WO2008087004A1

WO2008087004A1 - Use of provisionally decided symbols in a quadrature receiver

Info

Publication number: WO2008087004A1
Application number: PCT/EP2008/000250
Authority: WO
Inventors: Christian Bock
Original assignee: Micronas Gmbh
Priority date: 2007-01-16
Filing date: 2008-01-15
Publication date: 2008-07-24
Also published as: WO2008087004A8

Abstract

The invention relates to a method and a circuit arrangement for a decision on a symbol (D) on reception of received symbols (S) which are coupled to a quadrature signal pair (I, Q) in which the symbol (D) is associated with a specific nominal point of a multiplicity of nominal points (P (x, y) ) on a plane (x, y) which is covered by the quadrature signal pair, and each of the nominal points is located on a nominal radius of a multiplicity of nominal radii (Rl - R31), with at least one such nominal point (P (x, y) ) being located on each of the nominal radii, and with a provisional symbol (Dr) being decided for one reception point (xe, ye, µ) of the received symbol (S) on at least one predetermined defined auxiliary radius (Rh), with the auxiliary radius being selectable as a function of the nominal radii of the multiplicity of nominal points (P (x, y) ). In this case, it is advantageous if the provisional symbol (Dr) is decided at a position at the angle (α) of the reception point (Pe) of the received symbol (S) on the auxiliary radius (Rh).

Description

Method and circuit arrangement for deciding a symbol in the complex phase space of a quadrature modulation method

The invention relates to a method for deciding a symbol when receiving a signal coupled to a quadrature signal according to the preamble features of claim 1 or a circuit arrangement for carrying out such a method.

Recovery of a sample clock, including clock, sample timing, and sample rate recovery when receiving a signal coupled to a quadrature signal, often occurs according to a KH Mueller and MS Muller correlation method is described in "Timing Recovery in Digital Synchronous Data Receivers", IEEE Trans. Com., Vol. COM-24, No. 5, 516-531, May 1976 ". In this correlation method, energy components of a symbol are looked at with temporally adjacent signals (English, inter symbol interference). This method has the advantage of requiring only signals at the simple symbol rate, whereas many other methods use twice the symbol rate.

Fig. 11 shows one suitable for implementing the method

Circuit arrangement in which as an introductory component, an A / D converter 1 is shown. In the A / D converter, a clock signal t and a received received analog signal sa are input. The A / D converter 1 converts the analog reception signal sa into a digitized or digital signal sd. The digital signal sd is output to a quadrature converter or mixer 2, which converts the digitized or digital signal sd into baseband and splits it onto the two quadrature components I, Q. The signal output from the mixer 2 is applied to a gain controller 3 in a manner known per se. A signal amplified by the gain controller 3 is applied to a low-pass filter 4 supplied, the output signal of which is supplied to a sampling device 5, at which a sampling clock ti is applied for sampling. A signal sampled with the sampling device 5 is output to a further low-pass filter 6, preferably a Nyquist filter, and filtered by the latter and applied to an E-qualifier 7 (German: equalizer). The equalizer 7 outputs received signal values or received symbols S to a decision maker 8, which decides a symbol and outputs a decided symbol D. The decided symbol D is output on further circuits for further data processing, as is known per se. A circuit 9 for determining a phase difference Δφ can be applied to both the received signal value and the received symbol S and the ent ^¬ different symbol D in order to determine therefrom the standing between these loading phase difference Δφ. The phase difference Δφ is applied to a frequency control device 10, FC, which is designed as a carrier control device. The frequency control device 10 controls, in a manner known per se, a local oscillator 11, which outputs two carriers offset by 90 ° to the mixer 2, so that the mixer 2 receives the digital signal sd to provide the two quadrature components I, Q receives. In particular, these components can be implemented in a known manner and in a different order and configuration. Modifications in a manner known per se, however, can also be implemented with regard to the other components.

Both the signal values or symbols S output by the equalizer 7 and the decided symbols D output by the decision maker 8 are also applied to a clock control device 14. From this, the clock control device 14 determines, in a manner known per se, the sampling clock ti which is applied to the scanning device 5. In addition, the decided symbols D are applied as another input signal to the equalizer 7 for processing in a manner known per se. Fig. 12 shows a Mueller-Müller correlator for detecting errors at the sampling time. Shown are the block which forms the clock control device 14 in FIG. 11, as well as the decision maker 8 (English, slicer or data slicer). The undecided signals are correlated with the decided symbols D of the neighboring time. A subsequent low-pass filter LP generates a control voltage for the change of the sampling phase or of the sampling clock ti. In one branch, the received signal or undecided symbol S is correlated with the delayed decided symbol D, in the other branch the decided symbol D with the delayed received signal or the undecided symbol S is correlated.

