RU99623U1 - ADAPTIVE SIGNAL FILTER - Google Patents

Abstract

 1. An adaptive signal filter, consisting of an input device block, an adaptation circuit block, a dynamic error compensator block, and an identifier block, characterized in that the blocks of this device, by means of the electric and / or electronic elements included in them, implement the principle of filtering a scalar message with random identification dispersion identification unit, which allows to compensate for a priori unknown stream of interference of the processed signal caused by a priori unknown stream of measurement errors, and / or the error of the state of the signal source, and / or external noise. ! 2. The adaptive signal filter according to claim 1, characterized in that the processed signal is a data stream that determines the state of the object under study. ! 3. The adaptive signal filter according to claim 1, characterized in that the processed signal is a stream of data transmitted over communication channels. ! 4. The adaptive signal filter according to claim 1, characterized in that the processed signal is an electrical pulse. ! 5. The adaptive signal filter according to claim 1, characterized in that the processed signal is a perturbation of the magnetic field. ! 6. The adaptive signal filter according to claim 1, characterized in that the signal being processed is a perturbation of the electromagnetic field.

Description

Technical field

The utility model relates to data processing devices and can be used in devices for receiving information passing through a communication channel, as well as in systems for measuring the parameters of objects in which there is the possibility of distortion, i.e. distortions occur due to an error in the signal receiving system and / or an error in the state of the signal source and / or external noise.

State of the art

Analogues of the proposed method.

The method of adjusting the frequency transmission coefficients of the channels of a multi-channel receiver according to Russian patent No. 2239284 according to IPC Н04В 1/10, op. 10/27/2004, in which the filter coefficients are calculated by the input signal in the tuned channel and the ChKP is adjusted to the required ChKP. In the reference and tuned channels, initially the same digital filters with a finite impulse response (FIR) are used.

The disadvantage of the above method is that it does not allow you to take into account a priori unknown flow of external influences on the process of obtaining data, as well as signal distortions associated with measurement errors and the state of the signal source. In addition, the filter used is not adaptive, i.e. uses the initially set signal correction parameters, independent of changes in noise parameters.

Adaptive filtering system of audio signals to improve speech intelligibility in the presence of noise according to Russian patent No. 2163032 according to IPC G10L 21/02, op. 02/10/2001, in which, for selective modification of the digital signal frame, the following devices were introduced: an energy level estimator, a speech component detector, a noise estimator, operating when the speech component detector determines that the speech component does not form part of the frame, a lookup table containing a plurality inputs, wherein each input is equipped with an index indicating the level of noise estimates, and a filter for receiving a frame of a digital signal, the characteristics of which can be selected by entering the look-up table to which schenie and input corresponds to the updated noise estimate generated by the noise estimator.

The disadvantages of this device include the low adaptability of the correction signal, since it is based on a finite number of correction tables entered, as well as the need for an early assessment of possible incoming noise, which is not always possible.

Method and device for measuring the mass flow of material in a Coriolis mass flow meter according to the patent of Russia No. 2155325 according to IPC G01F 1/84, op. 08/27/2000, in which, when taking measurements using the ADC, periodically sample the analog output signals of the sensors connected to the vibrating flow tubes. In an embodiment, a plurality of adaptive filter plugs are used to amplify the signal. The weight adaptation means measures the parameters of the filter plugs to ensure that the delay band is adjusted to suppress unwanted noise components of the sensor output signals. The signal values from the filter plugs are supplied to the means for calculating the phase of the signal of each sensor.

The disadvantages of this device are: an unnecessarily large number of measuring elements, and hence the processed signals, as well as insufficient reliability of the resulting final signal, due to the lack of a reliable correction signal and the use of a simple signal sampling.

The closest in technical essence, the prototype of the utility model, is a device for digital filtering according to Russian patent No. 2055396 IPC G06F 17/17, op. 02/27/1996, in which the introduction of a number of logic elements into the known interference filter made it possible to organize adaptive filtering with the correction of deviations from the optimal filtering mode by matching the actual and calculated values of the variance of the measurement residual in the filter.

The disadvantages of this device are the low accuracy of adaptation of the correction signal, the low speed of the device, as well as insufficient resistance to a number of possible processed signals.

Utility Model Disclosure

The objective of the utility model is to obtain objective data about the studied objects or data transmitted via communication channels.

