SU1091173A1 - Correlator - Google Patents

Abstract

A CORRELOMETER containing a matching unit whose input is the information input of a correlometer, and the output is connected to the first information input of the comparators block, the second information input of which is connected to the output of a digital-to-analog converter, the input of which is combined with the first information input of the arithmetic unit and connected to the output of the first pseudo-random generator Ш1sel, the outputs of the comparators block and the second pseudo-random number generator are connected respectively to the second and third information and the inputs of the arithmetic unit, the output of which is connected to the first information input of the accumulation unit, the output of which is the output of the correlometer, which, in order to expand the resolvability of the correlometer, a constant memory unit, a comparison unit, the first and the second adders, the first and second counters, the shaper of the stochastic pulse flow, containing the register, the pseudo-random number generator, the switch, the first and second counters, the clock inputs of which are connected respectively to n The first and second outputs of the switch, the first control input of which is combined with the clock input of the pseudo-random number generator, the generator of the stochastic pulse flow and connected to the output of the first counter of the generator of the stochastic flow of pulses, the second (P control input of the switch is connected to the input of the initial I setup and the output of the second counter of the generator of the stochastic flow of pulses, the information inputs of which are connected to the corresponding outputs of the pseudo-random number generator the stopper of the stochastic pulse flow, the inputs of the initial CO setup of which are connected to the corresponding register outputs, the bit outputs of the pseudo-random number generator of the stoosohastic pulse flow are the output of the parallel code of the former and connected to the first input of the first adder, the information input of the switch is the input of the former generator and connected to the output of the clock pulse generator, the output of the second counter of the generator of the stochastic flow of pulses pulse output of the driver and connected to the clock inputs

Description

The first n of the second pseudo-random number generators, the comparators block and the first counter, the output of which is connected to the second information input of the accumulation unit, the input of the initial installation of the first counter, are combined with the clock input of the second counter and connected to the transfer output of the second adder, the discharge outputs of which are connected to the address input of the memory block, the input of the second sum 173

109

the torus is connected to the output of the first adder, the second input of which is combined with the address input of the window block. And connected to the output of the second counter, the output of the permanent memory unit is connected to the first input of the comparison unit, the second input of which is connected to the output of the second pseudo-random number generator, the output of the block comparison is connected to the fourth information input of the arithmetic unit.

