JPH0479497B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a C/N measuring circuit that measures carrier power to noise power ratio (C/N) on the receiving side of a digital communication system.

(Prior art) In digital communication systems, in order to evaluate the performance of a demodulator or the entire transmission path including the demodulator,
Bit error rate (BER) of the code decoded by the demodulator
The ratio of the energy per bit of transmitted information to the noise power density (E _b /N ₀ ) is defined as a figure of merit for the following equation (1).

E _b /N ₀ = C/N・B/R ...(1) Here, C/N is the carrier power to noise power ratio, B
is the equivalent noise bandwidth of the demodulator and R is the data transmission rate.

By the way, as is clear from equation (1), E _b /N ₀
To find this, it is necessary to measure the C/N. Therefore, for example, in the case of determining E _b /N ₀ in a satellite line, the conventional C/N measurement method uses a band pass at the front stage of the demodulator 72, which has a narrower band than the satellite repeater, as shown in FIG. A filter 71 is arranged to measure the C/N in the IF band of the input of the demodulator 72 into which the received modulated signal is input.

The measurement procedure begins with the bandpass filter 71.
The output power of the bandpass filter 71 is measured by transmitting an unmodulated wave or a modulated wave to the center of the band of C.
Find +N. Next, the output power of the bandpass filter 71 is measured by stopping the transmission of the unmodulated wave or the modulated wave, or by shifting the transmission carrier frequency so that the received signal is outside the band of the bandpass filter 71. demand. Then, by subtracting N from the C+N obtained earlier, C is obtained, and from both, C/
Find N. Note that B in this case is the equivalent noise bandwidth of the bandpass filter.

(Problems to be Solved by the Invention) However, the conventional C/N measurement method described above has the following various problems.

First, a bandpass filter whose accurate equivalent noise bandwidth is known is required for measurement, and a separate measuring device such as a power meter is also required.

In addition, in C/N measurement, unmodulated waves or modulated waves are transmitted, and operations to stop or shift the frequency are performed in the IF band, which not only makes the operation complicated, but also requires C/N measurement during operational conditions. Difficult to do.

Therefore, as a measure to deal with this problem and enable C/N measurement during operation, it is conceivable to estimate the power of the signal component and the noise component, respectively, from the received baseband signal sequence generated by the demodulator.

That is, if the signal component in the received signal is S(t) and the noise component is N(t), the noise component can be regarded as a Gaussian distribution, so the average () is ()=0. That is, by performing averaging processing on the received signal, the signal component S(t) can be extracted.

Note that since it is necessary to align the signal point positions, an operation is performed to obtain the absolute value of the received signal before the averaging process. Therefore, the signal power can be obtained by performing squaring processing on the averaged signal.
That is, P={()+()} ² = {()+()} ² = {()} ² ...(2) On the other hand, for the received signal {S(t)+N(t)} 2
After multiplication and averaging, the signal becomes {() + ()} ² = () ² + (
) ² +2()() =() ² +() ² +2(
)()……(3). In equation (3), the third term disappears due to the condition ( ) = 0, so the following equation (4) is obtained.

{()+()} ² = () ² + () ² ... (4) Thus, C/N can be determined from equations (2) and (4).

By the way, when the received signal is related to, for example, a binary digital signal, the probability density distribution of the received signal amplitude value on which Gaussian noise is superimposed is as shown in A and B of FIG. 3A.

That is, under high C/N conditions, the probability density distribution of the amplitude value of the received signal shows a Gaussian distribution like curve A centered at the signal point position S. In this case, even if the received signal at signal point position -S is returned to the signal point position S side by absolute value manipulation, the probability density distribution will be similar to that of curve A, so the received signal sequence after absolute value manipulation will be The average is approximately equal to the amplitude value at the signal point position S. On the other hand, under low C/N conditions, the noise power superimposed on the received signal increases, so the probability density distribution of the amplitude value of the received signal broadens at the base as shown in curve B and goes beyond the intermediate position 0 to the signal point position - It reaches close to S.

Then, when the received signal at the signal point position -S is returned to the signal point position S side by absolute value manipulation, the probability density distribution of the amplitude value of the received signal will change left and right around the signal point position S as shown in curve C. It becomes asymmetrical.
In other words, the average of the received signal sequence is different from the amplitude value at the signal point position S. As a result, the 6th
As shown in the figure, highly accurate measurements can be made under high C/N conditions, but when the C/N conditions drop below 7 dB, the values gradually deviate from the ideal value and the error increases.