For faster synchronization, a clock synchronization with Hilfssvmbolen is used according to DE 102 49 492 Al, in which the decider does not output symbols of the alphabet, but auxiliary symbols of the alphabet on nominal radii, but at the angle of the received signal to the correlator. Sollradien are such radii, on which occur symbols of the symbol alphabet.

DE 102 49 492 A1 describes a method and a circuit for generating an auxiliary symbol for adjusting a QAM demodulator (QAM: Quadrature Amplitude Modulation). The method is applied to receiving a digital signal coupled to a quadrature signal of two quadrature signal components. Corresponding predetermined positions of the received digital signal in the plane spanned by the quadrature signal pair, reference radii and range limits, in particular radius limits, are determined. By means of a scanning device which is dependent on symbol sampling times, preliminary symbols or corresponding signal values from the digital signal and their polar coordinates are scanned. From the polar coordinates, in particular the radius component, an associated setpoint radius is determined which, in conjunction with the angle component of the preliminary symbol, determines the polar coordinates of the auxiliary symbol in the plane of the quadrature signal pair. This auxiliary symbol replaces during the adjustment phase the decisive symbol in at least one decision-feedback control.

In higher-order QAM methods, some of these magnetic circuits may be so narrow that, on the one hand, their evaluation in conjunction with the usual interference is uncertain. On the other hand, since their regular contribution is low, this uncertainty hardly interferes, and the unsafe circular rings are further reduced in their effect by a suitable weighting of the rule information or completely masked out. Furthermore, it is known to admit circular rings which enclose the respective nominal radius more closely and thus capture it with greater certainty. If the measured radius lies outside of these narrow radius boundaries, then no auxiliary symbol is formed, because this would be too uncertain.

In a circuit arrangement for implementing such a method, a multiplexer with its signal inputs to a decision maker for preliminary auxiliary symbols and with an input to a symbol decision for decided symbols is have been connected and forming a switching device, which is controlled by a STEU ^¬ ereinrichtung then from the auxiliary symbol on a decided icon switches when the decided symbol is safely within the capture range of each decision feedback loop.

Particularly in the case of higher-order QAM methods with a large number of target radii on which the possible predefined positions for the digital signals lie, there is the problem that the assignment to a certain one of the target radii in the initial adjustment phase is not performed correctly, resulting in a significant extension of the settling time or possibly even to hanging in a loop resul ^¬ animals can.

Fig. 13 shows a representation of a first quadrant of 64-QAM. When receiving a signal as a draw symbol S, sketched by a point at the end of the radius ray, falls the decision not to the nearest symbol D °, z. To the left, but to the nearest nominal radius, which is outlined in bold. There, according to DE 102 49 492 A1, an auxiliary symbol is generated at the receiving angle α which, as a decided provisional symbol Dr, passes to the sampling rate recovery and other control loops of the receiver, in particular to the equalizer.

Such a method can also be implemented in an arrangement as shown in FIG. 2 with two decision makers 8, 12. Decisions in the upper decision maker 12 serving as auxiliary decision maker are made according to DE 103 44 756 A1 or EP 1 523 144 only or preferably according to radii. A switchover between the decided symbols D and Dr takes place in a switching device 13, X. Instead of a second decider 12 and the actual first decision maker 8 can be changed per se to make a decision only or preferably to radii.

For more significant symbol alphabets, e.g. according to 256-QAM, 1024-QAM, etc., the decision fields are very narrow, so that decided symbols and undecided signals do not differ much. Symbols, set radii R1, R2,..., R31 and Cartesian decision boundaries in the first quadrant of 256-QAM are shown by way of example in FIG. 14.

Thus, the control voltage for the sampling clock ti is useful only in the vicinity of the target time. This problem occurs in particular if the phase of the sampling clock ti is still far from its nominal value and if a small roll-off factor of the Nyquist filter is used, so that corresponding eye diagrams as decision criteria are very narrow.

With a carrier frequency offset, there is hardly any correct control information for a Cartesian decider. Here, the method according to DE 102 49 492 Al, which also in the case of a wrong phase angle of the local oscillator (English, carrier phase offset) or even a frequency offset (English, carrier frequency offset) still valid Regelinforitiationen supplies.

FIG. 15 shows a simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase in 256-QAM, in decisions with a rectangular discriminator, which is sketched by thin curves, and with auxiliary symbols according to DE 102 49 492 A1, which is represented by thick curve is outlined. In this case, a normalization of the two curves to the same slope in the central area. The RoIl off factor of the Nyquist filter is 12%, no carrier frequency offset is present, and an average of 40,000 symbol transitions has been used.

