JPS5821584A - Specific signal detection system - Google Patents

Abstract

PURPOSE:To achieve earier detection of a specific signal by determing the number of observed data whose computed values are above set level once the observed data sampling the detection outputs are memorized retroactively by a period corresponding to preset estimation signal data. CONSTITUTION:A data convertion circuit 1 converts an analog voltage into a specified signal level receiving an output Z(t) of an detector at the input thereof and derives obserbed data Zi sampling it at a fixed time interval so that they will be applied to an observed data memory circuit 2. The memory circuit 2 memorizes the observed data in the order of the sampling while an estimation signal data memory circuit 3 memorizes estimation signal data Si by one cycle. A computing circuit 4 receives and computes signals from the memory circuit 2 and 3 each time the sampling is done and the results are stored into a data meory circuit 5. The data of the memory circuit 5 are each compared with the set level V from a level setting device 7 by means of a comparator circuit 6. The number of the data is counted as larger than the set level and when the count exceeds a fixed value, a discrimination output is provided to indicate the presence of a specific signal.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for early detection of a specific signal buried in noise. This invention is used, for example, when a magnetic body passes through the earth's magnetism, and the magnetic change is detected as a specific signal.

Various methods for detecting a specific signal buried in noise have been proposed and employed in the past. One of them, the Subek)/I/ analysis method, stores the measured output for a certain period of time, uses all of that data to find the Subek)/l/ density, and the spectrum μ of a specific signal is present in it. Since the method detects whether or not a specific signal is detected, a considerable time delay occurs from the start of measurement to the detection of a specific signal.

Also, in signal detection methods that use statistical hypothesis testing methods.

As observed data Z (t) and estimated signal S (t),
9 time t when the integral value exceeds the set level.

It is determined that there is a specific signal from the object to be detected within the time width from time to+T, and the first determination is often delayed by at least T time. (T, period of estimated signal) As in the conventional method described above, if there is a time delay in detecting a specific signal, it is inconvenient when it is necessary to immediately take various blind actions in response to the signal at the time of signal detection. It is.

SUMMARY OF THE INVENTION Therefore, it is an object of the present invention to provide a specific signal detection method capable of eliminating the drawbacks of the conventional methods and detecting a specific signal as early as possible with as little time delay as possible.

The present invention will be described in detail below with reference to one drawing.

FIG. 1 is a characteristic diagram showing an example of observation data Z(t). In the figure, it is assumed that signal 5(t) is present between times 4 and 6. Observation data 2(1) shown in Fig. 1 is set to 0.
When sampling is performed at one time interval and the integral calculation of equation (1) is performed during one period of the signal S (t), a characteristic curve as shown in FIG. 2 is obtained. In FIG. 2, for example, L(t, 2.0) is from time 2.0 to time t(=4.0
) shows the product of observed data according to equation (1),
The characteristic curve shown by the solid line shows the integral process in the middle. If the level of this characteristic curve appears negative, it means that the observed data Z (i) is noise, and if the characteristic curve shows a positive level transition.

Indicates that it contains a specific signal.

In the conventional method, as mentioned above, a certain sump/L'l
For example, at time t = 4.0 at time I-, t = 4.0
signal 5(i)olil1 minute before [z(t)
;2. The first equation is calculated using O≦i≦4.01, and the calculated result value (indicated by ◎) is compared with the set level to determine the presence or absence of a signal. Similar determinations are made at the next point by shifting them in parallel for each sampling time. In this way, since the presence or absence of a signal is determined based on one calculation result at each time point, in the case of the signal S (t) shown in FIG.
As shown by the mark ◎ on the curve o), accurate determination cannot be made unless t=6°0, which results in a time delay of about 2.0 from the signal generation.

The specific signal detection method of the present invention utilizes not only the calculation result value according to equation (1) but also the integration process in the intermediate stage of obtaining the calculation result, that is, each curve shown in FIG. 2, for signal detection. be. For example, at time t = 4.5, each part of the integral process [L(t, 2.5) to L(t.

4.4) +t=ts) If the nd point exceeds the set level among the 20 points, it is determined that there is a signal. Note that in the above, L(t, ti) means.

FIG. 3 is a circuit block diagram showing one embodiment of the present invention.

In Fig. 3, 1 is a data conversion circuit which receives the detector output 2α) as an input, converts the analog voltage to a predetermined signal level, derives observation data Zi while sampling at a constant time width, and outputs the observation data Zi. Added to data storage circuit 2.

