SU741273A1 - Two-channel correlometer - Google Patents

Description

When deterministic quantization

. (-, 2

I1 g / 2r and D «12 the number of quantization levels is large (i.e. there is no coarse quantization) and the signal range is large. In the general case, the quantization error depends on the signal size and when designing devices it is necessary either to impose restrictions on the signal parameters, or to introduce the necessary accuracy margin for adverse situations. Therefore, deterministic-probabilistic quantization in both channels of the correlometer allows performing the rough quantization mode and eliminating the most unpleasant systematic error. However, the addition of an auxiliary random signal to the signal, in comparison with deterministic quantization, doubles the maximum and averaged error, due to which it is necessary to increase — the measurement time. Such a correlometer structure would be optimal, which would allow to obtain unbiased estimates of the instantaneous values of the signal with errors of Lu and X equal to the minimum possible, i.e. d / 2 and. If the quantization error is statistically independent of the magnitude of the signal, then such a correlometer would have no limit on the signal parameters.

The aim of the invention is to improve the measurement accuracy.

The goal is achieved by introducing a digital-analog converter, a random number generator and an adder, each input of which is connected to the output of a random number generator and others to the corresponding outputs of its own comparators, the output of the digital-analog converter is connected to the second input of a voltage divider. , the control inputs of the random number generator of the first and second channels are connected respectively to the output of the strobe pulse generator and to the output of the delay element.

The drawing shows the block diagram of the proposed correlometer.

The correlometer contains channels A and B of the pre-processing signal, each channel has a matching unit 1, a current source 2, a voltage divider 3 connected between the current source 2 and the output of the digital-to-analog converter 4. Block 1 is connected to the first inputs of the comparators 5, the second inputs of which connected to the tapes of the voltage separator 3. The magnitude of the current of the current source 2 and the voltage value of the divider determines the magnitude of the quantum. The random number generator 6 is connected to a digital-to-analog converter 4 and an input of the adder 7, to the other inputs of which are connected the outputs of the comparators 5. The adder 7 converts the unitary code of the comparators 5 to a binary code and subtracts the instantaneous value of the auxiliary random signal from the estimate of the instantaneous signal digital-to-analog converter 4. In block 8 digital processing connected to the inputs of adders 7, the instantaneous values of the signals and the result are multiplied Ats are averaged. The timing of the entire device is carried out by the generator 9 strobe pulses connected directly to the control inputs of the comparators 5 and the generator 6 of the random numbers of channel A, and through the delay element 10 to the control inputs of the comparators 5 and the generator b of the random numbers of channel B, to the control input block 8 digital processing. The output of the counter 11 is connected to the control input of the strobe pulse generator 9.

By adding an auxiliary signal to the signal with a uniform distribution, then quantizing this additive mixture and subtracting from the quantization result of the digital equivalent of the auxiliary signal, an unbiased estimate of the instantaneous signal value is obtained, the error of which is no more than ± d / 2, and the dispersion of the estimates is not more than g2 / i2. An important advantage is also the fact that the one-dimensional error distribution is uniform, independent of the signal, and the spectrum of the error is uniform, similar to white noise.

There is no clear summation of the signal with the auxiliary signal in the correlometer. This is accomplished indirectly by shifting the family of quantization reference levels. Therefore, the estimate of the instantaneous value of the signal x is determined according to the expression x ng d - d / 2 for positive signal values, or -ng + (g - x 9 / for negative signal values, where n is the number of comparators 5 that have been triggered. The last expression can be converted to x (ng + - d / 2) So, in the adders get the code according to the expression (ng + - d / 2), having a + or - sign.

The operation of the correlometer is more convenient to consider from the beginning of the measurement of the correlation moment. Under the influence of the strobe pulse generator 9, the comparators 5 are gated and the quantization result is fixed. At the same time, the digital equivalent of the instantaneous value of the auxiliary signal from the random number generator b to adder 7 is generated and a new random number is generated, which replaces the old number at the input of the digital toggle converter 4. Delay element 10 delays the strobe pulse by the time G, after which the channel B occurs a similar sequence of operations, except that the digital processing unit 8 receives a clock pulse r under the influence of which the estimates are multiplied and averaged. The number of clock pulses is calculated by the accumulating counter, AND, upon overflow of which the gage stops.

The introduction of new elements — adders and the execution of an auxiliary random signal generator in the form of serially connected random number generator and a digital-analogue of a new converter make full use of the potential possibilities of stochastic signal quantization in the Canash.x preprocessing of signals. The static measurement error, as compared with the prototype, with all other parameters being the same, decreases in two areas. The introduction of new elements leads to some complication of the device, however, at the modern level of digital integrated circuits, this is justified, since it is more profitable to improve the metrological parameters of the device not at the expense of the angular one, but at the expense of the discrete part of the device. For example, both adders and a digital processing unit can be replaced by a microprocess or performed on chips of medium integration, for example, K1 55 IPP. The execution of a random signal generator in the form of a random or pseudo-random number generator with a digital-to-analog converter also corresponds to the current trends in instrument making, since it allows one to guarantee the stability and accuracy of the signal's probabilistic characteristics. The frequency of changing the codes of the digital-to-analog converter is equal to the frequency of taking paired samples, i.e. The speed requirements of this node are not very high.

The main advantages of the proposed device are the increase in measurement accuracy and independence

statistical error of the signal parameters.

Claims (2)

1. Zhovinsky V.N., Arkhovsky V.F./, Correlation devices, M., Energie, 1974, p. 23, 199. five 2. Authors certificate of the USSR 515113, cl. G 06 F 15/34, 1975 ,(prototype).