Nothing Special   »   [go: up one dir, main page]

JPH08220215A - Signal form judging method - Google Patents

Signal form judging method

Info

Publication number
JPH08220215A
JPH08220215A JP7024928A JP2492895A JPH08220215A JP H08220215 A JPH08220215 A JP H08220215A JP 7024928 A JP7024928 A JP 7024928A JP 2492895 A JP2492895 A JP 2492895A JP H08220215 A JPH08220215 A JP H08220215A
Authority
JP
Japan
Prior art keywords
frequency
signal
pulse
residual
pulse signal
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.)
Granted
Application number
JP7024928A
Other languages
Japanese (ja)
Other versions
JP2607861B2 (en
Inventor
Shunichi Kohama
俊一 小浜
Shunji Ozaki
俊二 尾崎
Susumu Mizota
享 溝田
Masato Yamashita
正人 山下
Shinji Yanai
伸治 屋内
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.)
Japan Steel Works Ltd
Oki Electric Industry Co Ltd
Technical Research and Development Institute of Japan Defence Agency
Original Assignee
Japan Steel Works Ltd
Oki Electric Industry Co Ltd
Technical Research and Development Institute of Japan Defence Agency
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 Japan Steel Works Ltd, Oki Electric Industry Co Ltd, Technical Research and Development Institute of Japan Defence Agency filed Critical Japan Steel Works Ltd
Priority to JP7024928A priority Critical patent/JP2607861B2/en
Publication of JPH08220215A publication Critical patent/JPH08220215A/en
Application granted granted Critical
Publication of JP2607861B2 publication Critical patent/JP2607861B2/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Landscapes

  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

PURPOSE: To accurately detect a target signal sound by removing a noise such as an impact noise and an organismic sound from a signal sound detected by a sonar or the like. CONSTITUTION: Frequency analysis is performed on an input signal with every constant time interval, and a peak frequency whose intensity becomes maximum is found, and a pulse forming unit 17 forms a pulse signal as a time continuance of peak frequencies. A pulse judging processing unit 19 judges whether a pulse signal is PCW or LFM on the basis of a time gradient or a change width of a liner function by approximating a time change in a frequency of the pulse signal by a linear function. An estimative residual calculating unit 20 calculates the square sum of a residual to a liner function of a measured value of the peak frequency. A residual comparator 22 judges the pulse signal as an unknown signal which is PCW and LFM or does not belong to PCW and LFM by comparing the square sum of the residual and an output signal of a reference value generator 21 with each other.

Description

【発明の詳細な説明】Detailed Description of the Invention

【0001】[0001]

【産業上の利用分野】本発明は、例えば、ソナー等にお
いて、入力信号から形成したパルス信号の変調形式を判
定する信号形式判定方法に関するものである。
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a signal format determination method for determining the modulation format of a pulse signal formed from an input signal in a sonar or the like.

【0002】[0002]

【従来の技術】従来、このような分野の技術としては、
例えば次のようなものがあった。ソナー等が目標から放
射されるパルス音等を検知し、パルスの変調形式を推定
する場合、入力信号を一定の時間間隔で周波数分析して
強度が極大となるピーク周波数を測定し、このピークの
時間的連なりとしてパルス信号を形成して、このパルス
信号内での周波数の変化の仕方からパルス信号の変調形
式を判定するという方法が取られている。図2は、従来
の信号形式判定方法を実施するための信号形式判定装置
の機能ブロック図である。この図を参照しつつ従来の信
号形式判定方法を説明する。この信号形式判定装置は、
目標から放射されるパルス音を周囲雑音と共に入力する
入力端子11を有している。入力端子11は、周波数分
析器12の入力側に接続されている。周波数分析器12
は、入力端子11から入力された信号を一定の時間間隔
で周波数分析する手段である。周波数分析器12の出力
側は、周波数分析器12で周波数分析した結果を各周波
数ビン毎に短時間の時定数で積分する短時間積分器13
の入力側に接続されると共に、周波数分析器12で周波
数分析した結果を各周波数ビン毎に長時間の時定数で積
分する長時間積分器14の入力側に接続されている。
2. Description of the Related Art Conventionally, techniques in such a field include:
For example, there was the following. When a sonar or the like detects a pulse sound or the like emitted from a target and estimates the modulation format of the pulse, it analyzes the frequency of the input signal at regular time intervals, measures the peak frequency at which the intensity becomes maximum, and measures the peak frequency. A method is used in which a pulse signal is formed as a time series, and the modulation format of the pulse signal is determined from the manner in which the frequency changes in the pulse signal. FIG. 2 is a functional block diagram of a signal format determining apparatus for implementing a conventional signal format determining method. A conventional signal format determination method will be described with reference to FIG. This signal format determination device
It has an input terminal 11 for inputting a pulse sound radiated from a target together with ambient noise. The input terminal 11 is connected to the input side of the frequency analyzer 12. Frequency analyzer 12
Is a means for frequency-analyzing the signal input from the input terminal 11 at fixed time intervals. An output side of the frequency analyzer 12 is a short-time integrator 13 for integrating a result of frequency analysis by the frequency analyzer 12 with a short time constant for each frequency bin.
, And is connected to the input side of a long time integrator 14 that integrates the result of frequency analysis by the frequency analyzer 12 with a long time constant for each frequency bin.