16 shows a simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase in 256-QAM, decisions with a rectangular separator, shown as a thin curve, and with auxiliary symbols according to DE 102 49 492 A1, shown as a thick curve, FIG the normalization of the last image was retained and is also used in the other similar comparative measurements. A RoIl-off factor of the Nyquist filter is 12%, a carrier frequency offset is 1.6% of the symbol rate and an average of 40000 symbol transitions has been taken into account.

Although a correlation with decided auxiliary signal according to DE 102 49 492 A1 also provides a control signal for the sampling phase outside the narrow limits which a Cartesian decision maker offers. However, this signal is very noisy. As a result, only a very slow control of the sampling phase can be achieved.

17 shows a simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase at 256-QAM, with decisions with auxiliary symbols according to DE 102 49 492 A1, a RoIl-off factor of the Nyquist filter of 12% and one Carrier frequency offset of 1.6% of the symbol rate for 10 trials with 1000 symbol transitions each. US 2005/0220220 A1 describes a further method and a corresponding circuit for generating an auxiliary symbol. According to the method, an intersection point on a single auxiliary radius is determined for a received signal in a first step in the radial direction, namely in such a way that rotation is effected on the auxiliary radius up to a position in whose radial course an actual predetermined position is located for a possible icon. Auxiliary symbols are thus positioned with an auxiliary radius in the intersection of a radial axis which passes through an actually possible predetermined position of digital signals and output as preliminary symbols to the other components of the corresponding control circuit. In this case, the auxiliary radius can be dimensioned such that it itself does not intersect any of the possible predefined positions for actual digital signals or decidable symbols. An essential aspect of such a procedure is that the angle of the received signal is decided so that it falls on the angle of a set point of the respective Qm constellation. Such decided temporary auxiliary symbols are used in the described circuit arrangement only for the regulation of the equalizer.

EP 1523144 shows a further circuit arrangement and method for deciding symbols in a manner known per se.

The object of the invention is to propose a method or a circuit arrangement which enable the generation of a control voltage with preferably a correlation circuit according to Mueller-Müller and even with large phase error of the symbol sampling time and also work with not yet engaged carrier frequency and phase control.

This object is achieved by a method for deciding a symbol having the features according to patent claim 1 or by a circuit arrangement, in particular for carrying out a Such a method with the features according to claim 11. Advantageous embodiments are the subject of dependent claims.

Accordingly, there is preferably provided a method of deciding a decided symbol upon receipt of a received symbol coupled to a pair of quadrature signals, in which the symbol is assigned to a particular setpoint of a plurality of setpoints in each plane spanned by the pair of quadrature signals, and each of the plurality of setpoints corresponds to a setpoint radius of a plurality of Sollradien is assigned, wherein on each of the set radii at least one of the setpoint points, and a receiving point of the received symbol on at least a predetermined fixed auxiliary radius a preliminary symbol is decided, the auxiliary radius independently of the

Sollradien is selectable, wherein the provisional symbol is decided on a position below the angle of the reception point of the received symbol on the auxiliary radius.

In particular, a method is also possible which can be used if none of the plurality of setpoint points lies on the auxiliary radius and if the angle of the reception point is unequal to an angle below which one of the decided symbols lies in the defined plane.

Optionally, a method is also possible which can be used in addition to these then formally excluded parameter ranges if at least one of the plurality of setpoints lies on the auxiliary radius and if another such angle of the reception point is equal to an angle below that of one of the decided symbols in FIG the level defined, as it is known in each case.

In this case, a method is advantageous in which more than one such auxiliary radius is defined in the defined plane and the provisional symbol depends on the position of the received one Symbols in the plane spanned on one of the auxiliary radii or not decided.

In particular, the provisional symbol is only decided if the received symbol lies in at least one region of the plane which is defined for the auxiliary radius or the respective one of the auxiliary radii.

In this case, the provisional symbol is preferably decided only if the received symbol lies in at least one predetermined radius region of the defined plane. The provisional symbol is preferably decided only if the received symbol lies in at least one predetermined radius region of the plane spanned with a radius greater than half the distance of the maximum of the symbol positions along one of the axes of the quadrature signal pair, particularly preferably with a radius greater than the distance of the maximum of the symbol positions lies along one of the axes of the quadrature signal pair.

An undecided received symbol may preferably be optionally unselected as the preliminary symbol instead of the preliminary symbol on the auxiliary radius.

The preliminary symbol is preferably forwarded to subsequent process steps or components for sampling rate recovery. Thus, the preliminary symbol is preferably used for the equalizer and the determination of the sample clock.

If the threshold for sufficient adjustment accuracy is undershot or exceeded, a switchover is made between the forwarding of the preliminary symbol and a decided symbol for a desired position to subsequent method steps or components.

Alternatively, for example, the preliminary symbol and a decided symbol for a target position for subsequent driving steps or components are mixed, in which case means the relative adding.

