RU2164725C2 - Method for processing orthogonaly-shifted broad- band signals with structural noise correction - Google Patents

Abstract

FIELD: radio engineering; communication systems. SUBSTANCE: method depends on multiplication of input mixture by reference sync signal, filtering of multiplication results in N frequency channels followed by signal filtration, amplification, detection, and comparison with threshold levels; signals exceeding threshold levels are recovered and used for correcting respective signals in input mixture. EFFECT: improved noise immunity. 5 dwg

Description

 The present invention relates to radio engineering and can be used in communication systems with broadband signals.

 Known methods for correlation processing of signals with orthogonal frequency shift, used in a multicast system with the choice of an arbitrary subscriber, the disadvantage of which is low noise immunity to structural interference.

 The closest in technical essence to the proposed one is the method implemented in the device for correlation signal processing with orthogonal frequency shift, presented in the monograph by G. I. Tuzov “Static theory of the reception of complex signals”, M., “Sov. Radio”, 1977, p. 111, fig. 3.3, adopted as a prototype.

 The method adopted for the prototype is that the input mixture is multiplied with a reference broadband signal synchronous with the useful signals, the results of the multiplication - a convoluted signal - are filtered in N bandpass filters.

To implement this method, a device for correlation signal processing with an orthogonal frequency shift is presented, the structural diagram of which is shown in FIG. 1, where indicated:
1 - multiplier;
2 ₁ - 2 _N - band-pass filter;
3 - signal copy generator.

The prototype device contains a multiplier 1, the input of which is the input of the device, and the output is connected to the inputs 2 ₁ - 2 _N bandpass filters, the outputs of which are the outputs of the device. In addition, the output of the signal copy generator 3 is connected to another input of the multiplier 1.

In FIG. 2 shows an enlarged diagram of the device (with the aim of simplification), where it is indicated:
1 - block multiplication;
2 - filtration unit;
3 - signal copy generator.

 In FIG. 2 comb of N filters combined into a filtration unit.

 The device prototype works as follows. The input mixture enters the multiplier 1, where it is multiplied with the reference broadband signal, synchronous with the useful signal coming from the signal copy generator 3. The result of the multiplication - the convoluted signal - goes to the filtering unit 2, where it is filtered. Since all N signals differ in frequency shift, each signal is collapsed into its own filtering unit (at its own frequency). However, due to the imperfectness of the inter-correlation functions of the signals in the frequency domain, the products of the mutual correlation of the signal of this subscriber with the signals of other subscribers are present in each filtering unit. This is similar to how, in code division of channels, each signal is minimized by multiplying with the corresponding reference signal, but cross-correlation products in the time domain are present from all receivers.

 The disadvantage of the prototype method is the low noise immunity to structural interference.

выполняют после последовательного вычитания их входной смеси оценок более мощных сигналов (помех) 1, 2 ... (i-1)-го уровней, при этом оценку помех i-го уровня формируют за счет того, что перемножают свернутые узкополосные сигналы i-го уровня с опорным широкополосным сигналом.To eliminate this drawback, the method consists in multiplying the input mixture with a reference broadband signal synchronous with the useful signal, followed by filtering the result of multiplication (convolution) in N frequency channels, after multiplying and filtering in each of the N frequency channels discretely on the first (k -1) the steps measure the levels of folded narrowband signals, and the measurement of signals of the i-th level

after successive subtraction of their input mixture, estimates of more powerful signals (interference) of the 1st, 2nd ... (i-1) -th levels are performed, while the assessment of interference of the i-th level is formed due to the fact that the convoluted narrow-band signals of the i-th are multiplied reference with a broadband signal.

 The inventive method of processing broadband signals with an orthogonal frequency shift with compensation for structural interference is that the input mixture is multiplied with a reference broadband signal synchronous with the useful signals, the multiplication result (coiled signal) is filtered in N frequency channels, and the above operations are repeated stepwise k times . At the first (k-1) -th stages after filtering, the levels of convoluted signals in each frequency channel are discretely measured by comparing them with thresholds, while the threshold at the i-th stage is set below the threshold at the (i-1) -th stage. Orthogonal signals, the levels of which exceeded the thresholds, are multiplied with the reference signal, due to which they are converted to broadband.

 The reconstructed broadband signals of the first level are subtracted from the input mixture, the first difference signal is multiplied with the reference broadband signal, the result of the multiplication (convoluted signal) is filtered using a comb of N filters, then the signal levels in each frequency channel are measured, the convoluted signals of the second are extracted from the mixture (more low level), convert them to broadband signals, similar to the broadband signal in the input mixture.

 These reconstructed signals are subtracted from the first difference signal. The above operations are repeated (k-1) times (where k is the number of measured levels). After subtracting from the input mixture the signals of 1, 2 ... 3, (k-1) -th levels, the difference signal is multiplied with the reference pseudo-random sequence, then the result of the multiplication (convoluted signals) is filtered in N filter units.

