WO2010026560A2

WO2010026560A2 - Predistortion unit and method of predistorting signals

Info

Publication number: WO2010026560A2
Application number: PCT/IB2009/053922
Authority: WO
Inventors: Peter Bode; Ralf Burdenski; Roland Egon Ryter
Original assignee: Nxp B.V.
Priority date: 2008-09-08
Filing date: 2009-09-08
Publication date: 2010-03-11
Also published as: WO2010026560A3

Abstract

A predistortion unit for a polar transmitter is provided, wherein the predistortion unit comprises a first amplitude modulation look up table memory storing a first set of data representing an amplitude predistortion function, and a first phase modulation look up table memory storing a second set of data representing a phase predistortion function, wherein the predistortion unit is adapted to predistort nominal amplitude signal and nominal phase signal of a signal by applying the first set of data to the nominal amplitude signal and the second set of data to the nominal phase signal.

Description

PREDISTORTION UNIT AND METHOD OF PREDISTORTING SIGNALS

FIELD OF THE INVENTION

The invention relates to a predistortion unit for predistorting signals, in particular of amplitude modulation signals and phase modulation signals in polar transmitters.

The invention further relates to a method of predistorting signals, in particular of amplitude modulation signals and phase modulation signals in polar transmitters.

Moreover, the invention relates to a program element.

Further, the invention relates to a computer-readable medium.

BACKGROUND OF THE INVENTION

There is growing interest in polar transmitter concepts in mobile phones because they promise advantages over conventional linear up conversion transmitters, such as easier integration in CMOS, higher power efficiency and less external filtering.

Polar solutions for GSM/EDGE are already on the market and it appears to be just a question of time until solutions for UMTS/WCDMA appear and it appears likely that polar transmitter concepts will dominate the scene.

A particular problem in polar transmitter concepts are impairments of the power amplifier (PA) which can be compensated by applying predistortion functions in the digital baseband domain.

Thus, there may be a need to provide a device and a method of applying a predistortion function to amplitude modulation signals and phase modulation signals, in particular for digital signals. SUMMARY OF THE INVENTION

It is an object of the invention to provide a predistortion unit for a polar transmitter and a predistortion method which may enable an improved flexibility and speed of the predistortion process of signals in a polar transmitter, in particular for digital data.

In order to achieve the object defined above, a predistortion unit, a method of predistorting signals, a program element, and a computer-readable medium according to the independent claims are provided. Preferred embodiments are described in the dependent claims.

According to an exemplary aspect of the invention a predistortion unit for a polar transmitter is provided, wherein the predistortion unit comprises a first amplitude modulation look up table memory storing a first set of data representing an amplitude predistortion function, and a first phase modulation look up table memory storing a second set of data representing a phase predistortion function, wherein the predistortion unit is adapted to predistort nominal amplitude signal and nominal phase signal of a signal by applying the first set of data to the nominal amplitude signal and the second set of data to the nominal phase signal.

According to an exemplary aspect of the invention a transmitter is provided which comprises a predistortion unit according to an exemplary aspect of the invention and a local memory, wherein the local memory is adapted to store a representation of the amplitude predistortion function and/or a representation of the phase predistortion function.

In particular, the representations may be formed by predistortion functions which may be given as coefficients of the predistortion functions or as full set of data representing the predistortion units, for example. That is, coefficients of the amplitude predistortion function and/or coefficients of the phase predistortion function may be received and stored or the predistortion functions may be received and stored in a non compressed manner, e.g. as full sets of data which may also be called raw look up table (LUT) content. In case of a non compressed or uncompressed storing of full sets of data, the local memory, e.g. local RAM, may be initialized via a digital interface between a baseband chip and the transmitter during start up. This may be a suitable way since at the start up no stringent real time requirements exist. According to an exemplary aspect of the invention a method of predistorting a nominal amplitude signal and a nominal phase signal is provided, wherein the method comprises applying a nominal amplitude signal and a nominal phase signal to a predistortion unit, and applying a first set of data to the nominal amplitude signal and a second set of data to the nominal phase signal, wherein the first set of data represents an amplitude predistortion function, and wherein the second set of data represents a phase predistortion function.

According to an exemplary aspect a program element is provided, which, when being executed by a processor, is adapted to control or carry out a method according to an exemplary aspect. Such a program element may be for example a part of a firmware of a transmitter.