It is preferred to use a circuit arrangement for deciding a decided symbol when receiving a received symbol coupled to a pair of quadrature signals, in which the symbol is assigned to a specific desired point of a plurality of desired points in a plane spanned by the quadrature signal pair and each of the plurality of setpoint points is assigned to a desired setpoint. is associated with a plurality of target radii, wherein on each of the plurality of target radii at least one of the setpoint points, and a decision is designed to decide to a receiving point of the received symbol on at least a predetermined fixed auxiliary radius a preliminary symbol, wherein the auxiliary radius independently is selectable from the set radii, and wherein the decider decides the provisional symbol at a position below the angle of the reception point of the received symbol on the auxiliary radius.

This also makes it possible, in particular, for none of the multiplicity of setpoint points to lie on the auxiliary radius and for the angle of the reception point to be unequal to an angle below which one of the decided symbols lies in the defined plane.

Optionally, a circuit arrangement is also made possible which, in addition to these then formally excluded parameter ranges, takes into account if at least one of the plurality of setpoints lies on the auxiliary radius and if another such angle of the reception point is equal to an angle below that of one of the decided symbols in FIG spanned level, as it is known on its own.

Such a circuit arrangement with a switching device which is designed to be advantageous when the threshold falls below or exceeds a threshold for sufficient adjustment accuracy between the forwarding of the preliminary symbol and a decided symbol for a target position to subsequent ones Switch process steps or components. In this case, the switching device can be designed to receive symbols of a first decider and an auxiliary decider and to forward a selected one of these symbols. However, it is also possible a switching device, which controls a single decider suitable, so that this outputs decided either for an auxiliary radius preliminary symbols or symbols for target positions decided symbols.

However, provision may be made, e.g. also a circuit arrangement with instead of a switching device of a mixing device, which is designed to mix their results for a sufficient adjustment accuracy depending on the provisional symbol and a decided symbol for subsequent method steps or components.

Preferably, a clock controller is provided for providing a sampling clock to which the preliminary symbol is applied.

It is possible to use such a method or a circuit arrangement for carrying out such a method with preferably a single auxiliary radius. Advantageously, however, it is also possible to use a plurality of such auxiliary radii, to which certain regions, in particular radii regions within the plane spanned by the quadrature signal components, are assigned as decision collection regions. In particular, auxiliary symbols on a reduced set of radii for the clock synchronization of QAM signals are determined for adjusting a control loop. Surprisingly, the positioning of the provisional symbol as an auxiliary symbol at the angle of the received signal or symbol results in an unexpectedly rapid adjustment of the circuit arrangement, so that an immediate switching to decided symbols or a higher weighting of decided symbols can take place. Also particularly advantageous is the use of the provisional symbols, which were provisionally decided under the same angular constellation, such as the angular constellation of the received signals, not only as an input signal of an equalizer but also as an input signal for a clock control device for providing a sampling clock. This leads to a surprisingly fast adjustment of the sampling clock and correspondingly fast provision of received symbols or signal values for the decision maker or decision makers.

Thus, the method can be implemented in particular on a transmission of digital signals with QAM (Quadrature Amplitude Modulation), typical receiver concepts with complex downmixing into baseband, amplitude, carrier frequency / carrier phase and sampling rate control.

Especially with higher order modulation forms, e.g. QAM with 256 or 1024 symbols, the method is according to the method according to DE 102 49 492 Al modified so that auxiliary symbols are not generated on all desired radii, but only on some selected radii of reference as auxiliary radii, in extreme cases only on a single radius. There new symbols are generated at the angle of the input signal. The sample rate recovery circuit and the equalizer receive these preliminary symbols as "decided symbols". In extreme cases, the

Auxiliary decision-makers for the sampling rate control and the "decided symbols" equalizer are therefore available on just one radius.

Such a procedure can be implemented, for example. for complex digital modulation methods like QAM. They are used by the new radio, television and data services via cable and z. T. also used terrestrial.

An embodiment and modifications thereof will be explained in more detail with reference to the drawing. The same reference numerals in different of the figures refer to the same, like acting or equivalent components, sizes or functions, and will not be described repeatedly with reference to the explanations in descriptions of other of the figures. Show it:

Fig. 1 shows a plane spanned by quadrature signal components with symbols and Cartesian decision boundaries distributed therein in the first quadrant of 256-QAM to illustrate the basic idea of the preferred method;

Fig. 2 is an exemplary circuit arrangement for carrying out the preferred method;

3 shows a simulation of an output voltage of a Mueller-Müller correlator;

4 shows a further simulation of the output voltage with modified parameters;

5 shows a further simulation of the output voltage with differently modified parameters;

6 shows the plane spanned by the quadrature signal components according to a modified embodiment with an auxiliary radius and a decision area assigned to it;

FIG. 7 shows control voltages with decisions for only large radii of such a modified embodiment; FIG.