【0003】短時間積分器13及び長時間積分器14の
各出力側は、ピーク検出器15の入力側に接続されてい
る。ピーク検出器15は、短時間積分器13から送られ
た短時間積分結果及び長時間積分器14から送られた長
時間積分結果に基づいて強度が極大となるピークを検出
する手段である。ピーク検出器15の出力側は、入力端
子11から入力する入力信号を分析してピーク検出器1
5で検出されたピークに対応する正確な周波数を測定す
る周波数精測器16に接続されている。又、ピーク検出
器15の出力側は、ピーク検出器15から送られたピー
ク検出情報及び周波数精測器16から送られた周波数情
報により、同一パルスの成分と判断されるピークの時間
的連なりからパルスを形成するパルス形成器17の入力
側に接続されている。パルス形成器17の出力側は、パ
ルス形成器17で形成されたパルス信号の周波数変化の
仕方を線形回帰により推定する勾配推定器18の入力側
に接続されている。勾配推定器18の出力側は、勾配推
定器18で推定されたパルス信号の周波数変化の勾配値
と、予め設定された閾値とからパルス信号の変調形式を
判定するパルス判定処理器19の入力側に接続されてい
る。パルス判定処理器19の出力側は、パルス信号の各
種パラメータの出力端子20に接続されている。
The outputs of the short-time integrator 13 and the long-time integrator 14 are connected to the input of a peak detector 15. The peak detector 15 is a means for detecting a peak having a maximum intensity based on the short-time integration result sent from the short-time integrator 13 and the long-time integration result sent from the long-time integrator 14. The output side of the peak detector 15 analyzes the input signal input from the input terminal 11 and
5 is connected to a frequency measuring device 16 for measuring an accurate frequency corresponding to the peak detected. In addition, the output side of the peak detector 15 is based on the peak detection information sent from the peak detector 15 and the frequency information sent from the frequency precision measuring device 16, from the time series of peaks judged to be the same pulse component. It is connected to the input side of a pulse shaper 17 that forms a pulse. The output side of the pulse shaper 17 is connected to the input side of a gradient estimator 18 that estimates the manner of frequency change of the pulse signal formed by the pulse shaper 17 by linear regression. The output side of the gradient estimator 18 is connected to the input side of a pulse determination processor 19 that determines the modulation type of the pulse signal from the gradient value of the frequency change of the pulse signal estimated by the gradient estimator 18 and a preset threshold value. It is connected to the. The output side of the pulse determination processor 19 is connected to an output terminal 20 for various parameters of the pulse signal.

【0004】次に、この信号形式判定装置の動作を説明
する。入力端子11から入力された広帯域成分を含んだ
入力信号は、周波数分析器12で一定時間間隔dtでサ
ンプリングされ、周波数分析される。周波数分析の方法
としては、例えば、FFT(Fast Fourier Transform、
高速フーリエ変換)等の方法がある。即ち、入力端子1
1からの入力信号をx(t)とし、その離散形を次の
(1)式で表す。 x(ti ),ti =t0 +i・dt ・・・(1) 但し、 i=0,1,・・・ 一定時間間隔Δt毎に周波数分析結果を出力するものと
し、分析周期kにおける周波数サンプルfn の周波数分
析出力信号を次の(2)項のように表す。 X(fn ;k) ・・・(2) 但し、 n=0,1,・・・,N−1 周波数分析結果Y(fn ;k)は、実部と虚部との二乗
和の形にして次の(3)式で表す。 Y(fn ;k)=|X(fn ;k)|2 ・・・(3) この周波数分析結果Y(fn ;k)は、短時間積分器1
3及び長時間積分器14へ送られる。
[0004] Next, the operation of the signal format judging device will be described. An input signal including a wideband component input from the input terminal 11 is sampled at a predetermined time interval dt by the frequency analyzer 12 and subjected to frequency analysis. As a method of frequency analysis, for example, FFT (Fast Fourier Transform,
Fast Fourier Transform). That is, the input terminal 1
The input signal from 1 is x (t), and its discrete form is represented by the following equation (1). x (t i ), t i = t 0 + i · dt (1) where i = 0, 1,... It is assumed that a frequency analysis result is output at regular time intervals Δt, and The frequency analysis output signal of the frequency sample f n is represented as the following item (2). X (f n ; k) (2) where n = 0, 1,..., N−1 The frequency analysis result Y (f n ; k) is the sum of squares of the real part and the imaginary part. And expressed by the following equation (3). Y (f n ; k) = | X (f n ; k) | 2 (3) The frequency analysis result Y (f n ; k)
3 and sent to the long time integrator 14.

【0005】短時間積分器13では、周波数分析器12
から送られた周波数分析結果Y(fn ;k)を予め設定
された短時間の積分時定数τ1 で周波数ビン毎に積分
し、その結果をピーク検出器15に送る。一方、長時間
積分器14では、周波数分析器12から送られた周波数
分析結果Y(fn ;k)を予め設定された長時間の積分
時定数τ2 (但し、τ2 ≫τ1 )で周波数ビン毎に積分
し、その結果をピーク検出器15に送る。短時間積分器
13の積分時定数τ1 は、積分結果が入力信号の特徴を
失わないように短く設定され、長時間積分器14の積分
時定数τ2 は、ピーク検出器15のピーク検出における
背景の平均レベルを推定するために長く設定される。積
分は、有限の時間窓を用いてもよいし、瞬時値の影響が
指数関数的に減少する無限の時間窓を用いてもよい。例
えば、無限の時間窓を用いた場合、次の(4)式の形で
積分処理を実行する。 Y1 (fn ;k)=α・Y(fn ;k) +(1−α)・Y1 (fn ;k−1) Y2 (fn ;k)=β・Y(fn ;k) +(1−β)・Y2 (fn ;k−1)・・・(4) 但し、 Y1 ;短時間積分器3の出力信号 α;積分時定数τ1 から定められる定数 Y2 ;長時間積分器4の出力信号 β;積分時定数τ2 から定められる定数 ピーク検出器15は、短時間積分器13及び長時間積分
器14から各積分結果を入力し、次の(5)式に示すよ
うに両者の出力信号の比を予め定めた閾値γと比較す
る。 Y1 /Y2 ≧γ ・・・(5) この(5)式が成立し、かつ周波数軸上でピークになっ
ている場合には、ピーク検出器15は、この周波数近傍
にパルス信号の周波数成分が存在すると判断し、周波数
ビンの中心周波数fnp(k)又は周波数ビン番号n
(k)を周波数精測器16へ送り、又、周波数ビン番号
n(k)をパルス形成器17へ送る。
The short-time integrator 13 includes a frequency analyzer 12
Sent frequency analysis result Y from; integrated for each frequency bin in (f n k) of the short time is a preset integral time constant tau 1, and sends the result to the peak detector 15. On the other hand, the long-time integrator 14 converts the frequency analysis result Y (f n ; k) sent from the frequency analyzer 12 by a preset long-time integration time constant τ 2 (where τ 2 ≫τ 1 ). Integration is performed for each frequency bin, and the result is sent to the peak detector 15. Integration time constant tau 1 of short integrator 13, the integration result is set shorter so as not to lose the characteristics of the input signal, the integration time constant tau 2 long integrator 14, the peak detection of the peak detector 15 It is set long to estimate the average level of the background. The integration may use a finite time window or an infinite time window in which the influence of the instantaneous value decreases exponentially. For example, when an infinite time window is used, the integration processing is performed in the form of the following equation (4). Y 1 (f n; k) = α · Y (f n; k) + (1-α) · Y 1 (f n; k-1) Y 2 (f n; k) = β · Y (f n K) + (1−β) · Y 2 (f n ; k−1) (4) where Y 1 ; output signal α of short-time integrator 3; constant determined from integration time constant τ 1 Y 2 : output signal of long-time integrator 4 β: constant determined by integration time constant τ 2 Peak detector 15 receives each integration result from short-time integrator 13 and long-time integrator 14 and outputs the following ( As shown in equation (5), the ratio between the two output signals is compared with a predetermined threshold value γ. Y 1 / Y 2 ≧ γ (5) When the equation (5) is satisfied and the peak is on the frequency axis, the peak detector 15 determines the frequency of the pulse signal near this frequency. It is determined that a component exists, and the center frequency f np (k) of the frequency bin or the frequency bin number n
(K) is sent to the frequency analyzer 16, and the frequency bin number n (k) is sent to the pulse former 17.

