WO2002049262A1

WO2002049262A1 - Method for determining channel quality in digital transmission systems, in which the bit error rate is re-measured over a greater bit length, depending on the sum of the bit-noise ratio and the coding gain

Info

Publication number: WO2002049262A1
Application number: PCT/AT2001/000391
Authority: WO
Inventors: Robert Messaros; Hermann Wolf
Original assignee: Siemens Ag Österreich
Priority date: 2000-12-14
Filing date: 2001-12-10
Publication date: 2002-06-20
Also published as: AT410385B; ATA20822000A

Abstract

According to a method for determining the channel quality in a digital transmission system, both the bit error rate (BER) over a first number (n1) of incoming bits and the bit-noise ratio (E0/N0) of the output (E0) per bit to the spectral noise output (N0) are measured in a digital receiver. If the sum of the bit-noise ratio (Eb/N0) and the coding gain (CG) is below a lower limit (Vu) or above an upper limit (V0), the channel quality is determined to be inadequate or faultless. If this sum is between the lower and upper limits, the bit error rate (BER) is re-measured, over a second number (n2) incoming bits, this number being greater or less than the first number (n1).

Description

METHOD FOR DETERMINING CHANNEL GOODS IN DIGITAL TRANSMISSION SYSTEMS, WOBEI IN

DEPENDENCY OF THE SUM FROM BIT NOISE RATIO AND ENCODING GAIN THE BITFEHLER R RATE IS RE-MEASURED OVER A LARGER BIT LENGTH

The invention relates to a method for determining the channel quality of channels of a digital transmission system and to an apparatus for performing the method.

In the digital transmission of data via potentially unsafe transmission media, e.g. B. satellite data connections, close monitoring of the quality of the transmission process is required. This "quality of service" is specified as the bit error rate BER, namely as a proportion of those bits whose value has changed during the transmission.

Postal companies, satellite providers, cable television companies and terrestrial broadcasting companies use surveillance systems that have to monitor the bit error rate of hundreds of digital channels. Every reduction in the measurement time required leads to a reduction in costs, since fewer measurement instruments are then required, the operating costs are reduced and the bit error rate can be measured more often.

According to the prior art, the bit error rate is measured by transmitting a known data sequence during a certain observation time and comparing the received data with this sequence. In order to obtain a reasonable estimate of the bit error rate, the measurement time must be long enough to guarantee that enough incorrectly transmitted bits reach the receiver so that a sufficiently accurate estimate of the bit error rate can be achieved.

For example, if the bit error rate is 10 ⁶ , the receiver has to wait an average of 10 ⁶ bits in order to recognize a single wrong bit. Unfortunately, however, defective bits are not evenly distributed over time, but rather often occur in bursts or series of incorrect bits followed by relatively long, error-free sections. As a basis for measurements, it is assumed that at least ten wrong bits must be determined by the receiver. However, the time required for this depends on the quantity to be measured. The measurement time must therefore either be chosen so long that the lowest desired bit error rate can be measured or a noticeable percentage of the measurements will not result in any faulty bits at all or too few bits in order to arrive at a reliable estimate of the bit error rate.

An improvement in the time required for the measurements can alternatively be obtained by changing the expected magnitude of the bit error rate underestimated. There is a constant relationship between the desired signal strength and the undesirable but unavoidable noise that is added by the transmission medium. Extremely expensive devices, e.g. B. Spectrum Analyzers with prices from around EUR 20,000 can be used to measure the carrier-to-noise ratio CNR (Carrier to Noise Ratio), which indicates the ratio between the desired signal power and the unwanted noise power. Using the carrier / exchange ratio CNR, the bit error rate can be estimated using theoretical curves. For example, FIG. 1 relates to digital satellite video radio DVB-S (Satellite Based Digital Video Broadcasting), in which QPSK modulation (Quadrature-phase Phase Shift Keying) is used. The relationship between the bit error rate BER and the carrier / noise ratio is shown in FIG. 1, the latter ratio being plotted linearly in dB and the bit error rate logarithrristically in the range from 10 to ¹⁰ to 1.

