JPS58181336A - Signal detector - Google Patents

Abstract

PURPOSE:To suppress impulse and strong noise generated suddenly during the processing of addition and averaging, by using a prestage addition average value obtained through the addition and averaging at each number of times of basic units obtained by subdividing the total number of adding and averaging processing. CONSTITUTION:A prestage adding and averaging device 5 subdivides the total number of addition and averaging of signals A/D-converted at an A/D converter 2 as basic units, and processes the addition and averaging at each number of times of the basic units. This output is distributed into plural storage devices 8 by a distributor 7. A characteristic quantity extracting device 9 extracts characteristics from the prestage addition and average value at the number of times of each basic unit stored in the device 8. A similarity discriminator 10 compares the similarity of characteristics from an output of the device 9. An adding and averaging device 11 eliminates the prestage addition and averaging value of inferior similarity and processes only the remaining prestage addition and averaging values.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a signal detection device that performs signal detection using averaging processing, and in particular, when there is pulsed noise that suddenly occurs in the noise, it is possible to suppress interference caused by this pulsed noise. The purpose is to provide a signal detection device.

Generally, when noise is superimposed on a periodically changing signal, averaging processing is used to detect signal components from the signal.

This averaging process is as shown in FIG.

A processing method in which periodically changing signals (a) are added over time in synchronization with the period (b) to obtain an added signal (C), and divided by the number of additions to obtain an average value (d). It is. It is well known that by performing the averaging process, the signal-to-noise ratio is improved in proportion to the square root of the number of times the averaging process is performed, compared to when the averaging process is not performed.

However, the averaging process described above is an effective method for white noise by appropriately selecting the number of averaging operations, but pulse noise that occurs suddenly has a larger energy than white noise. Due to the high noise level, the effectiveness against this is extremely low. Therefore, in a conventional signal detection device using averaging processing, even if the number of averaging operations is sufficient to suppress white noise, pulse noise energy remains, making it difficult to suppress it. In addition, in order to sufficiently suppress pulse noise, it is necessary to disperse residual energy.1. This method requires a considerable number of averaging operations, and has the disadvantage that it takes a long time to detect the signal.

The present invention is intended to eliminate such drawbacks of the conventional art, and provides a signal detection device having a configuration that suppresses suddenly generated noise with strong pulse characteristics during averaging processing.

The operating principle of the present invention will be explained using FIG. 2.

The number of averaging operations that can sufficiently suppress normal white noise is set to N.
times. Therefore, it is assumed that the same signal is repeated N times and is generated periodically as shown in FIG. 2(a). This is subdivided into 8 basic unit times as shown in FIG. 2(b). This basic unit is called a frame, and n frames are created. That is, a X n =N. After being divided into frames, which are basic units, the average value of each frame is calculated as shown in Figure 2 (d), and the unique characteristics of the frame are determined from the average value as shown in Figure 2 (e). demand.

If pulse noise occurs here, it will be included in one of the frames, and the influence of this pulse noise will only extend to the next adjacent frame at most, and other frames will not be affected. . Therefore, the pulse noise will be absorbed into one of the frames, and the characteristics of that frame will lose similarity with the characteristics of other frames. When finally averaging all frames, this pulse noise can be eliminated by excluding frames with lost similarity from the averaging target.

The number of additions is small enough to suppress white noise.

It becomes possible to sufficiently suppress and eliminate the noise, and the overall signal-to-noise ratio can be improved.

In Fig. 2(e), if the third frame contains pulse noise, this f beam value is removed during the final averaging as a whole, and as a result, high-quality averaging processing is performed. It is possible to do this.

Next, one embodiment of the present invention will be described with reference to FIG. The signal input from the input terminal 1 is a periodically changing signal with noise superimposed on it.

It is assumed that the noise component includes pulsed noise that occurs suddenly. 2 is a mu/D converter.

The input signal is converted into a digital signal at a sampling period that does not lose the amount of information in the input signal.

- A signal period generator 3 is provided in order to perform addition at the repetition period of the force signal. Here, the signal period is set in advance on the transmitting side and the receiving side. 4 is a counter which inputs the output of the signal period generator 3 and counts the number of basic units for one frame.