Fig. 8 is a simulation of the output voltage of the Mueller-Müller correlator of such an embodiment;

9 shows the first quadrant of the plane spanned by the quadrature signal components to illustrate a further modified embodiment with two auxiliary radii; 10 shows the first quadrant of the plane spanned by the quadrature signal components to illustrate a still further modified embodiment with two auxiliary radii;

Fig. 11 is a circuit arrangement known per se for receiving QAM signals;

FIG. 12 shows components of a known Müller-Müller correlator for detecting errors at sampling instants; FIG.

13 shows the first quadrant of the plane spanned by the quadrature signal components of 64-QAM for the purpose of illustrating a method according to the prior art according to DE 102 49 492;

FIG. 14 shows the first quadrant of the plane spanned by the quadrature signal components to illustrate symbols, set radii and Cartesian decision boundaries according to 256-QAM; FIG.

FIG. 15 shows an inherently known simulation of the output voltage of the known Mueller-Müller correlator as a function of the sampling phase in 256-QAM using the method according to DE 102 49 492; FIG. ^'

16 shows a simulation of the output voltage with modified parameters and

17 shows a further simulation of the output voltage to illustrate fluctuations in a large number of experiments carried out.

Fig. 1 shows the first quadrant spanned by quadrature signal components Q, I of 256-QAM along the coordinate axes with coordinate values x, y. Are shown possible target positions P = P (x, y) on which sent and ideally also received symbols are. In addition, Cartesian decision boundaries are shown in the usual manner, which form a decision area for the respectively lying in this target position P upon receipt of a symbol. An essential element is a radius which runs as an auxiliary radius Rh, in particular with equidistant spacing around the origin of the coordinate system. The auxiliary radius is chosen so that it is also possible that none of the desired positions P (x, y) is on the auxiliary radius Rh. The choice of the auxiliary radius is thus not dependent on the desired positions.

Also shown is an example received signal value of a received symbol S, which assumes an exemplary position Pe with receive coordinates xe, ye in the first quadrant. By way of example, it could be assumed that this received signal value or this received symbol S, in the case of a control loop that has not yet been adjusted, actually corresponds to a desired position P ° (3,2) in an adjacent decision area and a symbol actually assigned to this desired position P ° (3,2) D ° would be assigned.

However, during the control phase of a corresponding circuit arrangement or, if necessary, also at later times, the received symbol S is assigned to a position on the auxiliary radius Rh independently of the Cartesian decision boundaries. At this time, the position on the auxiliary radius Rh is set so that the provisionally decided symbol Dr is positioned at the same angle α on the auxiliary radius Rh which corresponds to the angle α of the received symbol S.

Thus, in Fig. 1, set points P (x, y) for symbols, Cartesian decision boundaries, and a central radius as auxiliary radius Rh in the first quadrant of 256-QAM are shown, on which an auxiliary symbol serves as preliminary symbol Dr for a received complex signal value undecided symbol S is formed. FIG. 2 shows an exemplary circuit arrangement in which an A / D converter 1 is shown as an initiating component. A clock signal t and a received receive signal sa are input to the A / D converter. The A / D converter 1 converts the analog reception signal SA into a digita ^¬ ized or digital signal sd. The digital signal sd is output to a quadrature converter or mixer 2, which converts the digitized or digital signal sd into the baseband and splits it onto the two quadrature components I, Q. The signal output from the mixer 2 is applied to a gain controller 3 in a manner known per se. A signal amplified by the gain controller 3 is supplied to a low-pass filter 4, the output of which is supplied to a sampling means 5 to which a sampling clock ti is applied for sampling. A signal sampled with the sampling device 5 is output to a further low-pass filter 6, in particular a Nyquist filter, and filtered by it and applied to an equalizer (German: Entzer- rer) 7. The equalizer 7 outputs received signal values or received symbols S to a decision maker 8, which decides a symbol and outputs a decided symbol D. The decided symbol D is output on further circuits for further data processing, as is known per se. To a circuit 9 for determining a phase difference Δφ, both the received signal value and the received symbol S and the decided symbol D are applied in order to determine the phase difference Δφ between them. The phase difference Δφ is applied to a frequency control device 10, FC, which is designed as a carrier control device. The frequency control device 10 controls, in a manner known per se, a local oscillator 11 which outputs two carriers offset by 90 ° to the mixer 2, so that the mixer 2 supplies the digital signal sd for providing the two quadrature components I , Q receives. In particular, these components can in a known manner also in a different order and Implementation be implemented. However, modifications in a manner known per se can also be implemented with regard to the further components.

