ITMI970768A1 - PROVISION TO FIND A DISTORTION CRITERION FROM AN M-QAM SIGNAL - Google Patents

Description

DESCRIPTION

State of the art

An arrangement for deriving a distortion criterion from a demodulated and quadratic amplitude modulated signal (M-QAM, M = 4 ... 256) is known for example from DE 41 34 206 CI. In accordance with the current state of the art, in receivers for M-QAM signals, a distortion criterion is usually obtained from the deviation of the signal value, estimated by a symbol decision-making body and sent with the maximum probability, of the sampling values of the demodulated complex signal M-QAM that occur at the adaptive equalizer output of the baseband. A distortion criterion thus obtained therefore depends on the state of the adaptive equalizer of the baseband. If the distortion criterion is to be used for example to control a Diversity combiner, then for a distortion criterion, detected in accordance with the current state of the art, a coupling would occur between the regulation loop for the adaptive baseband equalizer. 'and the regulation circuit for the Diversity combiner. However, it is known that coupled control loops can be controlled with great difficulty, for example as regards the setting of control time constants.

The invention therefore has the task of indicating an arrangement for obtaining a distortion criterion not oriented to an adaptive band-pass equalizer.

Advantages of the invention

This problem is either solved by the features of claim 1, in that the arrangement forms products of sampling values of the complex demodulated M-QAM signal and the sampling values of the complex and conjugately demodulated M-QAM signal, which are out of phase with respect to the first-mentioned sampling values due to a symbolic rhythmic pulse, furthermore due to the fact that the arrangement from multiple products of sampling values forms an average value that provides the distortion criterion. Advantageously, in accordance with the sub-claims 2 and 3, it is possible to form the amount or the square of the average value.

An alternative solution to the problem posed results from claim 4. On the basis of this, the error signals are obtained by forming a difference between sampling values of the unequalized, complex and demodulated M-QAM signal, and the symbol values estimated in relation to these values sampling. From the squared amounts of several error signals an average value is then formed providing the distortion criterion.

Preferably the distortion criterion obtained according to claim 1 or to claim 4 is used to control a Diversity combiner.

The distortion criterion revealed according to claim 1 has the advantage that it is independent of a carrier frequency offset obtained in the receiver. It is also independent of the phase of the carrier. Both arrangements of claims 1 and 4 can be implemented as digital circuits.

Description of examples of implementation

On the basis of two embodiments shown in the drawing, the invention is illustrated in detail below.

In particular:

Figure 1 shows a schematic diagram of a transmission system for M-QAM signals,

Figure 2 shows a schematic of a first distortion detector, and figure 3 shows a schematic of a second distortion detector.

As can be deduced from the principle diagram represented in figure 1 and relating to a transmission system for M-QAM signals, a transmitter SD produces a signal S1 (t) in accordance with equation 1:

av indicates the symbol of complex data sent at the instant t = vT. T is the rhythmic pulse symbol and p (t) represents the pulse response of the side-transmission filtration. The properties of the transmission channel are detected by the HD transmission function (f). An additional noise component n (t) is added to the reception signal S2 (t) at the receiver input. n (t) takes into account the input noise (noise factor) of the receiver. An AGC1 automatic level adjustment with the gain factor adjusts the received signal to a constant signal level. HE (f) describes the transmission function of the receiver including the demodulator with carrier frequency derivation and rhythmic pulses. The demodulated signal S4 (t) has the form indicated in equation 2:

k (t) is the impulse response relative to the transmission function k (f) given in equation 3:

The signal S4 (t) is sampled in the sampling circuit A with the rhythmic pulse symbol T. From this the sampling values S5 [n] are obtained in accordance with equation 4 at instants t = n T

fca:

ta indicates in particular a phase shift of the sampling instant which depends on the revenue properties of the rhythmic pulses. After sampling

there is a discrete time and value signal, since sampling is usually done by means of an analog-to-digital converter.

A second AGC2 level adjustment with the gain factor Ύ2 scales the sampling values S5 [n] appropriately, from which the sampling values S6 [n] are obtained according to equation 5:

The main value of the impulse response is k (ta) and the relation n2 (nT + ta) = Y2n1 (nT + ta) holds. The second AGC2 level adjustment ensures that equation 6 is satisfied:

A phase adjustment of the carrier present in the receiver causes a phase rotation, so that the imaginary part of the main value. of the impulse response disappears according to equation 7:

if there are no distortions and ta has been chosen optimally, then the channel impulse response applies to the sampling values:

A "robust" distortion criterion VB1 or VB2 is obtained from the sampling values of the signal S6 [n] in a circuit block VD1, described in detail below, for which the carrier phase adjustment must not necessarily be inserted and not even an adaptive equalizer will necessarily have to provide reliable symbol decisions. The discrete time signal S6 [n] is equalized in an adaptive EZ baseband equalizer. This adaptive EZ baseband equalizer is not necessary to derive the distortion criteria in the circuit blocks VD1 and VD2. A discrete time equalized signal y [n] is applied to the output of the adaptive equalizer of the baseband EZ. A symbol decision maker SE decides on which symbol an-i from the entire alphabet of symbols M-QAM corresponds in a maximally probable way to the sampling value y [n] applied in the decision-making body SE of symbol. oral i in the decided symbol an-i takes into account the delay of the decisions of

symbol, depending on the EZ adaptive baseband equalizer.