According to an exemplary aspect a computer-readable medium is provided, in which a computer program is stored which, when being executed by a processor, is adapted to control or carry out a method according to an exemplary aspect. Such a computer-readable medium may be for example a memory for the firmware of a transmitter.

The term "predistortion function" may particularly denote a function which is used to calculate values used to determine an amount of predistortion which is imposed to a signal in order to compensate for a distortion which is imposed to the signal afterwards, e.g. during a transmission of the signal or during a processing of the signal, e.g. by a power amplifier. Such a predistortion function may be given in different forms, e.g. in the form of coefficients characterizing the function or in the form of a set of data or look up table content.

The term "set of data" may particularly denote pairs of data which represents the input value of a predistortion function and the corresponding output value of the predistortion function. For example, the first set of data and/or the second set of data may represent a full set of data, i.e. for each nominal amplitude signal value a corresponding output value exists and for each nominal phase signal value a corresponding output value exists, e.g. in the form of a result so that no calculation may be necessary contrary to solutions which stores coefficients of a function. These full sets of data may be stored in the form of look up tables (LUTs) in the respective look up table memories. Thus, a look up table may be a specific form in which sets of data can be stored. Thus, a look up table and the corresponding look up table content (sets of data) have to be distinguished from a function or coefficients which may be used for calculating the respective output value for a given input value. That is, a look up table according to the present invention may comprise a full set of data wherein the address word length of the look up table equals the word length of the nominal amplitude signal, which may be a digitized signal. Since the look up table content represents a full set of data it may not be necessary to perform an interpolation in real time when predistort the respective signals by using a small number of coefficients of interpolating functions.

By providing a method of predistorting according to an exemplary aspect of the invention it may be possible to provide an efficient method with respect to speed, necessary area in the predistortion unit, flexibility and the like, since in real time, i.e. during the predistortion process, only address to data mapping occurs in the predistortion unit. That is, no real time calculation may be necessary during the predistortion process. The calculation of the data sets stored in the look up table memories may be performed offline. Since no on line or real time calculation is necessary it may also be possible to avoid the usage of dedicated hardware performing the predistortion function. In particular, the method may be performed in a pure software solution or a hybrid solution. Due to the fact that the method according to an exemplary aspect of the invention may be easily implemented in software it may be possible to provide a flexible predistortion unit which may be used to support different types of power amplifiers, e.g. from different vendors although the distortion functions of such different power amplifiers may be significantly different. Due to the flexibility it may also be possible to reuse the predistortion unit in multimode transmitters that can be reconfigured to different standards. Furthermore, high sampling rates as they may be encountered in high bandwidth standards such as UMTS/WCDMA may be effectively handled since only an address to data mapping may occur in real time.

It should be noted that the predistortion unit may be adapted to predistort digital data or digital signals. Furthermore, the look up table memories may be RAMs acting as reconfigurable look up tables. The size and costs may be acceptable when using CMOS technology. Power amplifier impairments may depend on a number of parameters, such as temperatures, frequency and average RF output power so that the data sets stored in the look up table memories might represent a multidimensional space of predistortion functions, i.e. a multidimensional space of output values of predistortion functions. Next, further exemplary embodiments of the invention will be described.

In the following, further exemplary embodiments of the predistortion unit will be explained. However, these embodiments also apply for the transmitter, for the method of predistorting, for the program element and for the computer-readable medium.

According to another exemplary embodiment the predistortion unit further comprises a look up table computer adapted to compute the first set of data and/or the second set of data.

Thus, it may be possible to separate the computation of the first data set and/or the second data set and the applying of the same to nominal amplitude signal and nominal phase signal in time. Therefore, the procedure of predistortion may be less time critical and it may be possible to provide less computational resources for the predistortion. The look up table computer may also be called look up table processor.

According to another exemplary embodiment of the predistortion unit the look up table computer is adapted to perform the computation of the first set of data and/or the second set of data off-line.

That is, the look up table computer may perform the computation of the first and/or second data set in advance and sends the computed first and/or second data set to the respective look up table memory.

According to another exemplary embodiment of the predistortion unit the look up table computer is adapted to receive a signal representing the amplitude predistortion function and/or the phase predistortion function. In particular, the look up table computer may be adapted to receive the amplitude predistortion function and or the phase predistortion function in an uncompressed or compressed manner, e.g. by receiving coefficients characterizing the predistortion functions.