According to a first particularly preferred embodiment, the circuit arrangement has an additional auxiliary decision block 12 to which the signal values or received symbols S output by the equalizer 7 are applied. The auxiliary decision block 12 carries out a method according to the procedure which is described with reference to FIG. 1 in order to decide on the received symbols S or the signal values of the equalizer 7 provisional symbols Dr as auxiliary symbols and to output these.

Both the signal values or symbols S output by the equalizer 7 and the decided symbols D output by the decision maker 8 are applied to a clock control device 14 in the controlled state of the circuit arrangement. The timing controller 14 determines, in a manner known per se, the sampling clock ti which is applied to the sampling means 5. In this way, a clock control loop is formed. In addition, the decided symbols D are applied as a further input signal to the E-qualifier 7 for processing in a manner known per se.

Both the provisional symbols Dr of the auxiliary decision maker 12 and the decided symbols D of the decision maker 8 are applied to a switching device 13, X which optionally either one of the applied preliminary symbol Dr or the applied decided symbol D to the clock control device 14 and to the equalizer 7 creates. The decision in the switching device 13 is made such that in non-steady or no longer steady clock control or equalizer 7, ie in particular when switching on the clock control circuit or the equalizer 7 of the circuit as long as the preliminary symbols Dr are forwarded until sufficient stability and the decided symbols D with sufficient accuracy lie on their desired radii. The switching device 13 then forwards the decided symbols D to the equalizer 7 and / or the clock control device 14.

Instead of a circuit arrangement with the decision maker 8 for

Deciding the decided symbols D, the further decider as Hilfsentscheider 12 and the switching device 13 for switching between these, for example, a mixing device can be used to mix their symbols Dr or D.

Instead of a circuit arrangement with the decision maker 8 for deciding the decided symbols D and the other decision maker as Hilfsentscheider 12, for example, a circuit arrangement with only a single decider can be used, which is driven accordingly, either provisional symbols Dr or decided symbols D to decide. Such a single decider would then be driven by a control device to corresponding decisions, wherein the control device uses comparable criteria as the switching device 13, in particular Einregelphasen the decision criteria for deciding the provisional symbols Dr and in controlled circuit the decision criteria for the decided symbols D in Ent - to activate separators or to supply such decision criteria to the decider.

Such auxiliary symbols as provisional symbols Dr, given as decided symbols, in the clock control device 14, as a per se known MM correlator, give a very good regulation voltage behavior over the entire range of the possible position of the sampling phase.

FIG. 3 shows a corresponding simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase or the sampling clock ti at 256-QAM, in the case of decisions with a rectangular discriminator, shown as a thin curve Auxiliary symbols according to DE 102 49 492 A1, shown as a dashed curve, and this for symbols on a radius according to the method of FIGS. 1 and 2, shown as a thick curve. Ver ^¬ applies was an arbitrary normalization of the last signal and a roil-off factor of the Nyquist filter of 12% with no carrier frequency and phase offset. Each graph shows an average of 40000 symbol transitions.

The control voltages according to the method according to DE 102 49 492 A1 and according to the present method are, in contrast to a control voltage, which is formed with decisions from the rectangular decision, independent of the position of the carrier frequency and phase, as with reference to FIG 4 is outlined. Shown in FIG. 4 is a simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase in 256-QAM, in decisions with a rectangular separator, shown as a thin curve, decisions with auxiliary symbol according to DE 102 49 492 A1 as dashed curve, and decisions for symbols on only one radius GE measure the method of FIG. 1 and 2, shown as a thick curve, with arbitrary normalization of the last signal. The RoIl off factor of the Nyquist filter is 12% and the carrier frequency offset is 1.6% of the symbol rate, with each curve showing an average of 40000 symbol transitions.

With this method, it is also possible to generate correct control voltages outside of a narrow range around the correct sampling instant which, in addition, generate much less noise than those generated with the auxiliary decoder according to DE 102 49 492 A1.

5 shows control voltages with symbols according to the preferred method as a simulation of the output voltage of the Mueller-Müller correlator as a function of the sampling phase at 256-QAM, a RoIl-off factor of the Nyquist filter of 12% and a carrier frequency offset 1 , 6% of the symbol rate for 10 trials with 1000 symbol transitions each. In addition, several equidistant or otherwise selected radii with their own decision zones can be used.

Particularly advantageous is a method in which only signals with high absolute values are decided. All other signals are passed on undecided.

Fig. 6 correspondingly shows the first quadrant of 256-QAM according to such a modified embodiment. While in FIG. 1 all received signal values or symbols S of a position corresponding to their received angle α are assigned to the auxiliary radius Rh as preliminarily decided symbols, according to the embodiment shown with reference to FIG. 6, only those received signal values or symbols S are decided as provisionally Symbols Dr or auxiliary symbols associated with the corresponding position on the auxiliary radius Rh, which are in a decision area with a radius value greater than or equal to the auxiliary radius Rh. Other signal values or received symbols remain unchanged in a decision according to this embodiment.