However, as I said, this solution requires expensive instruments so that it cannot always be used, especially when cost issues are at the forefront.

Recently, however, receivers for digital signals have been created which are integrated on individual silicon chips and offer the functionality required here. Such receivers are obtained for digital video broadcasting DVB (digital video broadcasting), for digital audio broadcasting DAB (digital audio broadcasting), digital satellite radio DSR (digital satellite radio), VSAT (very small aperture terminal) and the like. These chips are relatively cheap and the price of complete receivers with such chips is in the range of EUR 1,000. They offer a very precise measurement or estimation of the carrier / noise ratio as well as the bit error rate, so that the above-mentioned relationships can be used efficiently.

The ETSI standard ETS 300421 describes, for example, the air interface for satellite-based DVB and the data backup strategy used. The knowledge imparted to the person skilled in the art is assumed here.

An object of the invention is to use the a priori knowledge of the carrier / noise ratio - based on a "Variance Per Sample" evaluation in the chip - to achieve an estimate of the bit error rate without additional devices.

This object is achieved with a method of the type mentioned at the outset, in which the following steps are taken according to the invention: a) Measuring both the bit error rate over a first number of incoming bits and the bit-to-noise ratio of the power per bit to the spectral noise power in a digital receiver.

b) ba) if the sum of the bit-to-noise ratio and the coding gain is below a lower limit, determine the channel quality as insufficient and the measured bit error rate as sufficiently precise, bb) if the sum mentioned is above an upper limit, determine the channel quality as error-free if the measured bit error rate is not observed, and bc) if the said sum lies between the lower limit and the upper limit, at least one more measurement of the bit error rate in the same channel, but over a second number of incoming bits, this second number being greater than the first number is, and determining the channel quality as a function of the new measurement result of the bit error rate.

The method according to the invention enables the measurement times in practice to be reduced by a factor of 50 in an extremely cost-effective manner, as a result of which the monitoring and thus the quality of existing connections can be significantly increased.

A refinement of the method can be achieved in that the area between the lower and the upper limit is divided by a third, middle limit, and the measurement in step bc) is carried out over a third number of incoming bits if the said sum between the is lower and the middle limit, this third number being greater than the first number, but the measurement in step bc) is carried out over a fourth number of incoming bits if the said sum lies between the middle and the upper limit.

Sufficient accuracy in practice with a measurement time that is not too long results if, in step bc), the measurement is repeated over a number of bits, which is derived from the theoretical bit error rate, which is due to the known relationship between the bit error rate and the ratio for the type of modulation used given is. It is expedient if the number of bits in step bc) is of the order of 10 =

BER will.

The first number of incoming bits is advantageously in the order of 10 ⁵ to 10 ⁶ and the second number is in the order of 10 ⁷ to 10 ⁸ . The Viterbi bit error rate is used advantageously and in many cases as the bit error rate.

A device with a receiver connected to an antenna device via a first interface is particularly suitable for carrying out the method, with a control unit which can be connected to the receiver via a second interface for inputting the channel frequency, the coding rate, the channel symbol rate and the number of bits to be measured into the receiver and with a database, which can be connected to the control unit via a third interface, for reading the channel center frequency, the coding rate, the channel symbol rate and the number of bits to be measured into the receiver.

The invention is explained in more detail below using an exemplary embodiment with reference to the drawing, in which show

1 shows the relationship between the bit error rate and the carrier / noise ratio for QPSK in a diagram,

2 shows the relationship between the channel bit error rate and the ratio Eb / NO in a further diagram,

3 shows a graphic representation of decision criteria for the method according to the invention,

Fig. 4 shows an example of a method according to the invention on a flow diagram, and

Fig. 5 in a simplified block diagram an apparatus for performing the method according to the invention.

