RU2128346C1 - Procedure of isolation of envelope of signal of random process - Google Patents

Abstract

FIELD: measurement technology, specifically, conversion of signals of random processes, automation of computer, home and medical equipment and telemetering systems. SUBSTANCE: invention is aimed at simplification of algorithm of operation SELECTION thanks to design of operation ANALYSIS in time region. Procedure of isolation of envelope of signal of random process consists in reading of moment of change of sign of first derivative of signal to read current magnitudes of coordinates of signal at given moment, in additional setting of algorithm of selection, in analysis of magnitude of signal at each moment of change of sign of first derivative by storage of magnitude of amplitude and by its comparison with subsequent magnitude. Comparison results are employed to make decision on sign of first derivative of envelope. In agreement with made decision data on amplitude and sign are fixed, information reading signal is formed, signal to read current magnitudes of coordinates of signal at given moment is extracted. EFFECT: simplified functional algorithm. 2 dwg

Description

 The invention relates to the field of measuring equipment, in particular, to the analysis of modulated signals and can be used in automation, radio engineering and code systems, for example, locks, control of household appliances and color visualization of musical / speech compositions.

 Known by the application N 63 46611 from 09.16.88 g, Japan, a data sampling method, which consists in extracting several discrete values of the input analog signal for each sampling period and sampling from these discrete data as data characterizing the whole sampling period of discrete data having maximum deviation relative to data extracted in the previous sampling period.

 The disadvantage of this method is the need for additional and relatively complex operations to select extreme values, which is associated with time costs and a delay in obtaining the final result when optimizing the results according to Shannon / Kotelnikov criteria.

The prototype according to the technical essence of the alleged invention is the method according to I. A. Lipkin "Fundamentals of statistical radio engineering, information theory and coding." M., Sov. radio, 1978, p. 22 by the Shannon and Kotelnikov theorems for a stationary signal, which consists in the fact that a continuous signal u (t), whose spectrum is bounded from above by a frequency F _in , is completely determined by its values counted at discrete points at intervals dt = l / 2F _in .

 The disadvantage of this method is the need to transmit redundant data with a low information value to restore the signal of a continuous random process, because the conditions of the theorems for real signals are not satisfied, since they represent processes that are time-limited, and therefore their spectrum is not limited.

 The technical result of the invention is to increase the information value of the reference when highlighting the envelope of the signal of a random process due to the construction of the operation is analyzed in the time domain.

 The technical result is achieved by the fact that in the method for extracting the envelope of the signal of a random process, which consists in sequentially counting data on the signal value, an additional algorithm for extracting the moment of changing the sign of the first derivative of the envelope of the random process is analyzed, the signal values are analyzed at each moment of changing the sign of the first derivative of the signal why remember the value of the amplitude at each given moment of the sign change of the first derivative of the signal and compare it with the subsequent value of the change of the sign of the first derivative signal based on the comparison decide the sign of the first derivative of the envelope, on the decision made is fixed on the amplitude data, sign and form data read signal, emit a signal for reading the current values of the coordinates of the signal at the moment.

 It is known that the value of the amount of information Stratonovich, Information Theory, Sov. Radio, M., 1975, pp. 325-333, the initial measure of which is the Hartley amount of information, is directly related to the number of sales (readings) and is "defined as the maximum benefit that can be gained from the Hartley the amount of information "(p. 304).

Подставив это разложение в выражение момента

с учетом центральных моментов

получим

Если функция φ(x) изменяется в указанном интервале достаточно плавно, то в разложении можно ограничиться учетом лишь первых двух членов. При этом
M{φ(x)} = m_f(j)~ φ(m_j)+ 0,5φ(m_j)D_j.
Отсюда видно, что матожидание достаточно плавной функции от случайной величины приближенно определяется матожиданием и дисперсией, роль высших моментов убывает с увеличением их порядка.The introduction of the operation sets the algorithm for isolating the moment of changing the sign of the first derivative of the envelope of the random process is necessary and sufficient to determine when the signal value reaches the highest value, which increases the information value of the transmitted result while eliminating data redundancy. It can be shown that lower moments are more important than higher ones. In most practical cases, the probability density P (x) has the most significant value in the interval located near the mathematical expectation m _j , the length of this interval being of the order of σ _j = VD _j . If in this interval the function φ (x) is sufficiently smooth, then it can be represented with the necessary degree of accuracy by the Taylor series in the vicinity of the point m _j :

Substituting this expansion into the expression of the moment

taking into account the central points

we get

If the function φ (x) varies quite smoothly in the indicated interval, then in the expansion we can restrict ourselves to taking into account only the first two terms. Wherein
M {φ (x)} = m _{f (j)} ~ φ (m _j ) + 0.5φ (m _j ) D _j .
This shows that the expectation of a sufficiently smooth function of a random variable is approximately determined by the expectation and dispersion, the role of higher moments decreases with an increase in their order.

