(19)
(11)EP 3 005 553 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
14.08.2019 Bulletin 2019/33

(21)Application number: 14730587.4

(22)Date of filing:  15.05.2014
(51)Int. Cl.: 
H03F 1/32  (2006.01)
H03F 3/24  (2006.01)
(86)International application number:
PCT/IB2014/061472
(87)International publication number:
WO 2014/188316 (27.11.2014 Gazette  2014/48)

(54)

MULTI-BAND RADIO-FREQUENCY DIGITAL PREDISTORTION

MEHRBANDIGE DIGITALE HOCHFREQUENZ-VORVERZERRUNG

PREDISTORSION NUMERIQUE RADIOFREQUENCE MULTIBANDE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 24.05.2013 US 201313902023

(43)Date of publication of application:
13.04.2016 Bulletin 2016/15

(73)Proprietor: Telefonaktiebolaget LM Ericsson (publ)
164 83 Stockholm (SE)

(72)Inventor:
  • ROLLINS, Mark Edward
    Stittsville, Ontario K2S 2B4 (CA)

(74)Representative: Zacco Sweden AB 
Valhallavägen 117 Box 5581
114 85 Stockholm
114 85 Stockholm (SE)


(56)References cited: : 
US-A1- 2003 179 831
US-A1- 2013 064 325
US-A1- 2012 128 099
US-B1- 8 391 809
  
  • SUNGHO CHOI ET AL: "Design of digital predistorters for wideband power amplifiers in communication systems with dynamic spectrum allocation", ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2011 IEEE INTERNATIONAL CONFERENCE ON, IEEE, 22 May 2011 (2011-05-22), pages 3204-3207, XP032001036, DOI: 10.1109/ICASSP.2011.5946703 ISBN: 978-1-4577-0538-0
  • SEYED AIDIN BASSAM ET AL: "Channel-Selective Multi-Cell Digital Predistorter for Multi-Carrier Transmitters", IEEE TRANSACTIONS ON COMMUNICATIONS, IEEE SERVICE CENTER, PISCATAWAY, NJ. USA, vol. 60, no. 8, 1 August 2012 (2012-08-01) , pages 2344-2352, XP011456615, ISSN: 0090-6778, DOI: 10.1109/TCOMM.2012.051012.100285A
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

Related Applications



[0001] This application claims the benefit of US patent application US 2014/348263 A1, filed May 24, 2013.

Field of the Disclosure



[0002] The present disclosure relates to digital predistortion and more specifically relates to digital predistortion in a multi-band transmitter.

Background



[0003] Modern wireless communication systems employing wideband Power Amplifiers (PAs) must be operated close to saturation to maximize power efficiency. This introduces significant non-linear signal distortion to wideband signals such as Code Division Multiple Access (CDMA) and Orthogonal Frequency Division Multiplexing (OFDM) signals due to their large peak-to-average power ratio characteristics. Techniques to mitigate this non-linear distortion include:
  • Backing off the PA operating point. However, doing so leads to larger size and cost to yield the same output power.
  • Feed-forward cancellation of the non-linear distortion with analog combining networks. This is effective but incurs the additional cost of feed-forward cancellation circuits and requires an additional low-power but highly linear Radio Frequency (RF) amplifier.
  • Baseband Digital Predistortion (BB-DPD), which employs Digital Signal Processing (DSP) techniques to impress an "inverse characteristic" of the PA on the transmitted signal to compensate for the non-linear distortion introduced by the PA.
BB-DPD has become by far the preferred approach for managing PA nonlinearities.

[0004] Utilizing adaptive BB-DPD to compensate for the non-linearity of the PA is a proven technology that enables high linearity, high efficiency power PA subsystems for single-band transmitters. However, in applications such as base stations of a cellular communications network that support multiple radio access technologies or multiple bands for the same radio access technology, multi-band transmitters are desirable. Conventional BB-DPD systems are not optimal for multi-band transmitters.

[0005] In order to address the complexity and power consumption of digital BB-DPD, RF Analog Predistortion (RF-APD) for single-band signals has been proposed by entities such as Scientera. RF-APD technology predistorts for non-linear distortion using analog delay cells and Gilbert cell multipliers. It is difficult to design these circuit elements with sufficiently wide bandwidth to address multi-band signals, particularly those with a wide separation between individual bands. An academic example of RF-APD, Yi et. al., "Analog Predistortion Linearizer for High-Power RF Amplifiers," IEEE Transactions on Microwave Theory and Techniques, Vol. 48, No. 12, December 2000, illustrates the technical challenge of implementing RF-APD for a single-band signal of only 30 megahertz (MHz) bandwidth. Finally, analog circuits suffer from thermal, voltage, and semiconductor process variations that must be compensated for using sophisticated on-line calibration circuitry.

[0006] In addition, an academic example of digital predistortion at RF for a single-band signal is described in Mark Sterling et al., "Direct Digital Predistortion on a Computer Controlled FPGA," IEEE International Conference on Acoustics, Speech and Signal Processing, 2007, Vol. 2, pp.II-369, 11-372, April 15-20, 2007. Here, the real-valued digital RF signal is predistorted by passing it through the following digital operations: a cubic non-linearity (implemented as a lookup table) followed by a gain adjustment followed by a time delay. This method makes no attempt to manage the digital aliasing of third order terms generated from cubing the single-band signal. Managing this aliasing becomes infeasible for multi-band signals whose third order distortion products can alias into signal bands of interest within the active Nyquist zone.
As such, there is a need for systems and methods for digital predistortion in a multi-band transmitter.