Reference numeral 6 denotes a pre-adding and averaging device which adds and averages the input signals converted into digital quantities.Under the control of the control device 6, the adding and averaging device 6 performs the adding and averaging until the basic unit number is counted by the counter 4. Furthermore, when the set number of times is reached, the control device 6 resets the counter 4 and performs control to distribute the output of the pre-adding and averaging device 6 to a plurality of storage devices 8 by the next distributor 7. The distributor 7 distributes the frames so that they can be stored in locations corresponding to the frame numbers in the storage device 8. After this distribution and storage, the control device 6 resets the pre-adding and averaging device 6 and starts adding and averaging for the first frame. After data has been stored in all locations corresponding to each frame number in the storage device 8, the control device 6 stops the function of the MU/D converter 2 and stops taking in data thereafter.

Reference numeral 9 denotes a feature extracting device, which extracts a feature specific to each frame from data sequentially stored in locations corresponding to each frame number in the storage device 8. Here, to extract the characteristic amount, it is sufficient to be able to determine whether or not pulse noise exists, and if we pay attention to the magnitude of its energy,
The presence or absence of pulse noise can be determined by characterizing the energy of the frame after pre-averaging. The output of this feature extracting device 9 is input to a similarity determining device 10 that determines the similarity of features. This similarity determination device 1
0, it is determined that a frame with abnormally large energy relative to the whole has a sufficiently high possibility of containing pulse noise, and that frame is determined to be a frame with low feature similarity.
Output that frame number. 11 is an averaging device;
The pre-added average value from the storage device 8 is finally added and averaged, but the addition and average is performed by excluding data corresponding to the frame number output from the similarity discriminating device 10. Therefore, the averaging device 1 obtained at the output 12
The output of No. 1 is obtained as a high-quality signal in which white noise is suppressed and suddenly generated pulse noise is also removed.

According to the configuration of the above embodiment, the pre-adding and averaging device 6
Performs averaging processing for each frame.

Similarity determination device 10 from the output corresponding to each frame
The values for each frame are averaged by excluding values with poor similarity. Frames with sudden pulse noise are excluded as having low similarity, so for example, when communication is performed using a wireless line, pulse noise such as from a car ignition system in the vicinity of the receiving side Even if there is a source, highly reliable communication is possible without being disturbed by this pulsed noise.

Furthermore, the total number of averaging operations is approximately the same as the number that can suppress so-called white noise, and compared to conventional devices, this embodiment can suppress not only white noise but also pulse noise extremely effectively with the same processing time. It has the excellent feature of being able to be suppressed.

As described above, according to the present invention, the total number of addition averages is subdivided into basic units, and the pre-added average value obtained by performing addition average for each number of times of this basic unit is used, and the characteristic of each basic unit is This method extracts minute quantities, determines the similarity of their mutual feature quantities, removes only the prefixed average value of the basic unit with lower similarity, and averages the remaining prefixed average values.
Even if pulsed noise occurs, it can be removed very easily and effectively because the similarity of the pre-summed average value of the basic unit in which it occurs is clearly inferior. Also, since the total number of addition averages is almost the same as the conventional one,
White noise can be removed in the same way. In other words, the present invention is advantageous in that it suppresses noise and reduces the processing time for detecting signals at least as low as that of the conventional method, while effectively removing not only white noise but also pulse noise, thereby enabling signal detection with fewer errors. This is a signal detection device.

[Brief explanation of drawings]

FIGS. 1(a) to (d) are waveform diagrams illustrating averaging processing by a conventional signal detection device; FIGS. Fig. 3 is a block configuration diagram of this embodiment.1...Input end, 2...Mu/D converter,
3...Signal cycle generation circuit, 4...Counter, 5...Pre-adding and averaging device, 6...
・Control device, 7...distributor, 8...storage device, 9...feature extraction device, 1o river...
Similarity determination device, 11...Additional averaging device. Name of agent: Patent attorney Toshio Nakao (1st person)
Let's figure it out

Claims (1)

[Claims] A pre-adding and averaging device that subdivides the total number of addition averages into basic units and performs addition and averaging for each number of basic units; a storage device that stores the output of this pre-addition and averaging device; A feature quantity extraction device extracts the features from the pre-added average value for each basic unit number, a similarity comparison device which compares the similarity of each feature from the output of this feature quantity extraction device, 1. A signal detection device comprising: an averaging device that removes a pre-added average value having an inferior degree and performs averaging processing on only the remaining pre-added average values.