 In addition, the introduced operation is necessary and sufficient for a new quality - adaptability. In the method algorithm, the dependence on the frequency properties of the input signal, i.e. it is operable in a wide range of frequencies and regardless of the shape / spectrum of the input signal (sinusoid, triangle, trapezoid...) provides the minimum necessary number of samples for its optimal recovery with the information value of each of them of the order of 0.95.

 The operation "allocate the moment of changing the sign of the first derivative of the input signal", consists in the fact that for the sequential formation of the reference voltage levels more or less, comparing them with the input signal, sequentially generating, depending on the results of the comparison, the voltage levels more or less and the input signal, additionally form a grid of frequencies f, set the maximum frequency f of successive formation of levels more or less, set the norms for the rate of change of the input voltage less-normal-more, setting the norms of permissible values of the input signal are checked, the sequence of formation of voltage levels more or less is checked and, as a result, the correspondence of the duration of the fact of the sequence deviation and the rate of change to the specified norms is determined, according to the deviations in the sequence of formation of more or less levels at the time of the deviation, they decide on the sign of the first derivative with the subsequent formation of a message about the sign of the first derivative of the input signal, by changing this message, a signal is generated about the moment the sign changes p the first derivative of the input signal, according to the results of determining the duration of the fact of deviation and the correspondence of the rate of change of the signal, they decide on the frequency f of the formation of reference levels more or less, select the conversion results when the signal reaches acceptable values, followed by blocking the formation of reference levels of limiting values until the input signal level decreases . The introduction of the operation “set the allocation algorithm” is necessary to reduce data redundancy with low information value. The greatest value in this case according to Boltzmann and Hartley is precisely the moment the sign of the first derivative of the envelope changes, the fact of separation of which simplifies the totality of subsequent operations. In particular, the selection algorithm may contain a criterion for processing only one of the components, for example, the upper one or separation for subsequent autonomous processing of the upper from the lower signal components. As examples, signals with one side component or electrocardiosignals (EX) can serve.

где K - коэффициент сжатия по числу отсчетов;
n - объем информационной части слова несжатых данных;
m_ин - объем информационной части слова сжатых данных;
m_сл - объем служебной информации, приходящей на один существенный отсчет.The introduction of the operation is “analyzed” is necessary and sufficient to obtain the optimal amount of signal data, sufficient to restore the information component of the random process signal without introducing known adaptive algorithms that work with a priori imposed restrictions, and regardless of the frequency of the input signal, it provides registration of coordinates of 2-3 points , which, in accordance with the Shannon / Kotelnikov theorems, is necessary and sufficient for signal recovery. It should be noted that the new quality allows for an increase in the compression ratio of the frequency band.

where K is the compression ratio by the number of samples;
n is the volume of the information part of the word uncompressed data;
m _in - the volume of the information part of the word compressed data;
m _SL - the amount of service information coming to one significant sample.

 On average, the value of the compression coefficient by the number of samples lies in the range of 5.. .100 times, overhead is reduced for the subsequent processing of the information obtained in this case.

 The introduction of the operation "remember the coordinates of the signal" is necessary and sufficient to save data about the signal during the operation "make a decision according to the results of comparison", because the signal has alternating sign du / dt.

 The introduction of the operation "compare the current value with the previous value" is necessary and sufficient to obtain the result of a change in the signal for the operation "make a decision".

 Introduction Operations make a decision about the sign of the first derivative envelope, according to the adopted decision it is necessary and sufficient for counting and generating a signal for reading data with the highest information value (according to Boltzmann or Hartley).

 Known in US patent N 4752897, CL. G 06 F 15/20, 1988, signal generation operations that convert a continuous process into digital form by discrete samples. These first digital signals are converted into the second, characterizing the lower frequency of the first digital signals, while the second digital signals determine the desired phenomena in a continuous process. A large amount of intermediate operations reduces the efficiency of decision-making on changes in a controlled process, i.e. It has the worst indicators of manufacturability, which limits its scope.

 The introduction of the operation "record data on the amplitude and sign" is necessary and sufficient to coordinate subsequent operations with the speed of receipt of information about the monitored signal.

 The introduction of the operation “form a signal for reading information” is necessary and sufficient to identify the fact of the receipt of data in subsequent mating operations.

Thus, the totality of the entered operations and the sequence of their implementation is necessary and sufficient to achieve the goal;
The formation of the necessary and sufficient amount of data to highlight the envelope of the signal of the controlled process, which is generally random, increase the value of the data to 95%, and also to ensure the reliability of the conversion results with fairly simple operations, the number of samples is determined only by the characteristics of the input signal and is formed at the time of change sign of the first derivative of the input signal.