The example in which the invention is explained in more detail here is based on a satellite-based, digital video broadcasting. An attempt is being made to measure or estimate the channel bit error rate CBER and the Viterbi bit error rate VBER in an adaptively intelligent manner. In this way, a meaningful and precise determination of the channel bit error rate and the Viterbi bit error rate is achieved, saving measurement time and measuring instruments as well as switching devices. The example shown is adapted to the performance and properties of modern satellite-based digital video broadcast receivers (DVB-S = Satellite Based Digital Video Broadcasting). The solution is also to use the heuristic knowledge of, for example, QPSK modulation and the satellite fading channel, with line of sight, even channel impulse behavior and the absence of significantly strong echoes. The basic principle is as follows: An estimate of the current channel conditions of the DVB-S with regard to the τp is obtained

Ratio - -, where E is the energy per bit and No is the (one-sided) spectral noise level ^• ** o τp tung, namely with k * T _sys t. This knowledge of the - ^ - conditions will

exploited to save measurement time, which will be explained further below.

P P

Based on these - - estimates, three areas are distinguished from -:

.tau..sub.p

A range of bad - - conditions. A short measuring time is sufficient for the bit error

No lerrate in order to obtain sufficiently precise values for the channel bit error rate and the Viterbi bit error rate, since enough Viterbi bit errors occur in a short time, which are determined by the DVB-S receiver. .tau..sub.p

A range of excellent - - conditions. It is believed that Viterbi bit errors

No hardly ever can be determined by the measuring instrument. This is because conventional DVB-S receivers have an upper count limit of no more than 10 ⁸ data bits. You can assume that the channel is working with virtually no errors, and you don't even try to measure the exact Viterbi bit error rate or channel bit error rate.

A range between the extreme values given by 2a and 2b. It is assumed here that the quality of the channel is still acceptable and that it pays to invest measurement time for measuring the Viterbi bit error rate (and the channel bit error rate), since the DVB-S receiver is likely to have a sufficiently high number of errors during the Finds measuring process. Only in this case does it make sense to invest the maximum measurement time for measuring the Viterbi bit error rate or the channel bit error rate.

The following detailed discussion of the method should take into account the data given in Tables 1 and 2, together with the fact that the current relationship

P nis - - usual DVB carrier is in the range above 7.5 dB, which, for. B. for the DVB-S

^ o channels of the EUTELSAT Hotbird satellite (13 ° East) is realized.

.tau..sub.p

Due to the fact that the ratio - - and the carrier / noise ratio CNR

- * ^» o are used in the same context, their context should be given as follows. Eb energy per bit

No (one-sided) spectral noise power, equal to k * T _sys tk Boltzmann constant

T _S yst noise temperature of the system

B Receiver bandwidth

Rb bit rate, for QPSK: R _b = 2 * B

^J b _ C B_ B

No; R, vNy 2 * B 2

, C _ Carrier _ spectral power density of the signal

N noise spectral power density of the noise

CG coding gain

The coding gain CG (coding gain) describes by how many dB the signal-to-noise ratio can be worse in order to obtain the same bit error rate BER as with an uncoded, e.g. B. with Niterbi coding transmission method. For example, DVB-S uses QPSK (Quadrature Phase Shift Keying) as the transmission method, which - as can be proven theoretically - has the same error rate as FSK (Frequency Shift Keying). See in this regard: M. Schwartz, W. R. Bennet, S. Stein, "Communication Systems and Techniques", McGraw-Hill Book Company, 1986, page 299, Fig. 7-5-1 "Error Rates for Several Binary Systems".

Table 1: Dependence of CBER, VBER and remaining BER on - and CΝR, measured values

I ts

e,

Table 2: Required ratio - - and CNR for VBΕR = 10 " ⁶ for different values of FΕC.

.tau..sub.p

The improvement, expressed by the required - - or CNR becomes coding gain

(Coding Gain) called. This can be obtained from the data sheets of the chipset from a DVB receiver manufacturer. The coding gain can be used to obtain the Viterbi bit error rate VBΕR from the channel bit febler rate CBΕR. For example, the coding gain, if an FΕC = - is used, for CBΕR = 10- ⁶ (see Table 2).

13.5 dB - 7.8 dB = 5.7 dB

The coding gain is used in the method described in more detail below.