Figure 2 shows a distortion detector VD1 which derives the '' robust "distortion criterion VB1 or VB2 from the sampling values S6 [n]. As long as the carrier in the receiver is not inserted on the carrier frequency of the receiving signal it still exists a frequency offset Δ f. Taking into account the frequency offset, the signal Sg [n] indicated in equation 5 takes the form indicated in equation 8:

From this complex signal S6 [n] a distortion criterion is obtained in the following way. The product is formed between an nth complex sampling value S6 [n] and a conjugated complex sampling value Sg [n-1] out of phase by a symbol.

As shown in Figure 2, each complex sampling value S6 [n] has a real part and an imaginary part. The real part of the delayed S6 [n-1] complex sampling value

by reason of a symbol, it is applied to the output of a delay member ZR and the imaginary part of a complex sampling value S6 [n-1] delayed by reason of a symbol, is applied to the output of a delay member ZE. The product (Re {Sg [n]} jlm {S6 [n]}). (Re {S6 [n-l]} -jIm (S6 [n-1]}) is realized with the multipliers MR1, MI1, MR2, MI2, M and the adders Λ1, A2, A3. By forming an average value through several products of sampling value, the expected value represented in equation 9 is obtained:

Cr2 indicates the predicted value ava * v- The sampling values of the noise due to the time lag in relation to the duration of the symbol T are statistically independent, so that the predicted value <n2 (nT + ta) n2 * ([nl] T + ta)> disappears, and thus the equation is obtained

10:

From the expected value <S6 [n] .S * 6 [n-1]>, applied to the output of the formator of average value MW or by forming the amount or by squaring it is possible to obtain the distortion criteria VB1 or VB2 indicated in equation 11 or 12:

VB1 = | <S6 [n] .S6 * [n-1]> 1 (11) VB2 = <S6 [n] .S6 * [n-1]> 2 (12) while the predicted value in equation 10 depends again from the frequency offset Δ f of the carrier phase adjustment, this dependence disappears by forming the amount or squaring according to equations 11, 12.

The real part is applied to the output of the adder A1 while the imaginary part of the product S6 [n] * S6 * [n-1] is applied to the output of the adder A2. An MWR mean value former forms the mean value of the real part and a MWI mean value form former forms the mean value of the imaginary part of the whole product. At both average values, either the amount or the square is formed in the QB circuit block to obtain either the distortion criterion VB1 according to equation 11 or the distortion criterion VB2 according to equation 12.

When the sampling values Sg [n] do not exhibit internal symbol interference (distortion), the products

T + ta) do not contribute a

and therefore to the respective distortion criterion VB1 or VB2. If, on the other hand, the sampling values S6 [n] are distorted then the distortion criteria VB1 or VB2 are higher than zero.

The distortion detector VD2 represented in Figure 3 by the error signal eìsi (n) represented in equation 13 derives a distortion criterion VB3 commanded decisionally:

The index i (i> 0) appearing in equation 13 takes into account the delay of the symbol decision by the EZ band-pass adaptive equalizer. The error signal eisi (n) is obtained by difference formation of the two complex signals

This

the difference is effected in the circuit shown in Figure 3, by means of the adders AR and AI, wherein delay devices ZR1 and ZII comprising the real part and the imaginary part of the complex signal S6 [n] in terms of time i. The multipliers MR and MI that follow the adders AR and AI form the squares of the real part and the imaginary part of the signal difference, and a subsequent adder A forms the sum of the real part squared and the imaginary part squared of the signal. differential. This sum signal corresponds to the square of the amount of the error signal.A mean value former MW1 from the square of the amount of the error signal forms a predicted value like that represented in equation 14:

This expected value of the error signal represents the distortion criterion VB3 and can be deduced from equation 15.

Due to the level adjustment (Ύ1 Ύ2) Re (k (ta)) = 1 and the phase adjustment of the

also applies

gold in equation 15

σ-2n is the variance of the noise. The value of <| e ^ s ^ [n] 2> is dominated in case of strong distortions

predominantly from the term ^

The noise component σ-2n, on the other hand, plays only a secondary role.

The distortion criteria VB1, VB2 or VB3, obtained in accordance with the methods described above, are especially suitable for controlling a Diversity combiner, in fact the distortion criteria VB1, VB2 and VB3 are independent of the adjustment quantities of an adaptive band-pass equalizer. There is therefore no coupling of the control loop for the diversity combiner and the control loop for the adaptive band-pass equalizer. Distortion components not adjusted by the Diversity combiner can also be further minimized independently by an adaptive band-pass equalizer.

Claims (5)

CLAIMS 1. Arrangement to derive a distortion criterion from a demodulated signal, modulated in quadratic amplitude (M-QAM, M = 4 ... 256), characterized by the fact that it forms products from signal sampling values (M-QAM) complex demodulated and conjugated complex demodulated signal (M-QAM) sampling values, which are out of phase with respect to the first-mentioned sampling values due to a symbolic rhythmic pulse, as well as by the fact that from multiple products of sampling value it forms a mean value providing a distortion criterion (VB1, VB2). Arrangement according to claim (1), characterized in that it forms the amount of the average value. Arrangement according to claim (1), characterized in that it forms the square of the mean value. 4. Arrangement to derive a distortion criterion from a demodulated signal and modulated in quadratic amplitude (M-QAM, M = 4 ... 256), characterized by the fact that it derives the error signal by forming a difference between the sampling value of the unequalized, complex, and demodulated signal (M-QAM), and the symbol values estimated with reference to these sampling values, which a symbol decision maker (SE) derives from the sampling values of the demodulated complex signal (M-QAM ), as well as by the fact that from the squared amounts of several error signals it forms an average value that provides a distortion criterion (VB3). Arrangement according to one of the preceding claims, characterized in that the distortion criterion (VB1, VB2, VB3) obtained therefrom is used to control a Diversity combiner.