According to another exemplary embodiment of the predistortion unit the computation of the first set of data and/or the second set of data comprises a decompression of a compressed signal.

In particular, the compressed signal may comprise coefficients characterizing the amplitude predistortion function and/or coefficients characterizing the phase predistortion function. By using a compressed signal to transmit the predistortion functions to the look up table computer it may be possible to reduce the amount of bandwidth resources which are necessary so that it may be suitable to use a link between a memory external to the predistortion unit and the predistortion unit having less bandwidth.

According to another exemplary embodiment the predistortion unit further comprises a second amplitude modulation look up table memory storing a third set of data representing an amplitude predistortion function, and a second phase modulation look up table memory storing a fourth set of data representing a phase predistortion function.

According to another exemplary embodiment of the predistortion unit the predistortion unit is adapted to write the third set of data to the second amplitude modulation look up table while the first set of data is used to predistort the nominal amplitude signal, and/or wherein the predistortion unit is adapted to write the fourth set of data to the second phase modulation look up table while the second set of data is used to predistort the nominal phase signal.

Thus, at one point in time one amplitude modulation look up table or the look up table content, namely the set of data, may be used to predistort the nominal amplitude signal, i.e. may be read from, while the other amplitude modulation look up table or the look up table content, namely the set of data, may be updated, i.e. a new or updated set of data may be written on.

According to another exemplary embodiment of the predistortion unit the predistortion unit is adapted to swap between a using of the first amplitude modulation look up table and the second amplitude modulation look up table and/or between a using of the first phase modulation look up table and the second phase modulation look up table.

That is, the predistortion unit may switch between the using of the first set of data and the third set of data and/or the using of the second set of data and the fourth set of data when performing the predistortion. In particular, the swapping may be performed in a smooth manner. That is, the swapping may be performed in such a way that no or only minor discontinuity in the predistorted amplitude and/or predistorted phase signal arises.

According to another exemplary embodiment of the transmitter the local memory is adapted to store the representation of the amplitude predistortion function and/or the representation of the phase predistortion function as coefficients and/or as full sets of data. That is, the local memory may be adapted to store either full sets of data so that the transmitter is suitable for being used in a case full sets of data are received by the transmitter for representing the predistortion functions, e.g. in case the respective interface has a high bandwidth, and/or may be suitable for being used in a case the predistortion functions are received by the transmitter in a compressed manner, e.g. by coefficients characterizing the predistortion functions.

According to another exemplary embodiment the transmitter further comprises a coordinate rotation digital computer adapted to generate the nominal amplitude signal and/or the nominal phase signal from a quadrature signal.

According to another exemplary embodiment the transmitter further comprises a modulator adapted to provide the quadrature signal to the coordinate rotation digital computer.

Summarizing, an exemplary aspect of the invention may be seen in providing a predistortion unit for predistorting digital nominal amplitude and nominal phase signals by using sets of data stored in look up table memories, wherein the sets of data represent full sets of data, i.e. for each input value a corresponding output value is stored, so that no real time calculation is necessary. The calculation of the sets of data may be performed offline in advance by a LUT computer of the transmitter. The LUT computer may be implemented in software on a processor which may lead to an increased flexibility. In case predistortion functions for calculating the sets of data are transmitted in a compressed state, i.e. in the form of coefficients, between an external memory on a baseband chip and the transmitter the restrictions imposed to the local memory size of the transmitter and the possible load of the respective digital interface between the baseband chip and the transmitter may be lessened. Such a predistortion unit may consider the following aspects:

• the predistortion unit may be flexible enough to support PAs from different vendors although the shape of g(.) and G(.) may differ significantly between PAs from different vendors, wherein g(.) and G(.) represent the distortion functions which are compensated for by the predistortion functions.

• g (.) and G(.) may depend significantly on several parameters such as frequency, temperature, variable biasing (to meet the full range of required average output power due to uplink power control); this may lead to a multidimensional set of predistortion functions • the predistortion unit may be reusable in multi mode transmitters that can be reconfigured to different standards

• high sampling rates as they are encountered in high bandwidth standards such as UMTS WCDMA may be handled

• system partitioning:

- large baseband memory

- limited memory in the transceiver, i.e. receiver + transmitter

- digital interface between baseband and transceiver, such as specified by the DigRF 2.5 G standard

• system constraints:

- limited bandwidth (BW) of digital interface between baseband and transceiver: 26Mbit/sec at most as specified by the DigRF 2.5G standard

- autonomy of the transceiver (independence from baseband chip) may be enabled.

