WO2004012409A1

WO2004012409A1 - Detection and compensation of constellation impairments

Info

Publication number: WO2004012409A1
Application number: PCT/GB2003/003282
Authority: WO
Inventors: James Rowe; Peter Garrett
Original assignee: Pace Micro Technology PLC
Current assignee: Arris Global Ltd
Priority date: 2002-07-25
Filing date: 2003-07-25
Publication date: 2004-02-05
Anticipated expiration: 2005-01-25
Also published as: GB0217219D0; AU2003255733A1

Abstract

The invention relates to a method for improving the quality of a signal transmitted from a broadcaster to at least one broadcast data receiver (BDR) at a remote location from the broadcaster. The signal includes digital data for reception and processing by the BDR. The invention relates to the steps of generating constellation error vectors for the symbols of the transmitted signal and then ascertaining gain and phase impairments present in the received signal. Unique averaged error vector signatures are obtained for said impairments and identify those impairments which have fixed or slowly varying gain and/or phase values.

Description

DETECTION AND COMPENSATION OF CONSTELLATION IMPAIRMENT

The invention to which this application relates is to the improvement and use of Quadrature Amplitude Modulation (QAM) demodulators. QAM modulators in practise, vary the phase and amplitude of a radio frequency (RF) carrier to encode several bits of information pet symbol, a symbol being a particular phase and amplitude state, for digital data which is transmitted to a location for processing from a remote broadcaster.

The data which is transmitted from the broadcaster is transmitted in a digital format; however, during the transmission, the data is carried in an analogue format such that, at some point after reception, analogue to digital conversion is required to take place. Typically, this occurs within a broadcast data receiver (BDR) provided at a user location. The BDR provides for the reception, processing and generation of video and/or audio data for display by the user.

Typically, there are a number of locations at which the analogue digital conversion can occur but, the preferred location is that where parallel streams of inphase (I) and quadrature (Q) samples occur.

During the transfer of the data between the broadcaster and the reception location, the communications channel used can add noise, usually termed Additive White Gaussian Noise — AWGN; interference which is usually modelled as discrete CW signals, and multipath (also known as echoes) . Even when received at the BDR, more noise can be added such as further interference which is internally generated spurii, IQ gain and/or phase imbalances and DC offsets. All of these impairments can act to reduce the quality of the received signal and, in locations where the received signal is already relatively weak, this further impairment due to the transmission system, may render certain services or functions unworkable and therefore un-viewable or un-listenable by the user.

One method of demodulation is for conventional broadcast data receivers to mix an RF carrier down to base band via one or more intermediate frequencies (IF) - A zero intermediate frequency (ZIF) receiver architecture mixes the carrier direct to the base band in-phase (I) and quadrature (Q) components using quadrature local oscillators (LO) at or near the carrier frequency. This approach confers many cost advantages but it is difficult to keep a good balance between the I and Q phases and gains as these depend on the matching of analogue electrical components of the signal path. As a result, there is often a residual DC component present due to break-through of the LO into the RF front end and these impairments are serious enough to render the ZIF architecture unsuitable for higher order QAM modulation schemes.

As these problems are already known, attempts have been made to try to overcome the problems as described.

The aim of the present invention is to provide a method whereby the detrimental effects on the transmitted signal can be reduced thereby improving the quality of the signal received and processed by the BDR and hence the quality of the services and/ or functions provided to the end user by the BDR.

In a first aspect of the invention, there is provided a method for improving the quality of a data signal transmitted from a broadcaster to a broadcast data receiver (BDR) at a remote location, said signal including digital data for reception and processing by the BDR to generate video, audio and/or auxiliary data and characterised in that the method includes the steps of generating constellation error vectors for the symbols of the transmitted signal, ascertaining any gain and phase impairment present in the received signal and obtaining unique averaged error vector signatures for said impairments, identifying those impairments which have fixed or slowly varying gain and phase values and using the signatures for those identified impairments in an adaptive method to reduce or remove the impairments from said signal.

In one embodiment, for each identified impairment, if there are imbalances between the gain and phase, then there is generated a separate metric for a shift in a positive sense and a negative sense. Typically these metric values are combined such that the metric value obtained is the difference between the scalar derived from positive shift and negative shift which is then allocated a sign.

Typically the adaptive structure can be generated to suit specific requirements and may be either iterative or single step.

Typically, in use, the method allows the removal of certain of gain and phase error impairments and the limit on the amount of gain and phase error that can be removed is typically defined by the quantisation noise of the conversion process at the broadcast data receiver. The quantisation noise of the digital analogue conversion therefore effectively sets the resolution of the grid on which the received constellation points sit.

In a second aspect of the invention there is provided a method for improving the quality of a signal transmitted from a broadcaster to a broadcast data receiver (BDR) at a remote location, said signal including digital data for reception and processing by the BDR and characterised in that an adaptive structure is adopted to allow the removal of selective gain and phase error impairments and subsequent improved digital form of the signal.