The invention relates to computing technology, in particular, to devices for processing information for a special purpose and can be used in various fields of science and technology. The permissibility of existing digital correlometers over time is methodologically limited by the speed of the element base used, which prevents the use of these devices for analyzing wideband signals. A multichannel sign correlometer, a process switching unit, two comparison units, a random signal generator, and a pulse generator are known. , a block of nominations and distributions, a block of n accumulations, a block of formation of zero ordinate, an element of RShI, two elements And, and a control trigger ij. The correlometer has a low resolution time capability and has low accuracy due to large statistical errors due to the quantization principle of input signals. The closest in technical essence to the present invention is a Correlometer comprising a clock pulse generator, two matching blocks, two comparators blocks, two pseudo-random number generators, two digital-to-analog converters, a delay-block, an accumulation block, and three adders, multiplication blocks, block fixed memory, forming in the aggregate the arithmetic unit 2 The correlometer implements the method of measuring the correlation function based on second-order stochastic quantization and periodic sampling input signals and has a fairly high accuracy. However, the sampling period determines the time resolution of the correlometer, which cannot be shorter than the time required to establish digital-analog converters and comparators, which limits the frequency range of the signal being processed. The purpose of the invention is to increase the resolution of the correlometer. This goal is achieved by the fact that a correrelometer containing a matching unit whose input is an information input of a correlometer and an output is connected to the first information input of the comparators block, the second information input of which is connected to the output of a digital-to-analog converter, the input of which is combined with the first information input of the arithmetic unit and connected to the output of the first pseudo-random number generator, the outputs of the comparators block and the second pseudo-random number generator are connected respectively with the second and third: information inputs of the arithmetic unit, the output of which is connected to the first information input of the accumulation unit, the output of which is the output of the correlometer, a fixed memory block, a comparison unit, the first and second adders, the first and second counters, a stochastic generator pulse flow, containing a register, a pseudo-random number generator, a switch, the first and second counters, clock inputs of which, respectively. u are connected to the first and second outputs of the switch, the first control input of which is combined with the clock input of the pseudo-random number generator of the stochastic flow generator and connected to the output of the first counter of the stochastic pulse generator, and the second control input of the switch is connected to the input. the initial installation and the output of the counter of the generator of the stochastic pulse flow, the second control switch input is connected to the input of the initial installation and the output of the second counter of the generator of the stochastic pulse flow, the information inputs of which are connected to the corresponding bit outputs of the generator of pseudo-random numbers of the generator of the stochastic pulse of pulses the initial setting of which is connected to the corresponding outputs of the register, and the discharge outputs of the pseudo-case generator Numbers of the stopper stochastic pulse flow generator are supplied by the parallel driver code and connected to the first input of the first adder, the switch information input is the driver input and connected to the output of the clock generator, the output of the second counter of the stochastic pulse flow generator is a pulse generator output and connected with clock inputs of the first and second generators of pseudo-random numbers, a block of comparators and the first counter, the output of which is connected to the second information input of the accumulation unit, the input of the initial installation of the first counter is combined with the clock input of the second counter and connected to the transfer output of the second mat, the separate outputs of which are connected to the address input of the permanent memory unit, the input of the second totalizer is connected to the output of the first sum of the torus, the second input of which is combined with the address input of the accumulation unit and connected to the output of the second counter, the output of the permanent memory unit is connected to the first input of the comparison unit, the second input of which is connected to the output of the second pseudo-random number generator, the output of the comparison unit is connected to the fourth information input of the arithmetic unit. The principle of operation of the proposed correlometer ensures the elimination of the effect of superposition of the high-frequency components of the signal under study, which imposes constraints on the implementation of the well-known method of representing correlation functions by means of digital samples of these functions determined at constant time intervals. The use of a stochastic stream with low discreteness as a discretization stream allows the order to increase the resolution of the correlometer and expand the analysis area and high frequency range at a relatively low average sampling rate of the signal under study. Figure 1 shows the block diagram of the correlometer; figure 2 - the structure of the former stochastic flow of pulses; on fig.Z - the structure of the arithmetic unit; 4 shows the structure of an accumulation unit. The correlometer contains (Fig. 1) the matching unit 1, the input i: oToporo is the input of the correlometer and the output is connected to the first input of the unit 2 of the comparators. The second input of the block 2 of the comparators through a digital-to-analog converter 3 is connected to the output of the first generator of 4 pseudo-random numbers. The output of the first generator 4 of pseudo-random numbers, as well as the outputs of block 2 of the comparators and the second generator of 5 pseudo-random numbers are connected by the corresponding information inputs of the arithmetic unit 6, the output of which is connected to the first information input of block 7 of accumulation. The output of block 8 of the associated memory is connected to the first input of block 9 of comparison, the second input of which is connected to the output of the second generator of 5 pseudo-random numbers. The output of the comparison unit 9 is connected to the corresponding information input of the arithmetic unit 