In summary, if the power of the signal component and noise component are estimated from the receiver low-band signal sequence generated by the demodulator, C/N measurement during operation can be performed without the need for measuring instruments or complicated operations. enable.

However, since absolute value manipulation is essential for estimating signal power, there is a problem that under low C/N conditions, the influence of aliasing increases, resulting in poor measurement accuracy.

The present invention was made in view of these problems, and its purpose is to accurately measure C/N even under low C/N conditions when performing C/N measurement during operation.
An object of the present invention is to provide a C/N measurement circuit that can perform measurements.

(Means for Solving the Problems) In order to achieve the above object, the C/N measuring circuit of the present invention has the following configuration. That is, the C/N measuring circuit of the present invention measures the carrier power to noise power ratio (C/N) on the receiving side of a digital communication system.
A C/N measurement circuit for measuring the C/N ratio; an absolute value circuit that receives a received baseband signal sequence as input and takes its absolute value; a weighting circuit that performs predetermined weighting on the output of the absolute value circuit; a first averaging circuit that performs an averaging process on the output of the weighting circuit; a first squaring circuit that performs a squaring operation on the output of the first averaging circuit; and a first squaring circuit that receives the received baseband signal sequence as an input. a second squaring circuit that performs the squaring operation; a second averaging circuit that performs averaging processing on the output of the second squaring circuit; the first squaring circuit and the second averaging circuit; a C/N calculation circuit that calculates and outputs the ratio (C/N) upon receiving the output of;
It is characterized by having the following.

(Operation) Next, the operation of the C/N measurement circuit of the present invention configured as described above will be explained.

First, the absolute value of the received baseband signal sequence is taken in an absolute value circuit, and then a predetermined value is determined in a weighting circuit in order to eliminate the contribution of the portion where the influence of aliasing that occurs under low C/N conditions is the greatest. Subjected to weighting processing. Here, the weighting function is
For example, the following can be considered. If x is the output signal of the absolute value circuit and TH is the threshold, (1) W(x)=x (2) W(x)=1 x>TH =0 x≦TH (3) W(x)=1 x>TH =-α x≦TH (4) W(x)=x ^2The output of this weighting circuit is input to the first averaging circuit, where short-term fluctuations (noise components) are removed and signal components are Extracted. Now, S(t) is the signal component, N(t) is the noise component, and W(u) (where u=|
If S(t)+N(t)|) is the weighting function,
Since the output of the first averaging circuit is |()+()|(), the output of the first square circuit is |()+()|() ² ...(5). In equation (5), ()=0, ()・
Since W(u)0, |()+()|() ² ≒()・() ² ...(6) This is taken as the signal power () ² .

This becomes one input of the C/N calculation circuit.

On the other hand, the received baseband signal sequence is input to the second averaging circuit via the second square circuit, so the output of the second averaging circuit is (()+()) ² ...(7) . S(t) and N(t) are uncorrelated,
When formula (7) is rearranged taking into account that (t)=0, it becomes (() ² +() ² . . . (8). This becomes the other input of the C/N calculation circuit.

In this way, the C/N calculation circuit can calculate the C/N based on equations (6) and (8), and can output the measured C/N.

As explained above, according to the C/N measurement circuit of the present invention, weighting processing is performed on the signal subjected to absolute value manipulation to reduce the influence of aliasing that occurs under low C/N conditions, and then averaging processing is performed. By doing this, it is possible to improve the accuracy of estimating signal power under low C/N conditions.
N measurements can be made.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings.

FIG. 1 shows a C/N measuring circuit according to an embodiment of the present invention. In FIG. 1, 101 is an absolute value circuit, 102 is a weighting circuit, 103 is a first averaging circuit, 104 is a first square circuit, 106 is a second averaging circuit, and 107 is a C/N calculation circuit. be.

In the above configuration, the received baseband signal sequence is applied to the input terminal 1 and supplied to the absolute value circuit 101 and the second square circuit 105, respectively.

The absolute value circuit 101 performs absolute value manipulation on the input signal and outputs the output signal to the weighting circuit 102 .

The weighting circuit 102 is configured as shown in FIG. 2, for example. This implements weighting processing using the weighting function shown in (3) above, and is basically composed of a comparator 201, a multiplier 202, and a selector 203. The comparator 201 compares the levels of the output signal of the absolute value circuit 101 and the threshold value TH, and provides the result to the selector 203 as a selection control signal. For example, (output signal) > (threshold TH)
In this case, the selection control signal is set to "1" level, and in the opposite case, it is set to "0" level.