The exemplary embodiments and aspects defined above and further aspects of the invention are apparent from the examples of embodiment to be described hereinafter and are explained with reference to these examples of embodiment. It should be noted that features described in connection with one exemplary aspect or embodiment may also be combined with the features of another exemplary aspect or embodiment.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be described in more detail hereinafter with reference to examples of embodiment but to which the invention is not limited.

Fig. 1 schematically shows a polar transmitter.

Fig. 2 schematically shows a predistortion unit according to an exemplary embodiment of the invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

The illustration in the drawing is schematically. In different drawings, similar or identical elements are provided with the same reference signs.

In the following, referring to Figs. 1 to 2 some basic principles of data processing in a device according to exemplary embodiments will be explained.

Fig. 1 serves for explaining some principles of predistorting in a transmitter and schematically shows a polar transmitter. In particular, Fig. 1 outlines a polar transmitter

100 for 8PSK modulation according to the GSM/EDGE standard. However, the invention is not limited to an 8PSK modulation.

In the following the main building blocks are discussed. The 8PSK modulator

101 is fed with octernary input symbols 150 at a rate of about 270 kHz. It produces I&Q output samples 151 and 152 at a rate of about 4.33 MHz, corresponding to an oversampling factor of 16.

The 8PSK modulator 101 can be configured to ramp the envelope of the I&Q signal up and down smoothly as it is suitable at the begin and end of time slot in a time division multiple access (TDMA) system such as GSM/EDGE. The I&Q signal may be normalized, meaning that its average power within the time slot is constant and no scaling is applied in order to affect the average RF power of the polar transmitter.

A coordinate rotation digital computer (CORDIC) 102 converts the I&Q signal to the nominal amplitude r₀ 153 and phase signals Φo 154. A predistortion unit 103 scales the nominal amplitude r₀ provided by the CORDIC by a factor k which chooses the power of the power amplifier (PA) output signal. Further, two memoryless nonlinear functions r₂ = H(k-r₀) and ΔΦ₂ = g(H(k-r₀)) are computed and ΔΦ₂ is subtracted from the nominal phase 155 provided by the CORDIC. HQ and g(.) are determined by the PA as will be explained below. The two output signals of the predistortion unit form the predistorted amplitude r₂ 156 and phase Φ₂ 157.

The dashed box 104 at the bottom left in Fig. 1 includes a differentiator 105, an FM DAC 106, a 2-point PLL 107 and further functional blocks which are not mentioned specifically. The input to this black box is the digital predistorted phase Φ₂ 157 and an output 158 is a constant envelope RF carrier that is modulated by a phase Φ₃ with Φ₃ = Φ₂. The dashed box at the bottom right of Fig. 1 depicts a behavioural model of a power amplifier (PA) 108. The AM is imposed via r₃, an amplitude control input signal to the PA that is typically the PA's supply voltage. For simplicity r₃ = r₂ is assumed. In the ideal case, the PA would act like a multiplier with a constant conversion factor between the amplitude control input r₃ 159 and the amplitude of the PA's RF output signal r₄ 160. In practice, this conversion factor is a non linear function r₄ = G(r₃) that is denoted as the AM/AM characteristic. In addition, the PA's phase shift is a non linear function ΔΦ₄ = g(r₃) 161 of the amplitude control signal. g(.) is denoted as the AM/PM characteristic.

Assuming that the distortion functions g(.) and G(.) are known, their effect can be cancelled or at least mitigated by means of predistortion. The predistortion unit has to simply imitate the function g(.) and the inverse function of G(.) which is denoted HQ.

Fig. 2 schematically shows a predistortion unit according to an exemplary embodiment of the invention which may be used as a predistortion unit 103 in the transmitter 100 shown in Fig. 1. The predistortion unit 103 comprises a pair of AM LUT 201 and 202, a pair of PM LUT 203 and 204, which are all connected to a LUT computer 205. Furthermore, the predistortion unit comprises a delay 206.

In particular, Fig. 2 outlines an implementation of the predistortion unit 103 according to an exemplary embodiment. The hatched parts represent the synchronous data path that is clocked at the sampling frequency, i.e. 4.33MHz in this example.