FIG. 7 shows corresponding control voltages with decisions for only large radii, that is to say for received signal values or symbols S which lie in the outer region of the first quadrant. For example, received signal values or symbols outside an auxiliary radius Rh are determined has a radial value equal to the outer position value of the Cartesian decision boundaries adjacent to the axes of the quadrature components Q, I or a corresponding desired position P (x, y). Shown here is a simulation of the output voltage of a Mueller-Müller correlator as a function of the sampling phase at 256-QAM with a RoIl-off factor of the Nyquist filter of 12% and a carrier frequency offset of 1.6% of the symbol rate at 10 Try with 1,000 symbol transitions each. FIG. 8 shows a simulation of the corresponding output voltage of the Mueller-Müller correlator as a function of the sampling phase at 256-QAM in the case of decisions with a rectangular decider, which are shown as a dotted curve and which in this case are not worth mentioning because of the carrier frequency control deviation Controlled variable provides, of decisions with an auxiliary symbol according to DE 102 49 492, which are shown as a dashed curve, and decisions for only large radii, which are shown as a solid curve according to a method of FIG. 6, with respect to the latter signal is an arbitrary Scaling was applied. In the experiments, a 12% RoIl-off factor of the Nyquist filter and a carrier frequency offset of 1.6% of the symbol rate were used, the curves showing an average of 40,000 symbol transitions each.

Alternatively, other decision areas are possible, in particular with more than one auxiliary radius set in this way. By way of example, FIG. 9 shows a first quadrant of 256-QAM, in which two auxiliary radii Rh, Rh * are arranged, whereby not necessarily set positions for decidable symbols must lie on them. In the embodiment illustrated by way of example with reference to FIG. 9, received signal values or received symbols S having a position in an outer radius range lying between the auxiliary radii Rh, Rh * are assigned to the outer of the two auxiliary radii Rh, while received signal values or symbols S * with a position within the radii range are assigned to the smaller of the two auxiliary radii Rh *. Again corresponding auxiliary symbols or temporary symbols Dr are decided at the corresponding angles α, α * of the received signal values or symbols on the auxiliary radii Rh and Rh *, respectively. By way of example, all received signal values or symbols S, S * are assigned to one of the two auxiliary radii Rh or Rh * for determining a corresponding setpoint position and a corresponding provisional symbol Dr or Dr *. The decision boundaries need not be in the middle between the radii, but otherwise can be chosen. The radii on which the auxiliary symbols lie do not need to lie in the middle or at the border to decision areas, but may be elsewhere within the complex plane.

Also, a method in which only signals whose values are within certain areas, ie at two radii z. B. within the smaller or outside of the larger radius, works. These signals are mapped below the received signal angle to the small or larger radius. In the area between the two radii, the received signal is not changed in the decider and simply output as a "decided symbol". Additionally possible are any circular rings as auxiliary radii for undecided areas.

Accordingly, FIG. 10 shows by way of example the first quadrant of 256-QAM with likewise two auxiliary radii Rh, Rh *, in which case, for example, all received signal values or received symbols remain in a radius range between these two auxiliary radii Rh or Rh *. In particular, signal values received in the radius range between the two auxiliary radii Rh or Rh * can be forwarded as undecided symbols to the sampling recovery or clock control device 14 and the equalizer 7.

Although the preferred circuit of FIG. 2 for large phase deviations from the sampling instant gives a much better control signal than a circuit with auxiliary symbol according to DE 102 494 92 A1 or even as one with square decision, yet for sampling times at or close to the ideal phase, the performance of other decision makers much better. This applies in particular to the rectangular separator with carrier frequency control and carrier phase regulation engaged. Therefore, the receiver by means of the switching device 13 in the locked state should preferably either a sampling rate control with a decision for an auxiliary symbol according to DE 102 494 92 Al, a Modified decision for an auxiliary symbol according to DE 102 494 92 A1 within the symbol boundaries of a symbol decision according to EP 1523144 or switched to a decision with a rectangular decision.

In general, it is also possible to process in two decision makers, a rectangular decision maker or decider according to DE 102 494 92 A1 and a decision maker of FIGS. 1 and 2, or more decision makers in different ways decided symbols in two or more clock control devices 14 and the Add control signals with a corresponding weight in a subsequent block. This means that, especially for the acquisition phase, there is a control voltage that always has the correct sign and a large slope in the range of ideal sampling. The weighting between the signals can be adjusted to the progress of the synchronization.

