RU2313902C1 - Method for relative amplitude demodulation of digital signals - Google Patents

Abstract

FIELD: technology for processing impulse electric signals, possible use as a decisive in end digital equipment, and also in regenerators and repeaters, where restoration of primary video-impulse shape of binary signals is supported.

SUBSTANCE: method includes comparing levels neighboring bits accumulated in RC-integrators, and gain in signal/noise ratio results from mutual compensation of noises during subtraction, and this effect increases with narrowing of pass bands of RC-accumulators. When polarity of parcels is changed at output of differential cascade - positive and negative drops, on repeat of sign pattern of parcels 2nd comparator is used, which disables output of differential cascade and connects memory of last recognized bit to output of device.

EFFECT: increased signal/noise ratio.

1 tbl

Description

The invention relates to techniques for processing video-pulse electrical signals and can find application in measuring equipment, long-distance wired and fiber-optic communications, as well as in radio and microwave communications.

An analogue of the invention can serve as a method for receiving a deterministic signal ([1], p. 451, Fig. 11.3, b), involving the use of a matched filter and threshold device.

The prototype of the invention is the "Method and device for analogue extraction and accumulation of pulse signals" (patent No. 2212111 from 09.10.03), where, in order to optimize convolution of the pseudorandom sequence, the difference in the nearest levels accumulated on the integrators during one bit is measured: (signal + interference ) and (interference) and summarize the level differences received on each bit. Such processing is sufficient for a known probing signal generated in an OTDR; another procedure is needed to recognize the sequence of unknown bits.

The essence of the invention is to compare, by subtracting the input voltage levels accumulated on the RC chains for the time τ of the current bit of the signal and the previous bit stored for the time τ. That is, two drives are needed that are connected alternately to the input of the device for a time τ and preserve the accumulated (over time τ) level for another time τ. At the end of each interval τ, in the interval from about 0.9τ to τ, the accumulated levels of H ₁ and H ₂ in the difference cascade P are compared and the polarity of the measured difference (H _{1 (2)} -H _{2 (1)} ) in the 1st comparator with a zero threshold.

In the simplest case, when receiving points (meander), it is enough to just connect the drive outputs alternately, through the interval τ, to the inverting and non-inverting inputs of the difference cascade.

When repeating the significance of the premises, the first of the repeated premises is critical. Its significance identified by the difference in accumulations (H _{1 (2)} -H _{2 (1)} ) applies to all others. In this case, a bit polarity repeater is used - the 2nd comparator with a switching threshold equal to 0.5H ₁ (H ₂ ), to which a voltage equal to | H _{1 (2)} -H _{2 (1)} | is applied. When the bits are repeated, the 2nd comparator disables the output of the 1st comparator, connecting simultaneously to the output of the device a memory record of the last recognized bit stored in the shift register. The threshold of 0.5 N is taken because on average the number of drops and repetitions is the same.

A brief assessment of the preference of the proposed method.

Здесь m_s(t) - среднее значение сигнала, σ_n(t) - среднеквадратичное значение шума, N_o -спектральная плотность шума, δ(τ) - дельта-функция. ТогдаWe find the signal-to-noise ratio q = m _s / σ _n at the output of the RC drive, provided that it is affected by a rectangular pulse of amplitude A _o and duration τ in the presence of white noise with a correlation function

Here m _s (t) is the average signal value, σ _n (t) is the rms noise value, N _{o is the} spectral noise density, and δ (τ) is the delta function. Then

где

. Полагая ατ<<1 (при этом exp(-ατ)≅1-ατ), придем к выражению

характерному для оптимального приема (E=A_o ²τ - энергия импульса).

Where

. Putting ατ << 1 (in this case exp (-ατ) ≅1-ατ), we arrive at the expression

characteristic for optimal reception (E = A _o ² τ is the pulse energy).

То есть выигрыш в отношении сигнал/шум по сравнению с q растет по мере сужения полосы пропускания фильтра RC, см. таблицу.Assuming that a signal with noise is received in the first interval τ, and only noise in the 2nd interval, and the input and output of the 1st drive in the interval τ≤t≤2τ are isolated from the housing, we obtain at the output of the difference stage q ₁ = ατA _o / σ (n ₁ -n ₂ ). Here, σ ² (n ₁ -n ₂ ) = 2σ _n ² (1-R) (see [1], p. 179), where R = exp (-ατ), that is, σ ² (n ₁ -n ₂ ) = α ³ τ ² N _o , a

That is, the gain in signal-to-noise ratio compared to q increases as the bandwidth of the RC filter narrows, see table.

ατ 0.1 0.01 0.001

2.236 7.072 22.36

The rms noise value σ (n ₁ -n ₂ ) also decreases by the same number of times when repeating the significance of the bits.

The source of information

1. Tikhonov V.I. "Statistical radio engineering", M., 1966

Claims (1)

The method of relative amplitude demodulation of digital signals, in which the input voltage is connected alternately, for a time τ of the bit duration to the input of two storage RC circuits, the accumulated level is stored for another time τ, at the end of each interval τ, the accumulated levels of H ₁ and H ₂ are compared difference cascade, and the outputs of the accumulating RC circuits alternately, through the interval τ, switch to the inverting and non-inverting inputs of the difference cascade and determine the polarity of the measured voltage difference (H _{1 (2)} -H _{2 (1)} ) in the first comparator with a zero threshold, thereby identifying the combinations of bits 1/0 and 0/1, and to recognize the significance of the repeated bits, the absolute value of the voltage difference | H _{1 (2)} -H _{2 (1)} | and compare it with a threshold of 0.5H ₁ (H ₂ ) in the second comparator, which, in case the threshold is not exceeded, turns off the output of the first comparator, simultaneously connecting to the device output a record of the value of the last recognized bit stored in the shift register.