6. The output of the generator 10 clocks with dinene input shaper 11 stochastic flow of pulses. The pulse output of the generator 11 of the stochastic flow of pulses is connected to the clock outputs of the 4 and 5 pseudo-random numbers of the comparators 2 and the first counter 12, and the parallel code output is connected to the first input of the adder 13. The second input of the adder 3 and the address input of the accumulation unit 7 are connected to the output of the second counter 14. The output of the adder 13 is connected with the input of accumulating adder 15 connected by the transfer output to the input of the initial installation of the first counter 12 and the clock input of the second counter 14, and the other outputs an address input unit 8 fixed memory. The output of the first counter 12 is connected to the second information input of the accumulation unit 7. The stochastic pulse flow generator 11 contains (FIG. 2) a register 16 whose outputs are connected to installation inputs of pseudo-random number generator 17, the outputs of pseudo-random number generator 17 are output of a parallel code of the stochastic flow generator 11 and connected to the corresponding information inputs of the first counter 18. The output of one of the counters 18 of the subtractor is connected to its own input of the initial installation and the corresponding control input of the switch 19, one of the outputs of which It is connected to the clock input of another counter 20 The output of counter 20 is connected to the clock input of the 17 pseudo-radiant number generator and another control input of the switch 19. The other output of the switch 19 is connected to the read input of the first counter 18, the output of which is the pulse output of the driver. The arithmetic unit 6 (FIG. 3) contains a constant memory element 21 in which a value code 1/2 is written, the first and second 22 and 23 adders, the first, second and third multiplication nodes 24-26, the third adder 27. The first inputs of adders 22 and 23 are connected and are connected to the output of the fixed memory element 21. The second input of the adder 23 is the first information input of the block, the outputs of the adders 22 and 23 respectively are connected to the first inputs of the multiplication nodes 24 and 25, the second input of the multiplication node 25 is combined with the first input of the multiplication unit 26 and is the second information ionic input unit, the second input unit 24, multiplying coupled to the second input node 26 multiplication, multiplying the output node 26 is connected to a first input of an adder 27, second and third inputs of which are connected to the outputs of nodes 24 and 25 multiplication is output block, respectively, 27 of the adder output. The second input of the adder 22 is the third information input of the block, the second input of the multiplication section 26 is the fourth information input of the block. The accumulation unit 7 holds the adder 28, the division node 29 and the memory element 30. The input of the adder 28 is the information input of the block, the output of the adder 28 is connected to the input of the division node 29, the output of which is connected to the information input of the memory element 27, the address input of which is the address input of the block. The output of the memory element 27 is the output of the block. The remaining blocks are well known and require no explanation. The correlometer estimates the each point of the correlation function of the studied signal x (t) and the reference signal y (a). The multiple-digit numerical samples of which are recorded in block 8 by the memory module are determined using the .M algorithm (.Aitl D. | iL -Z xk-hk-Puk, de (M (; ,,) fliC - the k-th stochastic time interval; A - volume of memory device 8; N - number of realizations of the reference signal y (a) used to calculate one correlogram point; output signals of generators of pseudo-random numbers 4 5 at the k-th instant of discretization; output signals of block 2 comparator c and block 9 of the comparison in the k time of sampling; half of the signal change range; the number of point estimates used to calculate one correlogram point; ... EZ is a whole part and Elw,., takes only .sO, -1, ..., P to integer values and in the right-hand side of expression (1) is present in an implicit form through the result of comparison p. The signal under study x (t) is centered, leads to the required scale –q, qj in matching block 1 and goes to the first input of the block of 2 comparators, on the second input of which from the generator 4 pseudo-case of the number of n are given uniformly distributed pseudorandom numbers 0, the converted digital to analog pre-forming 3. Stochastic quantization of quasi-simultaneously carried out in the ranges -qjO and o, q, i.e., unit 2 of the comparators contains two comparators. The positive values of the signal x (t) are compared with the reference levels qf., And the negative values are compared with the levels q (-1). The result of comparing the input signal x (t) with a pseudo-random level at the k-th sampling moment is determined by rule 1, if x (t) q 1c О, if (t) 7q () -1, if x (t) q (- l) The stochastic quantization of the reference signal y (t), 2jj multi-digit digital samples which are recorded in the block 8 of the permanent memory is carried out in digital form. Comparative block 9 contains two code comparators. The code y (ak) is compared with the code, on one comparator and with the code (the other. The result of the Pi comparison; at the k-th sampling time is defined as 1 if y (a) () n 0 if (1y + 1) y (ac) 1 (3) -1 if y (ak) the sampling times are determined by the stochastic pulse shaper I1. The stochastic pulse shaper 11 creates a sampling stream that is stochastic, described by the dispersion accumulation flow pattern, has a given dead- generation time during which sampling is not possible pulse, which is necessary to reconcile the low discreteness of the generated stream with a relatively low speed of the digital-analogue generator 3, block 2 of the comparators and the arithmetic unit 6. Shaper 11 generates a sequence of pseudo-random numbers and then transforms them into pseudo-random time intervals, at the time of the end of the sequence forms The pseudo-random time intervals that are generated are discrete and contain two components — children rminirovannuyu and pseudorandom. To implement the flow model with accumulation of dispersion, the formation of pseudo-random time intervals is carried out using the recurrent algorithm gu -ut; k 1,2, ... (4) where t. - the instant of the end of the k-ro pseudo-random time interval and the generation of the k-ro sampling pulse; clock interval; f- is the clock frequency; M is the counter module 20, which determines the constant component of the pseudo-random time interval, the so-called dead generation time, entered with, taking into account the dead time of blocks 2,3,6;