The AM LUT 201, 202 and PM LUT 203, 204 may be formed by RAMs acting as look up tables and hold the functions HQ and g(HQ) shown in Fig. 1. The delay 206 may simply be a register compensating the delay introduced by the PM LUT 203 and 204. Preferably there are two instances of both AM LUT and PM LUT in order to enable seamless and smooth updating: The on line instances of the LUTs are in read mode and act as the actual LUTs. The offline instances of the LUTs are written with new data under control of a LUT computer 205. The offline instances and the LUT computer are not drawn hatched because they do not form part of the synchronous data path and can be clocked at arbitrary rates. Once the filling of the offline instances with new data is finished, offline and on line instances can be swapped when required.

The LUT computer may essentially perform an expansion of compressed data. It computes the LUT contents from AM&PM coefficients that are stored in a local RAM 207 of the transmitter. The coefficients are worth a few dozens of Bytes while a typical LUT is worth about 1.5 kByte of RAM if 10 bit address space and a data word length of 12 bit or 1.5 Byte is used.

This data expansion (and previous compression) may be performed due to two main reasons:

• local RAM in the transmitter for multiple sets of raw, i.e. uncompressed, LUT data may be too large,

• insufficient bandwidth of digital interface between a baseband (BB) chip 208 and transmitter

It should be mentioned again that a multidimensional array of predistortion functions reflecting the PA' s dependency on parameters such as frequency, temperature and power level specific biasing may be necessary. If these predistortion functions were stored as raw images of the LUT content, this would require an amount of memory that may not be affordable in terms of local RAM in the transmitter. The raw LUT images may be stored in the large external memory of the BB chip but the transfer to the transmitter may not feasible due to the limited bandwidth of typical digital interfaces such as 26 Mbit/sec of the DigRF 2.5G interface.

A simple example for a data expansion algorithm is linear interpolation:

Samples from the predistortion curve represent compressed data and linear interpolation expands them to the full curves. The data expansion algorithm may be implemented in dedicated HW or it may be implemented in SW that is executed by a processor. In particular, the expansion is performed by the LUT computer in a predistortion unit according to an exemplary embodiment of the invention. It should be noted that a programmable processor may leave more freedom in the choice of the PA and in designing customized compression/expansion algorithms.

Thanks to compression, it may be possible to hold the AM&PM coefficients in a data base in the external memory of the baseband chip and send the respective set of AM&PM coefficients to the transmitter just in time. The small local RAM 207 in the transmitter holding the AM&PM coefficients may be double buffered although it may not be necessary because the data transfer from the BB to the transmitter occurs much faster than the computation of the new LUT content. Updating the LUTs will in general cause a discontinuity in the amplitude and phase signal resulting in unwanted spectral broadening of the TX signal. This may be avoided by using the envelope ramping feature of the 8PSK modulator by running through the following sequence:

• ramp the I&Q envelope down

• switch the LUTs

• if required, switch the PA biasing or other analogue parameters of the TX path

• ramp the I&Q envelope up

Although the invention is described above with respect to a polar transmitter having a 8PSK modulation, which is adopted by the GSM/EDGE standard, the invention is not limited to this modulation. Exemplary embodiments of the invention may be applicable in polar transmitters for arbitrary modulation schemes and standards.

Compared to a real time processing for predistortion by using coefficients and using linear interpolation between samples defining the coefficients the use of look up tables may have some advantages. Firstly, the computational load and hence the power consumption may be lower, since a linear interpolation has to be applied in real time to all samples of a transmitted signal, while the LUT computer instead may only compute the content of the LUT memories. For instance, an amplitude signal r₀ with 10 bit resolution demands for a LUT with 1024 entries. If this LUT is applied to more than 1024 samples, it may be computational more efficient than real time interpolation. This may particular be attractive for high bandwidth systems with high sampling rates because the maximum update rate for the LUTs (due to power level changes under uplink power control) is similar for most cellular standards and does not exceed a few kHz.

Secondly, due to the moderate computational complexity of the LUT computation, it may be possible to implement the LUT computer in software on a processor. This may mean a major advantage of the LUT approach according to the invention in terms of flexibility. If the shapes of the required predistortion functions turn out to be quite curvy, equidistant linear interpolation may require too many samples for just in time delivery via the DigRF interface with its limited bandwidth. The way out may be more powerful compression/expansion schemes that may be easily implemented in SW. Candidates include non equidistant linear interpolation and (non equidistant) cubic spline interpolation. The LUTs may be larger than dedicated HW for real time computation but the size in the context of an integrated transmitter with its coils and capacitors may still be acceptable.