【0006】周波数精測器16では、入力端子11から
入力した入力信号を詳細に周波数分析し、ピーク検出器
15から送られた中心周波数fnp(k)又は周波数ビン
番号n(k)に基づいてピーク周波数を精測し、その精
測値F(k)をパルス形成器17へ送る。ピーク周波数
の精測方法としては、例えば、周波数精測器16での周
波数分析結果の中から中心周波数fnp(k)に最も近い
周波数をもつ極大点を抽出し、この極大点及びその両隣
の周波数ビンの周波数分析結果から放物線補間により算
出する方法がある。パルス形成器17では、ピーク検出
器5からの周波数ビン番号n(k)及び周波数精測器1
6からの精測値F(k)を入力し、周波数ビン番号n
(k)に基づき、時間的かつ周波数的に連続したピーク
の時間的連なりを1つのパルス信号として形成する。こ
の方法としては、例えば、連続した分析周期に生じたピ
ークの周波数ビン番号n(k)と周波数ビン番号n(k
−1)とを比較し、その差が予め定めた基準値以下であ
れば同一パルスの要素と見做す方法がある。又、(k−
1)までの分析周期のデータから同一パルスと見做され
るピーク値の連なり(以下、トラックという)を形成
し、トラック内の周波数又は周波数ビンの時間的な変化
の仕方から分析周期kにおける精測値F(k)又は周波
数ビン番号n(k)を予測し、予測値と実際に得られた
値との比較により同一パルスに属するピークか否かを判
断する方法もある。
In the frequency precision measuring device 16, the frequency of the input signal input from the input terminal 11 is analyzed in detail, and based on the center frequency f np (k) or the frequency bin number n (k) sent from the peak detector 15. The peak frequency is precisely measured, and the measured value F (k) is sent to the pulse former 17. As a precise measurement method of the peak frequency, for example, a maximum point having a frequency closest to the center frequency f np (k) is extracted from the frequency analysis result by the frequency measurement device 16, and the maximum point and its neighboring neighbors are extracted. There is a method of calculating by parabolic interpolation from the frequency analysis result of the frequency bin. In the pulse shaper 17, the frequency bin number n (k) from the peak detector 5 and the frequency preciser 1
6 and input the exact measurement value F (k) from frequency bin number n
Based on (k), a temporal sequence of temporally and frequency-continuous peaks is formed as one pulse signal. As this method, for example, a frequency bin number n (k) and a frequency bin number n (k) of a peak generated in a continuous analysis cycle are used.
-1) is compared, and if the difference is less than or equal to a predetermined reference value, it can be regarded as an element of the same pulse. Also, (k-
A series of peak values regarded as the same pulse (hereinafter referred to as a track) is formed from the data of the analysis cycle up to 1), and the frequency in the track or the frequency bin in the analysis cycle k is determined based on the time variation of the frequency bin. There is also a method in which the measured value F (k) or the frequency bin number n (k) is predicted, and whether or not the peak belongs to the same pulse is determined by comparing the predicted value with an actually obtained value.

【0007】パルス形成器17は、同一パルスと見做せ
るピーク値の連なりを形成すると、その周波数精測値の
組F(k),k=m,・・・,m+L−1(m;ピーク
値の連なりの頭の分析周期の番号、L;ピーク値の連な
りの個数)を勾配推定器18へ出力する。勾配推定器1
8では、パルス内の周波数変化の仕方を次の(6)式で
表す。 F(t)=F0 +dF(t−tm ) ・・・(6) 但し、 tm ;ピークの連なりの頭の分析周期mの中心時刻 F0 ;初期周波数 dF;周波数勾配 上記(6)式で表したときの初期周波数F0 と周波数勾
配dFとを、パルス形成器17から送られた周波数精測
値F(k)の組から、次の(7)式及び(8)式により
推定する。
When the pulse former 17 forms a series of peak values that can be regarded as the same pulse, the set of frequency measurement values F (k), k = m,..., M + L-1 (m; peak) The number of the analysis cycle at the head of the series of values, L; the number of series of peak values) is output to the gradient estimator 18. Gradient estimator 1
In 8, the way of changing the frequency in the pulse is expressed by the following equation (6). F (t) = F 0 + dF (t−t m ) ... (6) where t m ; center time of analysis cycle m of peak continuous head F 0 ; initial frequency dF; frequency gradient (6) The initial frequency F 0 and the frequency gradient dF represented by the equation are estimated from the set of frequency accurate measurement values F (k) sent from the pulse former 17 by the following equations (7) and (8). To do.