 FIG. 1 is a timing chart explaining a method algorithm; FIG. 2 - a variant of the structure of the circuit of the device for extracting the envelope of the signal of a random process, where the ADC is made according to the patent N 2024194 of the Russian Federation, cl. H 03 M 1/48 1994

 The device of figure 2, contains the ADC 1, the memory unit 2, the allocation unit 3, the latching unit 4. The input device in through the serial connection of the ADC 1, the allocation unit 3, the latching unit 4, the memory unit 2 is connected to the output bus Output. The output bus of the ADC 1 is connected to the second input of the allocation unit 3, the second output of the allocation unit 3 is connected to the storage unit 2.

 We consider the operation of the method using the example of the signal of FIG. 1a with the upper envelope. In the initial state, for example, a criterion is set in the algorithm: the upper envelope is extracted, i.e., at the moment du / dt changes the sign of the first derivative of the signal from plus to minus (in Fig. 1a, points 2, 3, 4, 5, 6), with a trapezoidal component: points 5-6.

The amplitude obtained by this criterion is analyzed, as a result of the operation, the moment of du / dt-sign change of the first derivative of the signal of FIG. 1b. The increase in information value in the allocation of the envelope of the signal of a random process is based on the determination of the moments du ^o / dt of the change in sign of the first derivative of the envelope of fig.1b of this signal. The criterion for distinguishing the moment du ^o / dt is the change in sign of the amplitude increment at the given moments du / dt of the change in sign of the first derivative of the signal of Fig. An analysis by this criterion is a comparison of the amplitude value at time t _{i + 1} with the stored data of the previous t _i taking into account the sign du / dt. At each subsequent moment t _{i + 1, the} sign is confirmed and the corresponding coordinate is stored. At time t _{i + 1 of} FIG. 1b as a result of a change in the magnitude of the signal of FIG. 1a, the sign of FIG. 1f and the coordinates of the point t _i and form the read signal of the information of FIG. 1c. In this case, the data of the point t _{i + 1} is stored - the coordinate and sign. The du ^o / dt sign in FIG. 1f is represented in binary. The conversion algorithm allows you to select the trapezoidal components in the signal: points 5-6, which is reflected du ^o / dt = 0 fig.1f.

 Thus, information about the envelope of the signal is obtained in the form of a sequence of samples by which the envelope can be restored with a sufficiently high accuracy and accuracy using the appropriate algorithm. As shown above, the error determined by higher order components does not exceed 5-6%.

 Consider the operation of a variant of the device in figure 2.

The consequence of the signal at time t ₀ at the input of ADC 1 is the formation of a signal at its output, the sign of the polarity of the signal of figure 2 and the data on the output bus. Depending on the given selection algorithm, the selection unit 3 generates a signal of Fig. 1b, fixing the initial level of the input signal in it and the polarity of the signal in the fixing unit 4. The polarity of the signal of Fig. 1f is determined by the result of the comparison circuit in block 3 in the form of "<", ">" For the envelope of figa maximums “+/-” it will be, the sign “>”, for the envelope minima “- / +” it will be the sign “<". The operation of the comparison circuit in the allocation unit 3 initiates a signal at the input of the storage unit 2, which captures information in RAM 2; coordinates (amplitude, time) and du / dt sign. To the output of the device, the allocation unit 3 transmits the fact of the formation of information data for coupling the input speed into the mating system.

 The sequence of samples at time points 2, 4, 5, 8, 9 provides reading envelope information for transmission over the communication channel from the output of the device.

The positive technical and economic effect of this technical solution is determined by:
increasing the information value of the reference when highlighting the envelope of the signal of a random process,
exception of redundancy in the transmitted information,
presentation of information about the envelope of the signal in digital binary code,
the possibility of microelectronic performance, since there are no frequency-dependent components,
the scope of technical performance does not depend on the carrier frequency of the signal, which expands the scope.

 The action of the method and device variant is tested on mock-ups made on the IS series 564, 140, and is also modeled on a PC. The prototyping results were carried out for a carrier of the order of 500 kHz, simulation for a carrier of the order of 3 MHz. To compare capacity, prototyping was carried out with an amplitude-modulated signal and with an electrocardiogram, as a result of which fast and slow waves of the first and second types of human heart rhythm work were built.

 The results confirmed the efficiency and effectiveness of the operations introduced for use in various fields - technology and medicine.

Claims (1)

The method of extracting the envelope of the signal of a random process, in accordance with which the signal value data is sequentially counted, characterized in that the moment t _{i of} changing the sign of the first derivative of the signal is extracted for counting the current values of the signal coordinates, the signal value is analyzed at each selected moment t _i , for which remember the value of the amplitude and sign of the first derivative of the envelope of the mentioned signal and compare it with the value of the amplitude and sign of the first derivative of the envelope of the signal of the subsequent selected cient t _{i + 1} change in the first derivative signal mark, if a decision sign change of the first derivative of the signal envelope is fixed amplitude value and sign of the allocated time of t _i, as an information result on the envelope and generating information read signal, then the amplitude and sign of the derivative signal at the selected time t _{i + 1 is} stored for subsequent analysis.

Images