The technical results of the utility model are an increase in the accuracy, speed, adaptability and stability of the signal filter, as well as a decrease in the number of necessary measurements to obtain the corresponding reliability of the data.

Technical results are achieved in that the adaptive signal filter consists of an input device block, an adaptation loop block, a dynamic error compensator block, and an identifier block. This device, by means of connecting electrical and / or electronic elements in a known manner, implements the principle of filtering a scalar message with a random noise dispersion identification block, which allows to compensate for an a priori unknown noise stream of the processed signal caused by an a priori unknown flow of measurement errors and / or an error in the state of the signal source and / or external noise.

One of the types of signals processed by such a filter can be a data stream that determines the state of the object under study.

One of the types of signals processed by such a filter can be a stream of data transmitted over communication channels.

One type of signal being processed may be electrical impulses.

One of the types of the processed signal may be perturbations of the magnetic field.

One of the types of the processed signal may be perturbations of the electromagnetic field.

Utility Model Implementation

One of the possible ways to implement an adaptive signal filter may be to connect known electronic logic elements and / or elements of electrical circuits in a known manner, so that the resulting device implements the operation algorithm shown in figure 1, where:

1 is a block that implements the function x _{k / k-1} = F _{k / k-1} · x _{k-1 in a} known manner;

2 is a block that implements a function in a known manner.

;

3 is a block that implements the function ε _k = y _k -h _k · x _{k / k-1 in a} known manner;

4 - adaptation circuit block — adaptive recursive digital smoothing filter of a moving average with a compensator for dynamic filtering error;

5 is a block that implements a function in a known manner.

;

6 - block implements a function in a known manner

;

7 is a block that implements a function in a known manner.

;

8 is a block that implements a function in a known manner.

;

9 is a block that implements a function in a known manner.

;

10 is a block that implements the function x _k = x _{k / k-1} + γ _{x, k} · ε _{k in a} known manner;

11 is a block that implements a function in a known manner

ψ _{x, k} = (I-γ _{x, k} ) ψ _{x, k / k-1} ;

Where:

k Time reference number

D _{ε, k} Estimation of variance of input random sequence

β Moving average filter weight

ε _k Random variable in the input sequence

A priori estimate of the average input random sequence

ξ _k Posterior estimate of the average input random sequence

Δε _k Evaluation of the impulse disturbance in the input sequence (without taking into account the sign)

ζ _k Evaluation of the impulse disturbance in the input sequence (taking into account the sign)

A priori estimate of the random noise variance in the input sequence

δ _k Weighting factor

t _α Quantile for the α level of confidence

The random noise variance estimate in the input sequence

x _k State parameter vector

F _{k / k-1} Transition matrix of the state parameter vector

ψ _{x, k} Covariance matrix of errors in estimates of the state parameter vector

G _{k / k-1} Transition matrix of the vector of forming noise

ψ _{w, k} Covariance matrix of forming noise intensities

ε _k Vector of variations of measured parameters

y _k Measurement vector

h _k Observation matrix

ψ _{ε, k} Covariance matrix of the vector of variations of the measured parameters

γ _{x, k} Filter gain

γ _{w, k} Inverse gain of shaping noise

τ One-delay line

The implementation of this algorithm of the device will allow you to process several parallel signals without resorting to their normalization and other distortion, which in turn increases the accuracy and reliability of the data received at the output.

Claims (6)

1. An adaptive signal filter, consisting of an input device block, an adaptation circuit block, a dynamic error compensator block, and an identifier block, characterized in that the blocks of this device, by means of the electric and / or electronic elements included in them, implement the principle of filtering a scalar message with random identification dispersion identification unit, which allows to compensate for a priori unknown stream of interference of the processed signal caused by a priori unknown stream of measurement errors, and / or the error of the state of the signal source, and / or external noise. 2. The adaptive signal filter according to claim 1, characterized in that the processed signal is a data stream that determines the state of the object under study. 3. The adaptive signal filter according to claim 1, characterized in that the processed signal is a stream of data transmitted over communication channels. 4. The adaptive signal filter according to claim 1, characterized in that the processed signal is an electrical pulse. 5. The adaptive signal filter according to claim 1, characterized in that the processed signal is a perturbation of the magnetic field. 6. The adaptive signal filter according to claim 1, characterized in that the signal being processed is a perturbation of the electromagnetic field.