The (theoretical) behavior of the bit error rate BΕR of a coherently demodulated QPSK (corresponding to the channel bit fever rate for digital video broadcasting DVB) is given by:

erfc (x) = complementary Fe function

In practical cases, the bit error rate BER of a real receiver will be 0.5 dB to 1 dB worse than the theoretical curve for bit error rates around 10- ⁶ . In Fig. 2 the channel bit

EE error rate CBER above the ratio - ₇ - - on, again the - - ratio

N 'n0 N ^J, n0 is given in dB and is plotted linearly, whereas the channel bit febler rate moves logarithmically between 10- ¹⁰ and 1.

In other signal processing or modulation methods, such as. For example, "64 QAM" (Quadrature Amplitude Modulation) or DSSS (Direct Sequence Spread Spectrum), the corresponding theoretical relationships between bit error rate and Eb / Νo are used, whereby it is generally not the Viterbi bit error rate. The only requirement for the application of the invention is that the receivers used allow the determination of Eb / Νo.

Table 3 below shows the measurement times required for the EUTELSAT DVB channels.

Table 3

The method for determining or estimating the channel quality for satellite-based digital video broadcasting DVB-S thus uses the following steps, which are also shown in FIG. 3.

P a) determination, namely measuring the ratio - -, this determination by

No the DVB-S receiver is delivered. For this purpose, the DVB-S receiver is programmed to eg 10 ⁵ or 10 ⁶ data bits for the Viterbi bit error rate

C

To use calculations. The time it takes to complete one

Obtaining an estimate corresponds to the time it takes to read out the Viterbi bit error rate, which can be obtained within less than 0.2 seconds even for the carrier with the lowest character rate (see Table 3 above). At the same time, the bit error rate BER is also measured using the receiver. τp aa) If the sum of the ratio - and the coding gain CG, therefore = Eb / Νo +

No

CG is less than 9 dB (in practice less than 9.5 dB if the system accuracy of + _. 0.5 dB is taken into account), ie below a lower limit v _n (see FIG. 3) it is assumed that Viterbi bit error rate is greater than 10. ⁵ This follows from the theoretical relationship between bit error rate and Eb / No for the type of modulation used. In this case, the channel quality is rated as insufficient and no further measurements are carried out. The measurement carried out in step 1, which uses 10 ⁶ data bits, already gives sufficient accuracy for the channel bit error rate and the Viterbi bit error rate. τp bb) If the sum of the estimated ratio - - and the coding gain is greater

is greater than 11 dB (in practice greater than 11.5 dB, taking into account the system accuracy of + 0.5 dB, it is assumed that the Viterbi bit error rate is below 10 " ^7. In order to provide meaningful measurement results in this area would require more than 10 ⁸ data bits for the entire measurement process, which currently and in the foreseeable future cannot be supplied by DVB-S receivers available on the market, in which case the channel is considered to be error-free and the one measured by the receiver Bit error rate is ignored. Τp bc) If the sum of the estimated ratio - - and the coding gain between

see 9 dB and 11 dB, an additional measurement of the channel bit error rate and the Viterbi bit error rate is carried out. The DVB-S receiver is programmed to use 10 ⁷ data bits for its measurements, if

S = ^ - + CG <10 dB,

No τp, however, is 10 ⁸ data bits for S = - - + CG> 10 dB. The new measurement

result obtained for the bit error rate BER now offers the basis for the assessment of the channel quality.

In the measurements in step bc, the number of bits over which measurement is carried out, namely n ₂ , ri3 or r, is derived from the theoretical bit error rate BER or CBER, which is derived from the known relationship between bit error rate and noise ratio CNR or Eb / No (see Fig. 1, 2) for the type of modulation used.

In practice you go z. For example, do the following: If the assumed bit error rate BER = 10 " ⁶ , its reciprocal value gives 10 ^6. If a spread of 10 is taken into account, a number of bits of 10 ^{7 is obtained} for the measurement to be carried out.