91091

J5 is a pseudo-random number used to form a k-ro pseudo-random time interval.

Since the sampling of the studied signal x (t) is carried out in real time in accordance with the expression (4), and oporsch) the signal y (t) is quantized and discretized with the frequency f "-tg of the form recorded in block 8 If its volume is bounded by A, then the connection between the arguments of the functions x (t) and y (aj /) is

G (M UK-01)

.VSJ- H AK (

where t is the k-th moment of discretization of sig -

feCM + jf., u "

feCM4y., U

 ftt argument correlation l ggk jl

Noah functions.

The sum of the first two terms in expression (5) means the sum modulo A of the numerical equivalents formed by the BpeMfiHHbix intervals. The current address is formed by the accumulator of the counter 14, the adder 13 and the accumulating adder 15. To the first input of the adder 13 from the output of the parallel code of the generator 11 of the stochastic pulse stream receives the code of the pseudo-random component Jf. time interval. The second input of the adder 13 from the output of the counter 14 receives the code of the deterministic component M, which is stored in the counter 14 before the measurements start. Thus, the input of the accumulating adder 15 receives the numerical equivalent of M formed by the time interval (M) ut. The accumulating adder 5 provides the sum modulo A of the numerical equivalents of the formed time intervals, forming the same address as the sampling time ty; the studied signal x (t). At the transfer output of accumulating adder 15, overflow sigals are formed, which are counted by counter 14. Relation of the volume of accumulating adder 15 to well A determines the number of realizations N of the reference signal y (a) used to calculate one point

correlation function. The appearance of the transfer signal at the output of accumulating adder 5 increases by one the contents of counter 14, which means setting the next argument of the correlation function. Since the sampling pulse stream is stochastic, then the same number of realizations H drops out a different number of sampling pulses. To count the number of averages 1. | individual points of the correlogram, counter 12 counts the number of sampling pulses between the transfer signals of accumulating adder 15, which nullify the contents of counter 12

The output signals of block 2 of the comparator p of the first generator 4 of pseudo-random numbers X of the second generator 3 of pseudo-random numbers K and of the Comparison Chunk 9 are fed to the corresponding inputs of the arithmetic block 6, where they are processed according to the output in the curly brackets of the algorithm. The calculated point estimates of the correlation function enter the accumulation unit 7, in which they are summed up for each iftt argument of the correlation function, and then averaged taking into account the values of 8 coming from the output of counter 12, and stored. It should be noted that in accumulation unit 7, the division is made after the end of the measurement, therefore the performance of the correlometer does not decrease.

the formation of a stochastic pulse stream is essentially as follows. High-frequency clock pulses, the value of which is determined by the speed of the switch 19, counters 18 and 20 (when building these blocks on 100-series chips, the clock frequency is 100 MHz), the output of the clock pulse generator 10 goes to the information input of the switch 19 and from the first output are fed to the input of the counter 20. After arriving at the input of the counter 20 pulses at the output of the counter 20, a pulse is formed which, arriving at the second control input of the switch 19 and the clock input of the generator 17 Pseudorandom numbers, causing The switching of the clock sequence to the input of the first counter 18 and the generation of a new pseudo-random number. Progress 1091