Specific embodiments of the invention is now described with reference to the accompanying drawings, wherein:-

Figure 1 illustrates a typical QAM constellation;

Figure 2 illustrates a block diagram of a broadcast data receiver and demodulator apparatus;

Figure 3 illustrates a conventional ZIF receiver architecture;

Figure 4 illustrates one form of noise AWGN which can be present on transmitted signal;

Figure 5 illustrates a phase noise which can be present on a transmitted signal;

Figure 6 illustrates CW interference which can be present on a transmitted signal;

Figure 7 illustrates an IQ gain imbalance which can affect a transmitted signal;

Figure 8 illustrates an IQ phase imbalance which can be present on a transmitted signal;

Figure 9 illustrates a DC offset which can have an effect on the transmitted signal; Figure 10 illustrates the generation of error vectors for each impairment identified in a transmitted signal in accordance with the invention;

Figure 11 illustrates an adaptive structure in accordance with one embodiment of the invention; and

Figure 12 illustrates a 16 QAM example of the invention.

Turning firstly to Figure 1 , there is illustrated a typical example of a QAM constellation which comprises a QPSK which encodes two bits per symbol, giving a total of 2²= 4 possible symbols 2 that can be transmitted as part of the signal.

Figure 2 illustrates a typical broadcast data receiver architecture wherein the transmitted signal 4 is received by the BDR represented by the broken line 6 so that data which is transmitted and received enters the BDR for processing. Data first passes through gain 8, transferring to down conversion to base band 10 and analogue digital conversion 12. The converted data then passes through demodulator 14 as shown. Thus, in the broadcast data receiver there are parallel streams of in-phase and quadrature samples.

Figure 3 illustrates a zero intermediate frequency receiver architecture wherein the signal 4 enters amplifier 16 passing to mixers 18 and 20. From both mixers 18 and 20 the data passes to filters 22 and 24 respectively with the signal emitted from the filter 22 being the in-phase sample 26 and the signal leaving filter 24 being the quadrature sample 28 which in turn passes to the demodulator (not shown) . In this conventional system, the carrier is mixed directly to the base band IQ using quadrature local oscillator 30 at or near the carrier frequency but it is difficult to keep a good balance between the IQ phases and gains 26, 28 as these depend on the matching of the analogue components and there is often a residual DC component present due to breakthrough of the LO 30 into the RF front end. This can cause serious impairments which means that the same is unsuitable for higher order QAM modulation schemes like 16 QAM or higher.

To illustrate the problems which can occur with various impairments on a received signal, one can refer to Figures 4 to 9, each of which illustrates a particular type of impairment. However, it should be clear to the reader that it is unusual for only one particular type of impairment to have a bearing on the signal and therefore the illustrations of Figures 4 to 9 are provided for clarity purposes to illustrate the effect which each particular impairment would have if on its own, and these are illustrated with respect to the signal depicted in Figure 1. Thus, each of the figures illustrates a particular "signature" for each type of impairment, but frequently the impairment which occurs is due to a combination of these types of impairment.

Thus if one refers to Figure 4 which illustrates AWGN impairment, it will be seen that each symbol 2 has considerable interference around the same, thereby rendering the same unclear.

Figure 5, which illustrates phase noise impairment, illustrates how again there is significant interference for each symbol 2 which is spread in a substantial arc 30 for each symbol as illustrated.

Figure 6 illustrates how each symbol 2 is enlarged to form an annular arrangement 32 as illustrated. Figure 7 which shows IQ gain impairment, illustrates how the symbol locations 34 change with respect to the original symbol locations 2 indicated in broken lines .

Figure 8 illustrates IQ phase imbalance and again illustrates how the symbols 36 vary in location with respect to the original location as indicated in broken lines by the reference numeral 2.

Figure 9 illustrates DC offset impairment and illustrates how the symbol positions 38 vary with respect to the respective desired location 2 illustrated in broken lines.

Of these impairments, those shown in Figures 7, 8 and 9 are relatively slowly varying and are almost fixed such they are effectively constant over a number of symbol periods when compared to the characteristics of the impairment of Figures 4, 5 and 6. Because the impairments of Figures 7, 8, 9 are relatively slow moving they can almost be dealt with as being fixed with respect to on-going constellations for the purposes of ongoing signal reception, and because there are unique signatures for each type of impairment, i.e. each particular impairment is easily distinguished, it is possible to identify the type and magnitude of the gain and phase errors in the received constellation with respect to certain impairments and hence remove them and this is achieved by the current invention.

If one refers to figures 4 to 6, if the error vector between each received symbol and its closest ideal constellation is vector averaged or otherwise vector filtered over enough symbols, it will converge on the ideal location. However, with respect to the impairment shown in Figures 7 to 9, due to their slow changing nature the error vector for these impairments will give an average error vector that gives the vector of the impairment itself so that if there are any gain and phase impairments present in the received signal due to the impairment of Figures 7 to 9 then they will be present in the averaged error vectors. Thus, each impairment has a unique averaged error vector signature as illustrated in Figures 10a to e, with, for example, Figure 10a illustrating the error vectors for the impairment of Figure 7, Figure l Od illustrating the Figure 8 impairment vectors and Figure l Oe illustrating the Figure 9 vectors.

