is.

According to the present invention, since the detecting means is configured to detect the pulse intervals belonging to the sub-intervals of the pulse repetition interval by using the PRI filter, there is an effect that the detection accuracy of the pulse train can be improved.

According to the present invention, the moving sum of the complex trigonometric functions is calculated within the moving window moving with time based on the arrival time of the received pulse which constitutes the pulse interval belonging to the subdivided section of the pulse repetition interval. Since it is configured, even if the pulse train of the signal source included in the received pulse train is short and fluctuations are included in the pulse repetition interval, there is an effect that the pulse train can be accurately detected.

According to the present invention, since the moving window having the width proportional to the pulse repetition interval is used, there is an effect that detection can be performed with the same quality for various pulse repetition intervals.

According to the present invention, when the pulse interval belonging to the subdivided section of the pulse repetition interval is detected, the width of the subdivided section is set in accordance with the expected fluctuation width. There is an effect that can increase.

According to the present invention, since the number of pulses in the moving window is counted and the threshold value is set according to the number of pulses, there is an effect that the detection accuracy of the pulse train can be improved.

According to the present invention, the PRI filter is used to detect the pulse intervals belonging to the subdivided intervals of the pulse repetition intervals, so that the detection accuracy of the pulse train can be improved.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a pulse train detection device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing a pulse train detection method according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a width of a PRI bin.

FIG. 4 is a configuration diagram showing a pulse train detection device according to a third embodiment of the present invention.

FIG. 5 is an explanatory diagram showing an input pulse train.

FIG. 6 is an explanatory diagram showing an absolute value of an output of a moving window type complex adder.

FIG. 7 is an explanatory diagram showing thresholds set by a moving window type adder.

FIG. 8 is an explanatory diagram showing a detection signal output from an output terminal.

FIG. 9 is a configuration diagram showing a conventional pulse train detection device.

FIG. 10 is an explanatory diagram showing a detection result of a conventional pulse train detection device.

FIG. 11 is an explanatory diagram showing a detection result of a conventional pulse train detection device.

[Explanation of symbols]

11 input terminals, 12₁~ 12_K  Simple PRI filter
(Pulse interval calculation means, detection means), 13₁~ 13_K  
Complex trigonometric function calculator (moving sum calculation means), 14 ₁~ 14
_K  Moving window type complex adder (moving sum calculation means), 15₁~
15_K  Threshold processing unit (comparing means), 16₁~ 16_K  Out
Power terminal, 17₁~ 17_K  Moving window type adder
Dan).

Claims (10)

[Claims] 1. A pulse interval calculation means for calculating the pulse intervals of all pulse pairs with reference to the arrival time of a received pulse, and a preset pulse repetition of the pulse intervals calculated by the pulse interval calculation means. Detecting means for detecting pulse intervals belonging to the interval subdivisions and arrival time of received pulses constituting the pulse intervals detected by the detecting means, based on the moving sum of complex trigonometric functions within a moving window moving with time. A pulse train detection device comprising: a moving sum calculating means for calculating the moving sum and a comparing means for comparing an absolute value of the moving sum calculated by the moving sum calculating means with a predetermined threshold value. 2. The pulse train detection apparatus according to claim 1, wherein the moving sum calculation means uses a moving window having a width proportional to the pulse repetition interval. 3. The detection means, when detecting a pulse interval belonging to a subdivided section of a pulse repetition interval, sets the width of the subdivided section according to an expected fluctuation width. Item 2. A pulse train detection device according to item 2. 4. The comparing means counts the number of pulses in the moving window and sets a threshold value according to the number of the pulses, wherein the comparing means sets the threshold value.
The pulse train detection device described in the paragraph. 5. The pulse train detection according to claim 1, wherein the detection means detects a pulse interval belonging to a subdivided section of the pulse repetition interval using a PRI filter. apparatus. 6. The pulse intervals of all the pulse pairs are calculated with reference to the arrival time of the received pulse, and the pulse intervals belonging to a preset sub-interval of pulse repetition intervals among the calculated pulse intervals are detected. Then, based on the arrival time of the received pulse that constitutes the pulse interval, the moving sum of the complex trigonometric function is calculated within the moving window that moves with time, and the pulse train detection that compares the absolute value of the moving sum with a predetermined threshold value. Method. 7. The pulse train detection method according to claim 6, wherein a moving window having a width proportional to the pulse repetition interval is used. 8. When detecting a pulse interval belonging to a subdivided section of a pulse repetition interval, the width of the subdivided section is set according to an expected fluctuation width.
Alternatively, the pulse train detection method according to claim 7. 9. The pulse train detection according to claim 6, wherein the number of pulses in the moving window is counted, and the threshold value is set according to the number of pulses. Method. 10. The pulse train detection method according to claim 6, wherein a pulse interval belonging to a subdivided section of the pulse repetition interval is detected by using a PRI filter.