At the input, clock pulses reduce the contents of counter 18. After passing. pulses, the counter 18 will be in the zero state, which is indicated by the appearance of a pulse at j its output. This pulse determines the sampling time and, arriving at the first control input of the switch 19 and the input of the initial installation of the counter 18, the clock sequence switches to the input of the counter 20 and the initial installation of the counter 18 to the state U. The process repeats cyclically. Register 16 performs the function of storing the initial value 15 when it is necessary to repeat a sample of pseudo-random numbers.

In accordance with the above algorithm, the location of the generated pulse in time t is uniquely determined by C., the counter module 20th and the pseudo-random number Jf (, prerecorded in counter 18. At the output of the parallel 25 I driver code a parallel value code is issued. The generated pulse stream satisfies the condition

, / (i, rbKV (tK-tKM)

7312

which characterizes a stream with a small discreteness and a relatively large minimum period. This feature also allows you to mitigate the speed requirements of the 17 pseudo-random number generator, which can; be of the same order as the speed of the digital-to-analog converter 3 and block 2 of the comparators, since the maximum clock frequency of the generator 17 is limited to f-kfi / M, Mate,

The present invention allows the use of stochastic sampling properties, which include jB processing a wideband signal (with a relatively low average sampling rate. The upper frequency component of the signal under investigation and the discreteness of the correlation function argument are not dependent on the average sampling rate, but are determined by the discreteness of the stream of pulses which can be obtained quite small.

It should be noted that an increase in the resolution of the correlometer more than an order of magnitude compared to the prototype is achieved while maintaining the accuracy inherent in the prototype.

-H

AND

13

one

W

1

one

22

21

23

2

2

27

26

25

f1 / g3

Claims (1)

CORRELOMETER containing a matching unit, the input of which is the information input of the correlometer, and the output is connected to the first information input of the comparator unit, the second information input of which is connected to the output of the digital-analog converter, the input of which is combined with the first information input of the arithmetic unit and connected to the output of the first pseudo random number generator , the outputs of the comparator unit and the second pseudo-random number generator are connected respectively to the second and third information input arithmetic unit, the output of which is connected to the first information input of the accumulation unit, the output of which is the output of the correlometer, characterized in that, in order to expand the solvable capacity of the correlometer, a read-only memory unit, a comparison unit, the first and second adders, the first and second counters, a stochastic pulse stream generator, comprising a register, a pseudo-random number generator, a switch, the first and second counters, the clock inputs of which are connected respectively to the first and second at the outputs of the switch, the first control input of which is combined with the clock input of the pseudo-random number generator, the stochastic pulse flux generator and connected to the output of the first counter of the stochastic pulse flux generator, the second control input of the switch is connected to the input of the initial setup I and the output of the second counter of the stochastic pulse pulse generator whose information inputs are connected to the corresponding outputs of the pseudo-random generator pulse stream, the initial setting inputs of which are connected to the corresponding outputs of the register, and the bit outputs of the pseudorandom number generator of the stochastic pulse stream generator is the output of the parallel generator code and connected to the first inputs of the first adder, the information input of the switch is the input of the generator and connected to the output of the clock generator , the output of the second counter of the generator of the stochastic pulse flux is the pulse output of the shaper1091173> the generator and is connected to the clock inputs of the first and second pseudo-random number generators, the comparator unit and the first counter, the output of which is connected to the second information input of the accumulation unit, the initial setting input of the first counter is combined with the clock input of the second counter and connected to the transfer output of the second adder, bit the outputs of which are connected to the address input of the read-only memory block, the input of the second sum 1091 173 of the torus is connected to the output of the first adder, the second input of which is combined with the address input of the block Single accumulation .i connected to the output of the second counter, the output of unit-volatile memory connected to the first input of the comparator, the second input of which is connected to the output of the second random number generator, the output of the comparator is connected to a fourth data input of the arithmetic unit.