【数1】 [Equation 1]

【数2】 但し、 k=m,・・・,m+L−1、 m;ピーク値の連なりの頭の分析周期の番号 L;ピーク値の連なりの個数 (7)式を解くと次の(9)式が得られる。 F0 =(BC−AD)/(LB−A2 ) dF=(LD−AC)/(LB−A2 ) ・・・(9) 尚、入力端子11からパルス形成器17までの動作がパ
ルス信号形成処理である。勾配推定器18は、初期周波
数F0 と周波数勾配dFとを算出すると、これらの値を
パルスの時間幅ΔT=(tm+L-1 −tm )とともにパル
ス判定処理器19へ出力する。パルス判定処理におい
て、パルス判定処理器19では、勾配推定器18から入
力した各パラメータを用いて周波数勾配dF又はパルス
信号内の周波数変化幅ΔF=dF・ΔTを予め定めた閾
値と比較し、パルス信号が一定周波数の正弦波(Pulsed
Continuous Wave、以下、PCWという)であるか或い
は周波数が時間と共に線形に変化する周波数変調波(Li
near Frequency Modulation 、以下、LFMという)で
あるかを判定する。
[Equation 2] However, k = m,..., M + L−1, m; the number of the analysis cycle at the head of the series of peak values L; the number of the series of peak values By solving equation (7), the following equation (9) is obtained. Can be F 0 = (BC-AD) / (LB-A 2 ) dF = (LD-AC) / (LB-A 2 ) (9) The operation from the input terminal 11 to the pulse former 17 is a pulse. This is a signal forming process. Gradient estimator 18, calculating the initial frequency F 0 and the frequency slope dF, outputs time width of these values pulse [Delta] T = with (t m + L-1 -t m) to the pulse determination processor 19. In the pulse determination process, the pulse determination processor 19 compares the frequency gradient dF or the frequency change width ΔF = dF · ΔT in the pulse signal with a predetermined threshold using each parameter input from the gradient estimator 18, and The signal is a constant frequency sine wave (Pulsed
A continuous wave (hereinafter referred to as PCW) or a frequency-modulated wave (Li) whose frequency changes linearly with time.
near frequency modulation, hereinafter referred to as LFM).

【0008】[0008]

【発明が解決しようとする課題】しかしながら、上記の
信号形式判定方法では、次のような問題点があった。即
ち、入力端子11からPCWやLFM以外の過渡音が入
力された場合にも、パルス判定処理器19は、それをP
CW又はLFMとして判定してしまうという問題点があ
った。特に、広い周波数帯域をもつ衝撃音や生物音等が
入力端子に入力された場合、ピーク周波数の時間的変化
が大きいので、周波数勾配dFやパルス内の周波数変化
幅ΔFが大きく推定されてLFMと見做されることがあ
り、誤った判断となることがある。本発明は、以上述べ
たパルス判定処理器19がPCWやLFM以外の信号を
PCWやLFMとして判定してしまうという問題を解決
するために、周波数勾配の推定値の信頼性に対する判断
機能を設け、広い周波数帯域をもつ過渡音をPCWやL
FM以外の不明信号として判定し、衝撃音や生物音等の
雑音による誤認を除去する信号形式判定方法を提供する
ものである。
However, the above-described signal format determination method has the following problems. That is, even when a transient sound other than PCW or LFM is input from the input terminal 11, the pulse determination processor 19 outputs
There is a problem in that it is determined as CW or LFM. In particular, when an impulsive sound or biological sound having a wide frequency band is input to the input terminal, the temporal change of the peak frequency is large, so that the frequency gradient dF and the frequency change width ΔF within the pulse are estimated to be large, and the LFM and the May be deemed to be incorrect and may lead to incorrect decisions. In order to solve the problem that the pulse determination processor 19 described above determines a signal other than PCW or LFM as PCW or LFM, the present invention provides a determination function for the reliability of the estimated value of the frequency gradient, PCW and L for transient sounds with a wide frequency band
It is intended to provide a signal format determination method for determining as an unknown signal other than FM and eliminating false recognition due to noise such as impact sound or biological sound.

【0009】[0009]

【課題を解決するための手段】本発明では、前記課題を
解決するために、入力信号を一定の時間間隔毎に周波数
分析して強度が極大となるピーク周波数を求め、かつ該
ピーク周波数の時間的連なりとしてパルス信号を形成す
るパルス信号形成処理と、前記パルス信号の周波数の時
間変化を1次関数で近似して該1次関数の時間勾配又は
変化幅に基づいて該パルス信号がPCWか或いはLFM
かを判定するパルス判定処理とを、施す信号形式判定方
法において、次のような処理を行っている。即ち、前記
ピーク周波数の測定値の前記1次関数に対する残差の二
乗和を算出する推定残差算出処理と、残差の二乗和と、
予め定めた基準値と前記パルス信号形成処理により検出
されたピークの連なりの個数とから計算される残差の二
乗和の許容値とを比較してパルス判定処理の判定結果を
採用するか或いは棄却するかを判定する残差比較処理と
を、行うようにしている。
According to the present invention, in order to solve the above-mentioned problems, an input signal is subjected to frequency analysis at regular time intervals to determine a peak frequency at which the intensity becomes maximum, and the time of the peak frequency is determined. A pulse signal forming process of forming a pulse signal as a continuum, approximating a time change of the frequency of the pulse signal by a linear function, and determining whether the pulse signal is PCW or LFM
The following processing is performed in the signal format determination method for performing the pulse determination processing for determining whether or not. That is, an estimated residual calculation process for calculating a sum of squares of residuals with respect to the linear function of the measured value of the peak frequency, and a sum of squares of residuals,
The determination result of the pulse determination process is adopted or rejected by comparing a predetermined reference value with the allowable value of the sum of squares of the residual calculated from the number of consecutive peaks detected by the pulse signal formation process. And a residual comparison process for determining whether to perform the comparison.