The observation data storage circuit 2 sequentially samples the added observation data into storage areas Z i-n, Zi-n+1,
・・・・・・Memorized in Zi. These two observation data storage circuits are equipped with a storage area that stores observation data for at least one period of the estimated signal (n samples), and each sampling discards the data from the oldest point in time and stores the new data. Capture and memorize.

Reference numeral 6 denotes an estimated signal data storage circuit, which also includes a storage area for storing one cycle of estimated signal data. s2...Sn is memorized.

4 is an arithmetic circuit which receives signals from the observation data storage circuit 2 and the estimated signal data storage circuit 3;

The calculation shown in FIG. 4 is repeatedly performed at each sampling time.

5 is the result L(i, 1-j) calculated by the calculation circuit 4
This is a calculation result data storage circuit that stores . The calculation result data storage circuit 5 stores an integral value L(i, 1-n) from time ti-n to time ti, and an integral value L(i, 1-n) from time ti-n+1 to time j.
Integral process value L(i, i-n+1) +... up to i
9 Integral process value L(i,
i-2).

Integral process value L(i.

1-1) is stored. That is, as shown in Figure 4, 1
Each integral process value from the start time to time ti, which is shifted by one sampling time from time ti-n to time ti, is stored.

The calculation in the calculation circuit 4 at time 9, i.e., time i, is performed according to the following equation.

L(i, 1-j)=L(i-1,i-j+1)+5j
(Zi-Zi -1)-TSj2Δt
・・・・・・(3) j=1.2. ......n The first term L (i-1, i-j+1) of this equation is at time 2
This is the calculation result at time ti-1, that is, one sampling time before. Further, the second term can be calculated using the observation data Zi sampled at time point i and the estimated signal data Sj, and the third term can also be calculated from the estimated signal data. Therefore, the calculation result L at time i
(Li-j) is the calculation result at time i-i at time i-
It can be easily calculated by appropriately adding or subtracting the amount at time it from j.

In the calculation result data storage circuit 5. Calculation result L (i, 1-n) calculated using data for one cycle of the estimated signal
is erased during the first sampling.

A new calculation result for the next L(i, 1-1) is stored.

Each data stored in the calculation result data storage circuit 5 is sent to a comparison circuit 6, and is compared with each calculation result data.

It is compared with the set level A/V of the level setter 7.

Then, the comparator circuit 6 sets the calculation result data value to the set level/l.
/V is counted, and when the counted value reaches a point in time when it is greater than nd, a determination output indicating the presence of a specific signal is derived. Here, the value of nd may be determined in consideration of how much false detection rate is allowed.

As described above, according to the specific signal detection method of the present invention, the time integral of the product of the time differential of observed data and the estimated signal data and the square of the estimated signal data are expressed by shifting the beginning of the observed data by one data. The calculation is performed individually up to the time, and the number of calculation values obtained in relation to the difference between the two integral values using the same leading data is found to be equal to or higher than a preset level.

When this number is greater than or equal to a predetermined number, it is determined that a specific signal is present, so that the specific signal can be detected early in real time without taking approximately one period of the estimated signal from signal generation. Furthermore, since a plurality of calculated values are used for detection, the false detection rate is also reduced compared to conventional methods.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of observation data, FIG. 2 is a diagram showing the result of integral calculation of observation data, FIG. 6 is a circuit block diagram showing an embodiment of the present invention, and FIG. FIG. 3 is a diagram illustrating the concept of arithmetic processing for explaining the operation of the embodiment. 1; Data conversion circuit + 2N observation data storage circuit,
3; Estimated signal data storage circuit; 4; Arithmetic circuit; 5;
Operation result data storage circuit. 6; Comparison circuit; 7; Level setting device. Patent applicant Shimadzu Corporation Representative Patent attorney Shigeru Nakamura

Claims (1)

[Claims] A method for detecting a specific signal buried in noise, using observation data obtained by sampling the detection output. The estimated signal data set in advance according to the detected object is stored retroactively from the period boundary time corresponding to the detected object, and the product of the time differential of this observed data and the estimated signal data, and the estimated signal Each time integral of the square of the data is individually performed by shifting the beginning of the observed data by one data until the current time, and each calculated value obtained in relation to the difference between the two integral values using the same beginning data is set in advance. Find the number that is higher than the specified level, and when this number is higher than the specified number,
A specific signal detection method characterized by determining that a specific signal is present.