Summarizing, a gist of an exemplary aspect of the invention may be seen in providing a method of predistorting and a predistortion unit which uses relatively large look up table memories in order to store full data sets which represents predistortion functions. Since full data sets are used for predistortion it may be possible to avoid real time calculation for predistorting signals. Thus, amplitude modulation and phase modulation distortions which influence polar transmitters may be cancelled or at least mitigated by applying a predistortion, e.g. by using a non linear predistortion function to the amplitude and phase signal in the digital baseband domain. Arbitrary predistortion functions may be supported for application in arbitrary wireless standards such as GSM/EDGE and UMTS.

It should be noted that the term "comprising" does not exclude other elements or features and the "a" or "an" does not exclude a plurality. Also elements described in association with different embodiments or aspects may be combined. It should also be noted that reference signs in the claims shall not be construed as limiting the scope of the claims.

Claims

CLAIMS:

1. A predistortion unit (103) for a polar transmitter (100), the predistortion unit (103) comprises: a first amplitude modulation look up table memory (201) storing a first set of data representing an amplitude predistortion function, and a first phase modulation look up table memory (203) storing a second set of data representing a phase predistortion function, wherein the predistortion unit (103) is adapted to predistort nominal amplitude signal and nominal phase signal of a signal by applying the first set of data to the nominal amplitude signal and the second set of data to the nominal phase signal.

2. The predistortion unit (103) according to claim 1, further comprising: a look up table computer (205) adapted to compute the first set of data and/or the second set of data.

3. The predistortion unit (103) according to claim 2, wherein the look up table computer (205) is adapted to perform the computation of the first set of data and/or the second set of data off-line.

4. The predistortion unit (103) according to claim 2, wherein the look up table computer (205) is adapted to receive a signal representing the amplitude predistortion function and/or the phase predistortion function.

5. The predistortion unit (103) according to claim 2, wherein the computation of the first set of data and/or the second set of data comprises a decompression of a compressed signal.

6. The predistortion unit (103) according to claim 1, further comprising: a second amplitude modulation look up table memory (202) storing a third set of data representing an amplitude predistortion function, and a second phase modulation look up table memory (204) storing a fourth set of data representing a phase predistortion function.

7. The predistortion unit (103) according to claim 6, wherein the predistortion unit is (103) adapted to write the third set of data to the second amplitude modulation look up table (202) while the first set of data is used to predistort the nominal amplitude signal, and/or wherein the predistortion unit is adapted to write the fourth set of data to the second phase modulation look up table (204) while the second set of data is used to predistort the nominal phase signal.

8. The predistortion unit (103) according to claim 6, wherein the predistortion unit is (103) adapted to swap between a using of the first amplitude modulation look up table (201) and the second amplitude modulation look up table (202) and/or between a using of the first phase modulation look up table (203) and the second phase modulation look up table (204).

9. A transmitter (100) comprising : a predistortion unit (103) according to claim 1, and a local memory (207), wherein the local memory (207) is adapted to store a representation of the amplitude predistortion function and/or a representation of the phase predistortion function.

10. The transmitter according to claim 9, wherein the local memory is adapted to store the representation of the amplitude predistortion function and/or the representation of the phase predistortion function as coefficients and/or as full sets of data.

11. The transmitter (100) according to claim 9, further comprising: a coordinate rotation digital computer (102) adapted to generate the nominal amplitude signal and/or the nominal phase signal from a quadrature signal.

12. The transmitter (100) according to claim 11, further comprising: a modulator (101) adapted to provide the quadrature signal to the coordinate rotation digital computer (102).

13. A method of predistorting a nominal amplitude signal and a nominal phase signal, the method comprising: applying a nominal amplitude signal and a nominal phase signal to a predis- tortion unit (103), applying a first set of data to the nominal amplitude signal and a second set of data to the nominal phase signal, wherein the first set of data represents an amplitude predistortion function, and wherein the second set of data represents a phase predistortion function.

14. A program element, which, when being executed by a processor, is adapted to control or carry out a method according to claim 13.

15. A computer-readable medium, in which a computer program is stored which, when being executed by a processor, is adapted to control or carry out a method according to claim 13.