【0010】[0010]

【作用】本発明によれば、以上のように信号形式判定方
法を構成したので、パルス信号形成処理において、入力
信号を一定の時間間隔毎に周波数分析して強度が極大と
なるピーク周波数を求め、かつ該ピーク周波数の時間的
変化を表すパルス信号を形成する。次に、パルス判定処
理において、前記パルス信号の周波数の時間的変化を1
次関数で近似して該1次関数の時間勾配又は変化幅に基
づいて該パルス信号がPCWか或いはLFMかを判定す
る。推定残差算出処理において、前記ピーク周波数の測
定値の前記1次関数に対する残差の二乗和を算出する。
更に、残差比較処理において、前記残差の二乗和と、予
め定めた基準値と前記パルス信号形成処理により検出さ
れたピークの連なりの個数とから計算される残差の二乗
和の許容値とを比較してパルス判定処理の判定結果の信
頼性について判定し、前記パルス判定処理の判定結果に
基づいてパルス信号がPCWやLFMであると見做すか
或いはPCWやLFMに属さない不明の信号と見做すか
を判断するようにしている。従って、前記課題を解決で
きるのである。
According to the present invention, since the signal type determination method is configured as described above, in the pulse signal forming process, the frequency of the input signal is analyzed at regular time intervals to determine the peak frequency at which the intensity becomes maximum. And a pulse signal representing the temporal change of the peak frequency is formed. Next, in the pulse determination process, the temporal change of the frequency of the pulse signal is set to 1
It is approximated by a linear function to determine whether the pulse signal is PCW or LFM based on the time gradient or the variation width of the linear function. In the estimation residual calculation process, a sum of squares of the residual with respect to the linear function of the measured value of the peak frequency is calculated.
Furthermore, in the residual comparison process, the sum of squares of the residual, and an allowable value of the sum of squares of the residual calculated from the predetermined reference value and the number of peaks detected by the pulse signal forming process. To determine the reliability of the determination result of the pulse determination processing, and based on the determination result of the pulse determination processing, the pulse signal is considered to be PCW or LFM, or an unknown signal not belonging to PCW or LFM. I try to judge whether to consider it. Therefore, the above problem can be solved.

【0011】[0011]

【実施例】先ず、本実施例の基礎となっている原理を説
明する。パルス信号の周波数の時間的変化についての観
測値F(tk )(但し、k=m,・・・,m+L−1)
を(6)式で近似した場合、推定残差の二乗和Gは次の
(10)式で表される。
First, the principle underlying the present embodiment will be described. Observations for the temporal change in the frequency of the pulse signal F (t k) (where, k = m, ···, m + L-1)
Is approximated by the equation (6), the sum of squares G of the estimated residual is expressed by the following equation (10).

【数3】 (10)式中の観測値F(tk )の二乗和Hは次の(1
1)式で表される。
(Equation 3) The sum of squares H of the observed value F (t k ) in the equation (10) is given by the following (1)
It is expressed by the equation 1).

【数4】 パルス信号がPCWかLFMであったとすると、パルス
信号内での周波数の時間的変化は、従来の(6)式で表
わされ、(10)式で表される推定残差の二乗和Gは非
常に小さくなる。ところが、入力信号が広帯域の衝撃音
や生物音の場合、ピーク周波数は時間に対して線形には
変化せず、(10)式で表される推定残差の二乗和Gは
大きい値をもつ。
[Equation 4] Assuming that the pulse signal is PCW or LFM, the temporal change of the frequency in the pulse signal is expressed by the conventional equation (6), and the sum of squares G of the estimated residual expressed by the equation (10) is Very small. However, when the input signal is a wideband impact sound or biological sound, the peak frequency does not change linearly with time, and the sum of squares G of the estimated residual represented by the equation (10) has a large value.

【0012】ここで、パルス信号がPCWかLFMであ
る場合、(10)式の推定残差の二乗和Gの期待値E
(G)は、次の(12)式で表される。 E(G)=L・σf 2 ・・・(12) 但し、 L;パルス信号に属するピークデータの個数 σf 2 ;各ピークの周波数推定値の誤差分散 このσf 2 の値は、ピーク検出の閾値γ、周波数分析器
12での周波数分析幅、周波数精測器での周波数分析等
により変化する。従って、これらの値を基にピーク周波
数推定誤差分散の許容値σmax 2 を定め、(10)式の
推定残差の二乗和Gの許容値Gmax を次の(13)式で
表す。 Gmax =L・σmax 2 ・・・(13) 更に、次の(14)式に示すように、推定残差の二乗和
Gが許容値Gmax よりも大きくなった場合にパルス信号
の変調形式を不明とすれば、入力信号中の広帯域の衝撃
音や生物音を誤ってPCWやLFMと判定する可能性が
大幅に低減する。 G>Gmax ・・・(14) 図1は、本発明の実施例の信号形式判定方法を実施する
ための信号形式判定装置の機能ブロック図であり、従来
の図2と共通の要素には共通の符号が付されている。
Here, when the pulse signal is PCW or LFM, the expected value E of the sum of squares G of the estimated residual in the equation (10)
(G) is expressed by the following equation (12). E (G) = L · σ f 2 ··· (12) where, L; peak number of data sigma f 2 belonging to the pulse signal; error variance value of the sigma f 2 of the frequency estimates for each peak, peak It changes depending on the detection threshold γ, the frequency analysis width in the frequency analyzer 12, the frequency analysis in the frequency analyzer, and the like. Accordingly, the allowable value σ max 2 of the peak frequency estimation error variance is determined based on these values, and the allowable value G max of the sum of squares G of the estimated residual in Expression (10) is expressed by the following Expression (13). G max = L · σ max 2 (13) Further, as shown in the following equation (14), when the sum of squares G of the estimated residual becomes larger than the allowable value G max , the modulation of the pulse signal is performed. If the format is unknown, the possibility of erroneously determining a broadband impact sound or biological sound in the input signal as PCW or LFM is greatly reduced. G> G max (14) FIG. 1 is a functional block diagram of a signal format judging device for implementing the signal format judging method according to the embodiment of the present invention. Common symbols are assigned.