In order to minimize the required effort, instead of different correlators for the different methods, it is possible to use the different types of "decided symbols", in particular auxiliary symbols on a single radius according to FIGS. 1 and 2 in particular, and symbols with a rectangular separator or such as EP1523144 or DE 102 494 92 A1 before being used as a "decided symbol" in the Mueller-Müller correlator or in the equalizer. The switching device 13 then becomes a weighted adder from a switcher.

Claims

claims

A method of deciding a symbol (D) upon receipt of a received symbol (S) coupled to a quadrature signal pair (I, Q), wherein

in a plane (x, y) spanned by the pair of quadrature signals (I, Q), the symbol (D) is associated with a specific set point of a plurality of set points (P (x, y)) and each of the plurality of set points (P (x, y) is assigned to a setpoint radius of a plurality of set radii (Rl - R31), wherein at least one of the setpoint points (P (x, y)) lies on each of the plurality of set radii (Rl - R31), and

- a provisional symbol (Dr) is decided to a receiving point (xe, ye, α) of the received symbol (S) on at least one predetermined fixed auxiliary radius (Rh), wherein

- The auxiliary radius (Rh) is independent of the desired radii (Rl - R31) selectable, thereby e e n e c e e s that e

- the preliminary symbol (Dr) is decided at a position below the angle (α) of the reception point (Pe) of the received symbol (S) on the auxiliary radius (Rh).

2. The method of claim 1, wherein on the auxiliary radius (Rh) none of the plurality of setpoints (P (x, y)) is located.

3. The method of claim 1 or 2, wherein the angle (α) of the receiving point (Pe) is not equal to an angle below which one of the decided symbols (D) in the plane spanned (x, y).

A method according to any preceding claim, wherein more than one such auxiliary radius (Rh) is set in the spanned plane (x, y) and the preliminary symbol (Dr) is dependent on the position (Pe) of the received symbol (S) in the plane spanned (x, y) on one of the auxiliary radii (Rh or Rh *) or not decided.

A method as claimed in any preceding claim, wherein the provisional symbol (Dr) is decided only if the received symbol (S) is in at least one of the auxiliary radius (Rh) and the respective one of the auxiliary radii (RH, Rh *) the spanned level (x, y) is located.

6. The method of claim 5, wherein the provisional symbol (Dr) is decided only if the received symbol (S) is located in at least one predetermined radius range of the plane spanned (x, y).

7. A method according to claim 5 or 6, wherein an undecided received symbol (S) is undecided as the preliminary symbol instead of the provisional symbol on the auxiliary radius (Rh; Rh, Rh *).

A method as claimed in any preceding claim, wherein the preliminary symbol (Dr) is passed to subsequent process steps or components for sample rate recovery.

9. Method according to any preceding claim, in which, when the threshold for sufficient adjustment accuracy is exceeded or exceeded between the forwarding of the preliminary symbol (Dr) and a decided symbol (D) for a position of position (P (x, y)) is switched to subsequent process steps or components.

10. The method according to any one of claims 1 to 8, wherein the provisional symbol (Dr) and a decided symbol (D) for a setpoint position (P (x, y)) are added weighted for subsequent steps or components.

11. A circuit arrangement for deciding a decided symbol (D) upon receipt of a received symbol (S) coupled to a quadrature signal pair (I, Q), in which

- In a plane spanned by the quadrature signal pair (I, Q) plane (x, y), the symbol (D) a specific setpoint of a plurality is associated with setpoint points (P (x, y)) and each of the plurality of setpoints (P (x, y)) is associated with a setpoint radius of a plurality of set radii (Rl-R31), wherein on each of the plurality of set radii (Rl-R) R31) is at least one of the reference points (P (x, y)), and

a decision maker is arranged to decide on a reception point (xe, ye, α) of the received symbol (S) on at least one predefined auxiliary radius (Rh) a provisional symbol (Dr), wherein - the auxiliary radius (Rh) is independent of the Sollradien (Rl - R31) is selectable, characterized in that

- the decider (12) decides the provisional symbol (Dr) at a position below the angle (α) of the reception point (Pe) of the received symbol (S) on the auxiliary radius (Rh).

12. Circuit arrangement according to claim 11 with a switching device (13) which is designed, when falling below or exceeding a threshold for a sufficient control accuracy between seeing the forwarding of the preliminary symbol (Dr) and a decided symbol (D) for a desired position ( P (x, y)) to subsequent process steps or components to switch.

13. Circuit arrangement according to claim 11 or 12, having a mixing device which is designed to add weighted results for subsequent method steps or components for sufficient control accuracy as a function of the preliminary symbol (Dr) and a decided symbol (D).

14. Circuit arrangement according to one of claims 11 to 13, which is adapted to a clock control means (14) for providing a sampling clock (ti) to apply the provisional symbol (Dr).

15. Use of a circuit arrangement according to one of claims 11 to 14 for carrying out a method according to one of claims 1 to 9.