On the other hand, the multiplier 202 performs an operation on the output signal of the absolute value circuit 101 and the weighting constant -α, and supplies the result to one input of the selector 203 . As a result, in the selector 203, the output signal of the absolute value circuit 101 is directly supplied to the other input.
When the selection control signal is “1”, the absolute value circuit 101
The output signal of the multiplier 202 is output as is, and when the selection control signal is "0", the output of the multiplier 202 is selectively output.

In short, in this weighting circuit 102, when the received baseband signal sequence is related to a binary digital signal, as shown in FIG. A threshold value TH is set at a point where a suitable distance has passed, so that the distribution curve C becomes the distribution curve D under low C/N conditions. At this time, the weighting constant -α is appropriately set so that the contribution of the portion most affected by folding is reduced. Note that FIG. 3 shows the case where α=0.

The output of this weighting circuit 102 is input to the first squaring circuit 104 via the first averaging circuit 103, so the output of the first squaring circuit 104 is expressed by the above equation (6).
and the signal power () ² is obtained. This becomes one input of the C/N calculation circuit 107.

On the other hand, the received baseband signal sequence is transmitted to the second square circuit 1.
05 to the second averaging circuit 106, the output of the second averaging circuit 106 obtains the power shown by equation (8), which is then input to the C/N calculation circuit 10.
This is the other input of 7.

The C/N calculation circuit 107 basically consists of a subtraction circuit 401 and a conversion table 402, as shown in FIG. 403 is the output of the second averaging circuit 106. In the subtraction circuit 401, the subtraction process of () ² + () ² - () · () ² is performed from the above equations (6) and (8), and the signal 405 whose content is the noise power () ² is converted. is provided to table 402. A signal 404 is also input to the conversion table 402 . This conversion table 402 is ROM
10log (signal 404) - 10log (signal 405) is stored. In other words, the signal 404 and the same 40
5 as the address, a signal 406 containing the ratio (C/N) is output to the output terminal 7.

Finally, FIG. 5 is a characteristic diagram obtained with threshold value TH=0.25 and weighting constant -α=-0.5.

As is clear from the comparison with FIG. 6, the approximation accuracy has improved to 4 dB even under low C/N conditions.

(Effects of the Invention) As explained above, according to the C/N measurement circuit of the present invention, weighting processing is performed on the signal subjected to absolute value manipulation to reduce the influence of aliasing that occurs under low C/N conditions. Since averaging processing is performed later, the accuracy of estimating signal power under low C/N conditions can be increased, and even under low C/N conditions, C/N can be accurately calculated.
N measurements can be made.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of the configuration of a C/N measuring circuit according to an embodiment of the present invention, Fig. 2 is a block diagram of the weighting circuit, and Fig. 3 shows the received signal amplitude value when Gaussian noise is superimposed. Probability density distribution diagram, Figure 4 is a block diagram of the configuration of the C/N calculation circuit, Figure 5 is a characteristic comparison diagram when weighting processing of the present invention is added, Figure 6 is a diagram of absolute value manipulation and averaging processing. FIG. 7 is a characteristic comparison diagram for estimating signal power. FIG. 7 is a block diagram of a conventional C/N measurement circuit. 71... Band pass filter, 72... Demodulator, 101... Absolute value circuit, 102... Weighting circuit, 103... First averaging circuit, 104... First
square circuit, 105...second square circuit, 106
... Second averaging circuit, 107 ... C/N calculation circuit, 201 ... Comparator, 202 ... Multiplier, 20
3...Selector, 401...Subtraction circuit, 402...
...conversion table.

Claims (1)

[Claims] 1 C/N measuring the carrier power to noise power ratio (C/N) on the receiving side of a digital communication system
N measurement circuit; an absolute value circuit that receives a received baseband signal sequence as input and takes its absolute value; a weighting circuit that performs predetermined weighting on the output of the absolute value circuit; a first averaging circuit that performs an averaging process on the output; a first averaging circuit that performs a squaring operation on the output of the first averaging circuit;
a second squaring circuit that receives the received baseband signal sequence as input and performs a squaring operation; a second averaging circuit that performs averaging processing on the output of the second squaring circuit; ; the first square circuit and the second
A C/N measuring circuit comprising: a C/N calculation circuit that receives an output of an average circuit, calculates and outputs the ratio (C/N);