【0013】この信号形式判定装置では、従来の図2中
の勾配推定器18の出力側に、パルス形成器17の出力
信号に基づいて勾配推定の推定残差の二乗和を計算する
推定残差算出器20が接続されている。又、パルス形成
器17の出力側には、パルス形成器17の出力信号に基
づいて残差判定の基準値を生成する基準値生成器21が
接続されている。基準値生成器21の出力側には、推定
残差算出器20から送られた推定残差の二乗和と基準値
生成器21から送られた基準値とを比較する残差比較器
22が接続されている。残差比較器22の出力側には、
パルスの各種パラメータの推定結果を出力する出力端子
23が接続されている。尚、この信号形式判定装置中の
各構成要素2〜22は、個別回路で構成されるか或いは
コンピュータを用いたプログラムで実現される。
In this signal type determination apparatus, an estimated residual for calculating the sum of squares of the estimated residual of the gradient estimation based on the output signal of the pulse former 17 is provided at the output side of the conventional gradient estimator 18 in FIG. The calculator 20 is connected. Further, a reference value generator 21 for generating a reference value for residual determination based on the output signal of the pulse former 17 is connected to the output side of the pulse former 17. The output side of the reference value generator 21 is connected with a residual comparator 22 for comparing the sum of squares of the estimated residuals sent from the estimated residual calculator 20 and the reference value sent from the reference value generator 21. Has been done. On the output side of the residual comparator 22,
An output terminal 23 for outputting estimation results of various parameters of the pulse is connected. Each of the components 2 to 22 in this signal format determination device is configured by an individual circuit or realized by a program using a computer.

【0014】次に、図1の動作を説明する。この信号形
式判定装置では、入力端子11から周波数精測器16ま
での動作は従来と同一であるが、パルス形成器17以降
の動作が異なっている。即ち、パルス信号形成処理にお
いて、パルス形成器17は、同一パルスと見做せるピー
ク値の連なりを形成すると、ピーク値の連なりの個数
L、分析周期kの中心時刻tk 、及び周波数精測値の組
F(k),k=m,・・・,m+L−1(m;ピーク値
の連なりの頭の分析周期の番号、L;ピーク値の連なり
の個数)を勾配推定器18及び推定残差算出器20へ出
力すると共に、ピーク値の連なりの個数Lを基準値生成
器21へ出力する。勾配推定器18では、パルス信号内
の周波数変化の仕方を(6)式で表し、パルス形成器1
7からの精測値F(k)から、(7)式及び(8)式に
より初期周波数F0 と周波数勾配dFとを推定する。勾
配推定器18は、初期周波数F0 と周波数勾配dFとを
算出すると、これらの値をパルス信号の時間幅ΔT=
(tm+L-1 −tm )とともにパルス判定処理器19へ出
力する。又、初期周波数F0 、周波数勾配dF、及び
(8)式中のCを推定残差算出器20へ出力する。パル
ス判定処理において、パルス判定処理器19では、勾配
推定器18から入力した初期周波数F0 、周波数勾配d
F、及びパルス信号の時間幅ΔTの各パラメータを用い
て周波数勾配dF又はパルス内の周波数変化幅ΔF=d
F・ΔTを予め定めた閾値と比較し、パルス信号がPC
WであるかLFMであるかを判定する。この判定結果
は、初期周波数F0 及び周波数勾配dFの推定結果と共
に残差比較器22へ出力する。
Next, the operation of FIG. 1 will be described. In this signal format judging device, the operation from the input terminal 11 to the frequency measuring device 16 is the same as the conventional one, but the operation after the pulse former 17 is different. That is, in the pulse signal forming process, when the pulse forming unit 17 forms a series of peak values regarded as the same pulse, the pulse forming unit 17 determines the number L of the series of peak values, the center time tk of the analysis cycle k , and the frequency measurement value F (k), k = m,..., M + L−1 (m: number of the analysis cycle at the head of the series of peak values, L: number of the series of peak values) In addition to outputting to the difference calculator 20, the number L of the series of peak values is output to the reference value generator 21. In the gradient estimator 18, the manner in which the frequency in the pulse signal changes is expressed by equation (6).
The initial frequency F 0 and the frequency gradient dF are estimated from the precise measurement value F (k) obtained from the equation (7) and the equations (7) and (8). After calculating the initial frequency F 0 and the frequency gradient dF, the gradient estimator 18 calculates these values as the pulse signal time width ΔT =
Output to (t m + L-1 -t m) with pulse determination processor 19. Further, it outputs the initial frequency F 0 , the frequency gradient dF, and C in the equation (8) to the estimated residual calculator 20. In the pulse determination process, the pulse determination processor 19 calculates the initial frequency F 0 and the frequency gradient d input from the gradient estimator 18.
F and each parameter of the time width ΔT of the pulse signal, the frequency gradient dF or the frequency change width ΔF = d in the pulse is used.
F · ΔT is compared with a predetermined threshold.
It is determined whether it is W or LFM. This determination result is output to the residual comparator 22 together with the estimation result of the initial frequency F 0 and the frequency gradient dF.