The structural diagram of a device that implements the inventive method is shown in FIG. 3, where indicated:
1 - multiplier;
2 ₁ , 2 ₂ , 2 ₃ - the first, second and third filtering unit;
3 - signal copy generator;
4 ₁ , 4 ₂ - the first and second meter of signal levels;
5 ₁ , 5 ₂ - the first and second block of signal recovery;
6 ₁ , 6 ₂ - the first and second block interference compensation;
7 ₁ , 7 ₂ - the first and second block of multiplication and delay signals;
8 - switching unit.

The device comprises a series-connected multiplier 1, a first filtering unit 2 ₁ , a first signal level meter 4 ₁ , a first signal recovery unit 5 ₁ , a first interference compensation unit 6 ₁ , a first signal multiplier and delay unit 7 ₁ , a second filtering unit 2 ₂ , and a second a signal level meter 4 ₂ and a switching unit 8, the output of which is the output of the device. The output of the signal copy generator 3 simultaneously with the corresponding inputs of the multiplier 1, the first and second signal recovery block 5 ₁ and 5 ₂ , the first and second block of multiplication and delay of the signals 7 ₁ and 7 ₂ . The inputs of the multiplier 1 and the first block of noise compensation 6 _{1 are} combined and are the input of the device, and the output of block 6 _{1 is} connected in series through the second block of noise compensation 6 ₂ , the second block of multiplication and delay of signals 7 ₂ and the third filter block 2 _{3 are} connected to one of the inputs switching unit 8. In this case, the output levels of the first measuring signal is coupled to a corresponding input switching unit 8, and a connection point of the other output meter April ₁ and the first input unit 5 ₁ is connected to another input of switch 8. The output of the second measure the I signal levels on April ₂ simultaneously coupled to one input of switching unit 8, and through a second recovery unit signals 5 ₁ - to another input of the second interference compensation unit ₆ February.

 The device operates as follows.

 To simplify, consider the case when the signals are divided into three levels: 1st — maximum, corresponding to the nearest subscribers; 2nd — medium, 3rd — minimal, corresponding to the most distant subscribers.

где T - длительность информационного сдвига), поступает на блок 1, где перемножается с опорной псевдослучайной последовательностью, поступающей с выхода генератора 3.An input mixture containing signals of N subscribers having the same carrier frequencies and the same structure of pseudorandom sequences used in their formation and differing in the frequency shift of the carrier frequencies (equal for adjacent channels

where T is the duration of the information shift), enters block 1, where it is multiplied with the reference pseudorandom sequence coming from the output of the generator 3.

As a result of multiplication, the broadband signals are collapsed into narrowband, shifted in frequency, which are fed to the input of block 2 ₁ .

In block 2 ₁ , narrow-band signals are filtered using a comb of N filters, after which they simultaneously enter the input of block 4 ₁ .

In block 4 ₁ in each frequency channel, the signals are amplified, detected and compared with the first (highest) fixed threshold. The commands for exceeding the threshold and narrowband signals that exceed the threshold are given to the control and signal inputs of block 8, respectively.

Narrow-band signals that exceeded the threshold at the same time are fed to block 5 ₁ , where they are converted to broadband signals similar to those in the input mixture.

 This is achieved due to the multiplication (phase manipulation) of narrow-band signals with a reference pseudorandom sequence coming from the generator 3.

The restored broadband signals of a large level (exceeding the first threshold) from the output of block 5 ₁ are fed to block 6 ₁ , where they compensate for the corresponding signals in the input mixture. From the output of block 6 _{1, the} input mixture, from which high-level signals are excluded, goes to block 7 ₁ , where it is multiplied with the reference signal of block 3. As a result of multiplication, the broadband signals are folded into narrowband signals, which are filtered by a comb of N filters in block 2 ₂ and then they go to block 4 ₂ , where the signal levels are compared with the second (average) threshold.

Commands about exceeding the threshold and signals whose levels have exceeded the 2nd threshold are given from the outputs of block 4 ₂ to the control and signal inputs of block 8, respectively. Narrow-band signals are sent simultaneously to block 5 ₂ , where they turn into broadband signals similar to the corresponding signals in input mixture, due to phase manipulation of their reference pseudorandom sequence of block 3.

The recovered signals compensate for the corresponding average level signals in the input mixture in block 6 ₂ , the input of which receives the input mixture from block 6 ₁ , from which the signals of the first level are excluded. From the output of block 6 _{2, the} input mixture in which the signals of the first and second levels are compensated are fed to block 7 ₂ , where small-level signals are convolved, and then fed to block 2 ₃ , where they are filtered. From the output of block 2 ₃ , low-level signals are fed to the input of block 8. In block 8, using control commands from blocks 4 ₁ and 4 ₂ , switching and decoupling of the signal outputs of blocks 4 ₁ , 4 ₂ and 2 _{3 are} carried out, providing connection to each output of the device, the output of one of the blocks 4 ₁ or 4 ₂ , or 2 ₃ and the exclusion of the mutual influence of blocks 4 ₁ , 4 ₂ and 2 ₃ .

 Consider the hardware implementation of the blocks.

