SU558407A1 - Adaptive intersymbol distortion equalizer - Google Patents

Description

one

The invention relates to the transmission of signals of discrete information and can be used in channels with phase shift in short-wave and ultra-wavelength radio communications in the presence of intersymbol patterns.

An intersymbol distortion adaptive equalizer is known in which a linear transversal filter is used to generate an error signal (echo), whose tap control determines the degree of correspondence of the error signal synthesized by the transversal filter to the error signal contained in the input signal, for which correlating the compensated signal with the input signal.

Based on the magnitude and sign of the correlation obtained, an adjustment signal is generated to adjust the transversal filter taps.

However, the well-known adaptive equalizer, due to the impossibility of a linear filter combating with large intersymbol interference, has a low corrective ability in the transmission of discrete information over radio channels with multipath propagation.

An adaptive intersymbol distortion equalizer is known, which contains a demodulator, the output of the latter is connected to the first input of the first subtraction node, to the second input of which through the first connected strobe node, the first

the decisive node, the zero-sampler, the second subtraction node, the error generator components of the error signal are connected to the adder, the output of the first gating node is connected via the second subtraction node to the second input of the zero-count generator, and the second output of the first crucial node is connected to the second one via the shift register the inputs n of the driver components of the signal error | Ki-.2.

However, this adaptive equalizer has a low accuracy of correction of information signals.

The purpose of the invention is to improve the accuracy of the correction of information signals.

To do this, the second gating node, the threshold node, the second decision node and the switching node are entered into the adaptive intersymbol distortion equalizer, and the output of the first gate node is connected to the first input of the commutation node via the threshold node, and to the second input of the switching node via the second crucial node, the other inputs of the latter are connected respectively to the outputs of the driver components of the error signal.

3

The drawing shows a structural electric circuit of the proposed adaptive equalizer of intersignal distortion.

Adaptive equalizer of distortions distortion contains demodulator 1, the output of the latter is connected to the first input of the first subtraction node 2, to the second input of which through the serially connected first gating node 3, the first crucial node 4, the zero reading unit 5, the second subtraction node 6, the n reference number 7 The components of the error signal are connected to the adder 8, the output of the first gating node 3 is connected through the second subtraction node 6 to the second input of the zero-count generator, and the second output of the first decision node and 4 is connected through the shift register 9 to the second inputs of the formers 7, except that, the equalizer contains the second gate node 10, the threshold node 11, the second crucial node 12 and the switching node 13, while the output of the first gate node 3 is connected to the first input of the switching node 13 through the threshold node 11, and the second input of the switching node 13 through the second crucial node 12, to the other inputs of which are connected respectively the outputs of the n drivers 7, and the outputs of the switching node 13 are connected to the inputs of the shift register 9.

Adaptive equalizer works as follows.

The demodulated signal from the output of the demodulator 1 is fed to the first input of the first subtraction node 2, to the second input of which a total error signal is fed from the adder 8. The components of the error signal are formed by the shaper 7 components of the error signal, the inputs of which are given the residual signal from the output the second subtraction node 6 and the signals from the outputs of the shift register 9. The compensated signal from the output of the first subtraction node 2 through the second gating node 10 is fed to the first input of the threshold node 11 and through the first gating node Ani 3 to the input of the first decision node 4, the output signal of which is fed to the forms, the zero-count transmitter and the input of the shift register 9.

A residual signal from the output of the second subtracting node 6 is applied to the second zero-form generator input. The output of the threshold node 11 is connected to the shift register 9 via the switching node 13. The levels of the error signal components are controlled by the second decisive node 12, the output of which controls the switching node 13. information is derived from the output of shift register 9.

But if the signal components are formed by n formers 7 based on the values of the information symbols, the presence of an erroneous information symbol, obtained at the output of the first decision node 4 at the / s-th preceding clock interval and located in the k-th discharge of the shift register 9, causes to the formation of the / s-th component of the error signal with the opposite phase in relation to the phase of this component, and its output signal of demodulator 1, as a result, instead of subtracting this component of the error signal from the output signal, modulator 1, they are summed up, followed by a sharp jump in the signal level at the output of the first subtracting node 2. By registering this jump with a threshold node 11 and then inverting the protective information symbol with the inverter of the shift register 9, the accuracy of the information signal correction at the input of the first decisive node 4 is improved. the trigger level of the threshold node 11 must be such that the probability of occurrence of signal level jumps,

registered by threshold node 11, under the influence of noise, was significantly less likely to exceed the threshold due to an erroneous signal correction. This is achieved by selecting the trigger level of the threshold node 11, which is 2-3 times greater than the signal level at the input of the first decision node 4. The threshold node 11 is made with the following threshold varying according to the change of the signal level at the input of the first slashing node 4.

The accuracy of the correction of the information signal depends on the correctness of determining the bit number of the shift register 9, in which at the time of registration of the level jump by the threshold node 11 there is an information symbol, which is used by the second decisive node 12, which measures the levels of the error signal at the outputs of drivers 7 and fixing bit number of shift register 9, if the level of the signal at the output of the shaper 7 component of the error signal corresponding to this bit is at least 1.5 times higher than the level of b drove the input of the first decision node 4. When there are several signals at the outputs of the formers 7, the error signal components are at levels 1.5 times greater than the signal level at the input of the first decision node 4, the second decision node

12 fixes the minimum number of the driver.

At the time of registration of the jump by the threshold node 11, the control voltage from the output of the second decisive node 12 is fed to the input

switching node 13 connects the output of the threshold node 11 to the inverter input of the shift register 9, which is included in that bit of the shift register 9, whose number is fixed by the second decisive node 12.

The inverter of the shift register 9 inverts the polarity of the information symbol recorded in this shift register register 9. To receive the input of the first decision node 4 signal after correcting the error information in the corresponding register register 9, the second node of the gated 10 is executed lead in advance with respect to the operation of the first node of strobing 3.

Comparative tests show the high efficiency of this device, i.e., the probability of error at the output of the adaptive equalizer is, on average, one way compared to the probability of error: at the output of the adaptive equalizer without a correction device.

Claims (2)

1. The patent of Great Britain N ° 1334250, cl. H 04B 3/16, 1974 2. The US patent, US 3614623, cl. H 04B 1/10, 1E7 g. (Prototy).