【0015】推定残差算出処理において、推定残差算出
器20では、パルス形成器17からピーク値の連なりの
個数L及び精測値F(k)(但し、k=m,・・・,m
+L−1)を入力し、(11)式により観測値F
(tk )の二乗和Hを計算し、この観測値F(tk )の
二乗和Hと勾配推定器18から入力した初期周波数
0 、周波数勾配dF、及び(8)式中のCとを用い
て、(10)式により推定残差の二乗和Gを算出して残
差比較器22へ出力する。基準値生成器21では、パル
ス形成器17からピーク値の連なりの個数Lを入力し、
予め設定されている周波数推定誤差分散の許容値σmax
2 から、(13)式により推定残差の二乗和Gの許容値
max を算出して残差比較器22へ出力する。残差比較
処理において、残差比較器22では、推定残差算出器2
0からの推定残差の二乗和Gと基準値生成器21からの
許容値Gmax とを比較し、次の判定基準(a)及び
(b)に基づいてパルス信号の形式を判定し、中心周波
数や勾配等の推定結果と共に出力端子23から出力す
る。 (a)G>Gmax のとき;パルス信号の形式を不明とす
る。 (b)G≦Gmax のとき;パルス信号の形式をパルス判
定処理器19で判定したとおりとする。
In the estimation residual calculation process, the estimation residual calculator 20 outputs the number L of the series of peak values and the measured value F (k) (where k = m,..., M
+ L-1), and the observation value F is calculated by the equation (11).
The sum of squares H of (t k ) is calculated, the sum of squares H of the observed value F (t k ), the initial frequency F 0 input from the gradient estimator 18, the frequency gradient dF, and C in equation (8) Is used to calculate the sum of squares G of the estimated residual according to the equation (10), and outputs it to the residual comparator 22. In the reference value generator 21, the number L of the series of peak values is input from the pulse former 17,
The preset allowable value σ max of the frequency estimation error variance
2, (13) to calculate the allowable value G max of square sum G of estimation error and outputs it to the residual comparator 22 by formula. In the residual comparison process, the residual comparator 22 uses the estimated residual calculator 2
The sum of squares G of the estimated residual from 0 is compared with the allowable value Gmax from the reference value generator 21 to determine the form of the pulse signal based on the following criteria (a) and (b). The data is output from the output terminal 23 together with the estimation result such as the frequency and the gradient. (A) When G> Gmax : The format of the pulse signal is unknown. (B) When G ≦ G max : The format of the pulse signal is as determined by the pulse determination processor 19.

【0016】以上のように、本実施例では、パルス信号
内での周波数の時間変化特性の観測値F(tk )に対し
て推定残差の二乗和Gを求め、推定残差の二乗和Gが許
容値Gmax を越えた場合にパルス信号の信号形式を不明
としてPCWやLFMと区別するようにしたので、パル
ス判定処理器19がPCWやLFM以外の信号をPCW
やLFMとして判定して入力信号中の衝撃音や生物音等
の雑音がPCWやLFM等の信号音に統合される可能性
が大幅に低減され、目標の信号音を正しく分析する追尾
システムを構成できる。尚、本発明は上記実施例に限定
されず、例えば、目標から放射されたレーダ信号を受信
するシステム等にも応用が可能である。
As described above, in this embodiment, the sum of squares G of the estimated residual is obtained for the observed value F (t k ) of the time change characteristic of the frequency in the pulse signal, and the sum of squares of the estimated residual is obtained. When G exceeds the allowable value Gmax , the signal format of the pulse signal is determined to be unknown and is distinguished from PCW or LFM, so that the pulse determination processor 19 converts the signal other than PCW or LFM into PCW or LFM.
And the possibility of noise such as impact sounds and biological sounds in the input signal being integrated into signal sounds such as PCW and LFM is greatly reduced, and a tracking system that correctly analyzes the target signal sound is constructed. it can. Note that the present invention is not limited to the above-described embodiment, and can be applied to, for example, a system that receives a radar signal radiated from a target.

【0017】[0017]

【発明の効果】以上詳細に説明したように、本発明によ
れば、パルス信号内での周波数の時間的変化の推定値の
信頼性に対する判断機能を設けて周波数の時間変化特性
の観測値に対して推定残差の二乗和を求め、推定残差の
二乗和が許容値を越えた場合に信号特性を不明としてP
CWやLFMと区別するようにしたので、入力信号中の
衝撃音や生物音等の雑音がPCWやLFM等の信号音に
統合される可能性が大幅に低減され、目標の信号音を正
しく分析する追尾システムを構成できる。
As described in detail above, according to the present invention, a judgment function for the reliability of the estimated value of the temporal change of the frequency in the pulse signal is provided to obtain the observed value of the frequency change characteristic of the frequency. On the other hand, the sum of squares of the estimated residuals is calculated, and if the sum of squares of the estimated residuals exceeds the allowable value, the signal characteristic is regarded as unknown and P
Since it is distinguished from CW and LFM, the possibility that noise such as impact sound and biological sound in the input signal is integrated into the signal sound such as PCW and LFM is greatly reduced, and the target signal sound is correctly analyzed. A tracking system can be configured.

【図面の簡単な説明】[Brief description of drawings]

【図1】本発明の実施例を示す信号形式判定方法を実施
するための信号形式判定装置の機能ブロック図である。
FIG. 1 is a functional block diagram of a signal format determination device for implementing a signal format determination method according to an embodiment of the present invention.

【図2】従来の信号形式判定方法を実施するための信号
形式判定装置の機能ブロック図である。
FIG. 2 is a functional block diagram of a signal format determination device for implementing a conventional signal format determination method.

【符号の説明】[Explanation of symbols]

11 入力端
子 12 周波数
分析器 13 短時間
積分器 14 長時間
積分器 15 ピーク
検出器 16 周波数
精測器 17 パルス
形成器 18 勾配推
定器 19 パルス
判定処理器 20 推定残
差算出器 21 基準値
生成器 22 残差比
較器 23 出力端
DESCRIPTION OF SYMBOLS 11 Input terminal 12 Frequency analyzer 13 Short-time integrator 14 Long-time integrator 15 Peak detector 16 Frequency preciser 17 Pulse former 18 Gradient estimator 19 Pulse judgment processor 20 Estimation residual calculator 21 Reference value generator 22 residual comparator 23 output terminal

───────────────────────────────────────────────────── フロントページの続き (72)発明者 溝田 享 東京都港区虎ノ門1丁目7番12号 沖電気 工業株式会社内 (72)発明者 山下 正人 東京都港区虎ノ門1丁目7番12号 沖電気 工業株式会社内 (72)発明者 屋内 伸治 東京都港区虎ノ門1丁目7番12号 沖電気 工業株式会社内 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor, Ryo Mizoda 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Masato Yamashita 1-7-12 Toranomon, Minato-ku, Tokyo Oki Inside the Electric Industry Co., Ltd. (72) Inventor Shinji 1-7-12 Toranomon, Minato-ku, Tokyo Inside Oki Electric Industry Co., Ltd.