Each of the figures 10a to e illustrate the particular error vector signature for particular impairments. These signatures are orthogonal in that a scalar metric can be formed that only contains information about that impairment and this metric is formed by summing the magnitudes of the vector products of the signature vectors and the vector averaged error vectors for each constellation point which is equivalent to calculating the metric for each received constellation point and then averaging.

In the case of the gain and phase imbalances, then for each there is a separate metric for a shift one way and a shift the other way and, to be practically useful these are combined such that the metric is the difference between the scalar derived for positive shift and negative shift, in other words, assigned value.

Examples of the equations for each of the Figures l Oa-c are hereby given;

Figure 10a - GQ_t = GQ₄ = (°_α); GQ₂ = GQ₃ = {° );

Figure 10b - GI₁ = GI₂= (¹ ₀) ; GI₃=GI₄= C )

Figure 10c and l Od - P^ O; P₂= ( ); P₃= ( ); P₄= ( )

Figure l Oe - D₁=D₂=^:D₃=D₄= (^α ₀) and an example of the metric value obtained is C₁.P₁ + C₂.P₂+C₃.P₃ ⁺C₄.P₄ = metric

With this impairment information then the same can be used in an adaptive structure as shown in Figure 11 which illustrates an adaptive structure incorporating a control algorithm which can be either iterative or single step . The provision of the metrics give the sign of the impairment and its magnitude which can be input into the adaptor structure shown.

The invention can be used with respect to higher order QAM signals as is illustrated with regard to Figure 12 which illustrates a 16 QAM example.

In the QPSK example of Figure 1 , there is a large space between constellation point which means that it only needs a low number of bits for each symbol sample so as to not affect receiver performance. This means that only relatively large gain and phase problems can be resolved and hence removed but QPSK is not too susceptible to these problems in any case. The higher order modulation schemes like 16 QAM and above are more susceptible to gain and phase imbalances but have to be sampled using more bits which means greater resolution and finer control over gain and phase.

Thus in accordance with the present invention there are provided two types of signal impairment, those whose error vectors tend to zero after vector averaging or filtering over many symbols and those that don't. The latter group, such as those in Figures 7 to 9, include fixed or slowly varying gain and phase impairments and the resulting error vectors for these particular impairments for the signatures for each of the impairments which can then be used in a suitable adaptive structure to reduce or remove the same. As the gain and phase impairments of a received digitally modulated signal leave unique signatures of distortion on a demodulated constellation then each signature which is unique to the particular impairment can be separately resolved from the constellation samples and averaged over a suitable number of samples to give a scalar metric of the magnitude of each impairment. This metric can then be used to subtract the effects of the corresponding impairment from the incoming demodulated constellation. As a result, the ZIF architecture can now be used with high order QAM modulation schemes as the resulting impairments to reception such as gain and phase errors, can be removed in the digital domain. The removal of these errors in the digital domain is far less costly than designing them out in the analogue domain.

Claims

1. A method for improving the quality of a data signal transmitted from a broadcaster to a broadcast data receiver (BDR) at a remote location, said signal including digital data for reception and processing by the BDR to generate video, audio and/ or auxiliary data and characterised in that the method includes the steps of generating constellation error vectors for the symbols of the transmitted signal, ascertaining any gain and phase impairment present in the received signal and obtaining unique averaged error vector signatures for said impairments, identifying those impairments which have fixed or slowly varying gain and phase values and using the signatures for those identified impairments in an adaptive method to reduce or remove the impairments from said signal.

2. A method according to claim 1 characterised in that for each identified impairment if there are imbalances between the gain and phase values, a separate metric is generated for the shift in a positive sense and a negative sense.

3. A method according to claim 2 characterised in that the metric values are combined such that the metric value obtained is the difference between the scalar derived from the metrics for the positive shift and negative shift which is then allocated a sign.

4. A method according to claim 1 characterised in that the adaptive method used is generated to suit specific requirements and may be either iterative or single step.

5. A method according to claim 1 characterised in that the removal of certain of gain and phase error impairments and the limit on the amount of gain and phase error that can be removed is defined by the quantisation noise of the conversion process at the broadcast data receiver.

6. A method according to claim 5 characterised in that the quantisation noise of the digital to analogue conversion sets the resolution of a grid on which the received constellation points sit.

7 A method for improving the quality of a signal transmitted from a broadcaster to a broadcast data receiver (BDR) at a remote location, said signal including digital data for reception and processing by the BDR and characterised in that an adaptive method is adopted to allow the removal of selective gain and phase error impairments and subsequently improve the digital form of the signal.

8. A method according to claim 7 characterised in that the selected impairments are any or any combination of IQ gain impairment, IQ phase impairment, DC offset impairment.

9 A method according to claim 7 characterised in that an error sector vector signature is provided for each selected impairment.

10 A method according to claim 7 characterised in that when used with 16 QAM a set of four "representative" vectors, one from each quadrant are provided for each selected impairment.