In FIG. 4 shows a block diagram of a signal level meter 4 ₁ and 4 ₂ , where it is indicated:
41 ₁ - 41 _N - delay element;
42 ₁ - 42 _N - key;
43 ₁ - 43 _N - amplifier;
44 ₁ - 44 _N - detector;
45 ₁ - 45 _N - comparison block with a threshold.

The block of the signal level meter 4 ₁ (4 ₂ ) contains N identical rulers, each of which contains a delay element 41 and a key 42 connected in series, and the output of the key 42 is the output of the block, which is connected to blocks 5 and 8. The input of block 4 is connected to the combined the inputs of the delay element 41 and the amplifier 43, the output of which through a series-connected detector 44 and the comparison unit 45 is connected to another input of the key 42 and at the same time is the output of the block 4, which is connected to the block 8.

 Block 4 operates as follows. The narrow-band signal from the output of each of the band-pass filters of the filtering unit 2 is amplified in block 43, detected in block 44 and compared with a threshold in block 45. The command of block 45 about exceeding the threshold opens a key 42, which transmits a narrow-band signal (coming to the key 42 from block 2 through the delay element 41) to the signal inputs of blocks 5 and 8. At the same time, commands of blocks 45 about exceeding ("1") or not exceeding ("0") thresholds are supplied to the control inputs of block 8.

The signal recovery block 5 ₁ (5 ₂ ) contains a series-connected adder and a multiplier, to the reference input of which a signal is supplied from block 3 through a delay element, the delay value of which is selected so that synchronization of the multiplied signals is ensured.

Block 6 ₁ (6 ₂ ) is a subtractor, to the second input of which a signal can be supplied through a delay element and an attenuator, which ensure the alignment of the signals at its inputs in phase and amplitude.

Blocks 7 ₁ , 7 ₂ are made in the form of a multiplier, to the reference input of which the reference signal is supplied through an adjustable delay element that ensures synchronization of the multiplied signals.

The block diagram of the switching unit 8 is shown in FIG. 5, where indicated:
81 _1-N , 85 _1-N - inverters;
82 _1-N , 86 _1-N , 87 _1-N - keys;
83 _1-N , 84 _1-N , 88 _1-N - amplifiers.

The switching unit 8 operates as follows. Narrow-band signals from block 4 ₂ are fed to the signal inputs of keys 82 ₁ - 82 _N , from the outputs of these keys signals through amplifiers 83 ₁ -83 _N are fed to the inputs of the device, which also receives signals from block 4 ₁ through amplifiers 84 ₁ -84 _N and the signals from block 2 ₃ coming through series-connected keys 86 ₁ -86 _N , 87 ₁ -87 _N and amplifiers 88 ₁ -88 _N. Key management 82 ₁ -82 _N and 87 ₁ -87 _{N is} carried out by commands from block 4 ₁ , coming through inverters 81 ₁ -81 _N. When the command “1” appears on the control outputs of block 4 ₁ , the keys 82 ₁ -82 _N and 87 ₁ -87 _N are closed, not passing signals from block 4 ₂ and 2 ₃ to the corresponding outputs of the device, providing only signals from block 4 ₁ . The control commands from block 4 ₂ prohibit the passage of the corresponding signals from block 2 ₃ to the outputs of the device.

Amplifiers 83 _1-N , 84 _1-N and 88 _1-N ensure the passage of signals in one direction and eliminate the mutual influence of blocks 4 ₁ , 4 ₂ , 2 ₃ , 5 ₁ and 5 ₂ .

 The prototype method is based on the simultaneous correlation processing of signals of near and far subscribers, while powerful signals of near subscribers have an interfering effect on the reception of signals from remote subscribers. The disturbing effect is due to the presence of outliers of the cross-correlation function of broadband signals in the frequency domain.

In accordance with the monograph edited by VB Pestryakova "Noise-like signals in information transmission systems", Moscow, Sov. Radio, 1978, p. 383, permissible excess of the power of one structural noise over a signal, determined by their inter-correlation properties, at B = 10 ³ (where B is the signal base ), does not exceed 20 dB. With an increase in the number of structural interference, this figure decreases.

 The inventive method involves performing correlation processing of the signals of remote subscribers after compensation from the input mixture of signals of near subscribers. The phased compensation of structural interference provides an additional gain in noise immunity to structural interference of 20-40 dB, while the allowable level of interference for the considered example is 40-60 dB, which is much more than for the prototype.

Claims (1)

 A method for processing broadband signals with an orthogonal frequency shift with structural interference compensation, based on multiplying the input mixture with a synchronous reference signal, filtering the multiplication result in N frequency channels, characterized in that after filtering the signals are amplified, detected, compared with the first (high) threshold, high-level signals that exceed the first threshold (high-level structural noise), restore and compensate for the corresponding signals in the input mixture, which again make the above sequence of operations, compensating for structural noise of the second, third, and so on levels, after compensation from the input mixture of structural noise, it is multiplied with a synchronous reference signal, the result of the multiplication is filtered in N frequency channels.

Images