Claims (1)

【特許請求の範囲】[Claims] 【請求項1】 入力信号を一定の時間間隔毎に周波数分
析して強度が極大となるピーク周波数を求め、かつ該ピ
ーク周波数の時間的連なりとしてパルス信号を形成する
パルス信号形成処理と、 前記パルス信号の周波数の時間変化を1次関数で近似し
て該1次関数の時間勾配又は周波数変化幅に基づいて該
パルス信号が一定周波数の正弦波か或いは周波数が時間
と共に線形に変化する周波数変調波かを判定するパルス
判定処理とを、 施す信号形式判定方法において、 前記ピーク周波数の測定値の前記1次関数に対する残差
の二乗和を算出する推定残差算出処理と、 前記残差の二乗和と、予め定めた基準値と前記パルス信
号形成処理により検出されたピークの連なりの個数とか
ら計算される残差の二乗和の許容値とを比較して前記パ
ルス判定処理の判定結果を採用するか或いは棄却するか
を判定する残差比較処理とを、 行うことを特徴とする信号形式判定方法。
1. A pulse signal forming process for frequency-analyzing an input signal at regular time intervals to obtain a peak frequency at which the intensity is maximum, and forming a pulse signal as a time series of the peak frequency; A temporal change in the frequency of a signal is approximated by a linear function, and the pulse signal is a sine wave having a constant frequency or a frequency-modulated wave whose frequency changes linearly with time based on a time gradient or a frequency change width of the linear function. A pulse determination process for determining whether the residual is the sum of squares of a residual of the measured value of the peak frequency with respect to the linear function; and a sum of squares of the residual. And a permissible value of the sum of squares of the residual calculated from a predetermined reference value and the number of consecutive peaks detected by the pulse signal forming process. The determination result and residual comparison determines whether or reject adopted, signal format determination method, which comprises carrying out.
JP7024928A 1995-02-14 1995-02-14 Signal format judgment method Expired - Lifetime JP2607861B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP7024928A JP2607861B2 (en) 1995-02-14 1995-02-14 Signal format judgment method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP7024928A JP2607861B2 (en) 1995-02-14 1995-02-14 Signal format judgment method

Publications (2)

Publication Number Publication Date
JPH08220215A true JPH08220215A (en) 1996-08-30
JP2607861B2 JP2607861B2 (en) 1997-05-07

Family

ID=12151773

Family Applications (1)

Application Number Title Priority Date Filing Date
JP7024928A Expired - Lifetime JP2607861B2 (en) 1995-02-14 1995-02-14 Signal format judgment method

Country Status (1)

Country Link
JP (1) JP2607861B2 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009512857A (en) * 2005-10-21 2009-03-26 レイセオン カンパニー Sonar system and method with low probability of affecting marine mammals
JP2014032080A (en) * 2012-08-02 2014-02-20 Nec Corp Active sonar apparatus, active sonar signal processing method, and signal processing program thereof

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009512857A (en) * 2005-10-21 2009-03-26 レイセオン カンパニー Sonar system and method with low probability of affecting marine mammals
JP2014032080A (en) * 2012-08-02 2014-02-20 Nec Corp Active sonar apparatus, active sonar signal processing method, and signal processing program thereof

Also Published As

Publication number Publication date
JP2607861B2 (en) 1997-05-07

Similar Documents

Publication Publication Date Title
US7567204B2 (en) Method for determining noise floor level and radar using the same
JP3788322B2 (en) Radar
US20060195279A1 (en) Method and system for the detection and/or removal of sinusoidal interference signals in a noise signal
EP3665494B1 (en) System and method for interference detection in a rf receiver
Abratkiewicz et al. Radar signal parameters estimation using phase accelerogram in the time-frequency domain
AU2019229361B2 (en) Input count rate estimation in radiation pulse detectors
JP2607861B2 (en) Signal format judgment method
EP1697721B1 (en) Method and apparatus for obtaining information about the size distribution of macroscopic particles in a liquid
JP2005031083A (en) Method and device for finding delay between signals, and closed-loop feedback system using the method
US11656166B2 (en) Method and apparatus for detecting particles
JP2970950B2 (en) Inter-vehicle distance measuring device having threshold value determining means
RU2685048C1 (en) Method of determining places of inhomogeneities and damage of power transmission lines
RU2331893C1 (en) Method of discrete component separation in signal spectre and device for its implementation
EP3640672B1 (en) Systems and methods for distance independent acoustical differential signature detection
KR102036790B1 (en) Condition Assessment System and Method for Cables via Stepped-Frequency Waveform Reflectometry
KR20180135721A (en) FMCW radar interference cancellation algorithm for automotive vehicle
JP3394412B2 (en) Pulse sound detection method and apparatus
JP2607854B2 (en) Signal tracking method
JP2002311131A (en) Object detector for vehicle
JP3075272B2 (en) Automatic target tracking method and device
JP2004069467A (en) Object detector for vehicle
CN111521978B (en) Radar signal target angle detection method and device, computer equipment and medium
JP2866292B2 (en) Signal detection method
RU2787743C2 (en) Count rate assessment at input in radiation pulse detectors
JPH07128349A (en) Speed detector

Legal Events

Date Code Title Description
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 19961224

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

R250 Receipt of annual fees

Free format text: JAPANESE INTERMEDIATE CODE: R250

S533 Written request for registration of change of name

Free format text: JAPANESE INTERMEDIATE CODE: R313533

R350 Written notification of registration of transfer

Free format text: JAPANESE INTERMEDIATE CODE: R350

EXPY Cancellation because of completion of term