Only the measurements with 10 ⁸ data bits take a little longer than if one had worked with 10 ⁸ data bits right from the start, whereas all other measurements require significantly less measurement time. For this reason, the measuring devices used be used significantly more efficiently. Qualitatively poor DVB channels will predominantly correspond to states according to point aa), although it can be assumed, however, that practically all DVB channels are error-free most of the time, this means the state of points bb and bc. In addition, most channels with S> 10 dB will be practically error-free.

Fig. 5 shows a practical device, with the help of which the method according to the invention can be carried out.

An antenna SEA, here a satellite antenna, can receive signals from a transmitter SAT, here a communications satellite, with the aid of a preamplifier LNB, the latter providing the input signal SE for an receiver RCR at an interface SCI.

Furthermore, a control unit STE is provided which is connected to the receiver RCR via a further, second interface SC2 or to a database DAB via a third interface SC3.

Via the second interface SC2 between the receiver RCR and the control unit STE, the channel center frequency, the coding rate, namely the measure for the FEC ("forward error correction"), the channel character rate and the number n of the incoming bits to be measured can be set; at this interface SC2, among other things, the bit error rate BER and the bit-noise ratio Eo / N can be read out o.

The database DAB contains all the data required for the settings on the receiver RCR, which is therefore not entered manually into the receiver each time, but is read into the control unit STE from the database DAB via the third interface SC3 and then fed to the receiver. These are the channel center frequency, the FEC, the channel symbol rate and the number of bits n.

Claims

PATENT CLAIMS

1. Method for determining the channel quality of channels of a digital transmission system, characterized by the following steps:

a) Measuring both the bit error rate (BER) over a first number (nl) of incoming bits and the bit-to-noise ratio (Eo/Νo) of the power (Eo) per bit to the spectral noise power (Νo) in a digital receiver.

b) ba) if the sum of the bit-noise ratio (Eb/Νo) and the coding gain (CG) is below a lower limit (v _u ), determining the channel quality as insufficient and the measured bit error rate (BER) as sufficiently accurate, bb) if the said sum is above an upper limit (v _σ ), determining the channel quality as error-free if the measured bit error rate (BER) is ignored, and bc) if the said sum is between the lower limit (v _u ) and the upper limit ( v ₀ ), at least one new measurement of the bit error rate (BER) in the same channel, but over a second number (n ₂ ) of incoming bits, this second number (n ₂ ) being greater than the first number (ni), and Determination of the channel quality depending on the new measurement result of the bit error rate.

2. The method according to claim 1, characterized in that the area between the lower (v _u ) and the upper (v ₀ ) limit is divided by a third, middle limit (v _m ), and the measurement in step bc) over a third number (113) of incoming bits is carried out if the said sum lies between the lower (v _u ) and the middle (vm) limit, this third number (113) being greater than the first number, but the measurement in step bc ) is carried out over a fourth number (ru) of incoming bits if the sum mentioned lies between the middle (v _m ) and the upper limit (v ₀ ).

3. The method according to claim 1 or 2, characterized in that in step bc).

ben is.

1. The method according to claim 3, characterized in that the number (n ₂ ) of bits in step bc) is chosen in the order of 10.

5. Method according to one of claims 1 to 4, characterized in that the first number (ni) of incoming bits is in the order of magnitude of 10 ⁵ to 10 ⁶ and the second number (n ₂ ) in the order of magnitude of 10 ⁷ to 10 ⁸ .

6. Method according to one of claims 1 to 5, in which the Viterbi bit error rate is used as the bit error rate.

7. Device for carrying out the method according to one of claims 1 to 6 with a receiver (RCR) connected to an antenna device (SEA, LNB) via a first interface (SCI), with a receiver (SC2) connected via a second interface (SC2). RCR) connectable control unit (STE) for entering the channel center frequency, the coding rate (FEC), the channel character rate and the number (n) of bits to be measured into the receiver and with a database (DAB), which is connected via a third interface (SC3). the control unit can be connected to read the channel center frequency, the coding rate (FEC), the channel character rate and the number (n) of bits to be measured into the receiver.