[0007] Document US 2003/179831 A1 may be construed to disclose methods of balancing network packet traffic among multiple checking functionalities (CFs). A network has at least one client operatively connected to at least one source switch and multiple available CFs operatively connected to at least one destination switch. Each available CF has predetermined, but possibly different inspection capabilities. A source switch receiving packets from a client inspects each packet and can optionally choose an available CF having at least the minimum necessary inspection capabilities to inspect the particular packet, and tunnel the packet to the chosen CF.

[0008] Document US 2013/064325 A1 may be construed to disclose a multi-band signal being generated by combining two or more input signals separated in frequency. The input signals are combined either before or after predistortion depending on the bandwidth of the multi-band signal. If the bandwidth of the multi-band signal is less than a predetermined bandwidth threshold, the input signals are combined and predistortion is applied to the combined signal to generate the multi-band signal. If the bandwidth of the multi-band signal is greater than the bandwidth threshold, the individual input signals are predistorted and subsequently combined to generate the multi-band signal.

[0009] Document "Design of digital predistorters for wideband power amplifiers in communication systems with dynamic spectrum allocation", Sungho Choi et.al., Transactions on IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), pages 3204 to 3207, may be construed to disclose a predistortion technique for dynamic spectrum allocation systems such as cognitive radio (CR). The system model considered occupies a small band at a time, but the center frequency can be changed in the wide range of frequency. In this scenario, the front-end filter may not eliminate the harmonics of the power amplifier (PA) output. The PD reduces the spectral regrowth of the fundamental signal at the carrier frequency (ωo) and removes the harmonics (2ωo, 3ωo, ···) at the same time. The PD structure is composed of multiple predistorters (PDs) centered at integer multiples of ωo. The PD at ωo is for removing spectral regrowth of the fundamental signal, and the others are for harmonic reduction. In the PD structure, parameters of PDs are found jointly.

[0010] Document US 2012/128099 A1 may be construed to disclose predistorted transmitters operable over a wide range of frequencies including a plurality of predetermined frequency bands. The transmitters include a programmable digital up-converter and a programmable digital down-converter, an ADC, a DAC, a power amplifier and at least one analog filter arranged along a transmit signal path and a feedback signal path.

[0011] Document "Channel-Selective Multi-Cell Digital Predistorter for Multi-Carrier Transmitters", Seyed Aidin Bassam et.al., IEEE TRANSACTIONS ON COMMUNICATIONS, vol. 60, no. 8, pages 2344 to 2352, may be construed to disclose a channel-selective multi-cell processing predistortion technique that compensates for the nonlinearities of multi-carrier transmitters. The technique uses independent processing cells to compensate for the intra-band and inter-band distortions of nonlinear transmitters. This frequency-selective feature of the technique significantly reduces the minimum sampling rate requirements of analog-to-digital and digital-to-analog converters, which are a critical issue for conventional digital predistortion (DPD) techniques dealing with wideband signals.

[0012] Document US 8,391,809 B1 may be construed to disclose a system for multi-band predistortion having a predistortion actuator applying a first and a further adjustable distortion to a first band input and at least one further band input, respectively, based on the first band input and at the least one further band input to provide a first band output and a second band output, respectively. The predistortion actuator drives an output port configured to be coupled to an amplifier. The system further has a processor coupled to an evaluation input, which is couplable to an output of the amplifier. A coefficient calculation block is coupled to a processor output, the first band input, and the at least one further band input. The coefficient calculation block calculates first and the further adjustable distortion coefficients based on the first band input, the at least one further band input and the processor output.

Summary



[0013] According to the invention, there are provided a method and an apparatus according to the independent claims. Developments are set forth in the dependent claims.

[0014] Systems and methods for Radio Frequency Digital Predistortion (RF-DPD) in a multi-band transmitter are disclosed. In one embodiment, the multi-band transmitter includes a digital upconversion system configured to digitally upconvert digital input signals to provide digital radio frequency (RF) signals. Each digital input signal, and thus each digital RF signal, corresponds to a different band of a multi-band transmit signal to be transmitted by the multi-band transmitter. The multi-band transmitter also includes a RF-DPD system configured to digitally predistort the digital RF signals at RF to provide predistorted digital RF signals, and to directly synthesize intra-band and inter-band distortion products, and a combiner circuitry configured to combine the predistorted digital RF signals to provide a multi-band predistorted digital RF signal. By performing digital predistortion at RF, sampling rates required for digital upconversion of each of the digital input signals is substantially reduced as compared to a similar system in which digital predistortion is performed at baseband prior to digital upconversion. In addition, complexity of the digital upconversion is reduced in that there is no need to digitally upconvert inter-band distortion products as needed in a similar system that utilizes baseband digital predistortion.

[0015] Preferably, the multi-band transmitter further includes a multi-band signal processing system configured to process the multi-band predistorted digital RF signal to provide a multi-band predistorted analog RF signal and Power Amplifier (PA) circuitry configured to amplify the multi-band predistorted analog RF signal to provide the multi-band transmit signal. Preferably, the multi-band predistorted digital RF signal is a complex signal, and the multi-band signal processing system includes circuitry configured to receive the multi-band predistorted digital RF signal and output a real part of the multi-band predistorted digital
RF signal, digital-to-analog conversion circuitry configured to convert the real part of the multi-band predistorted digital RF signal from digital to analog and output a corresponding analog signal, and filtering circuitry configured to filter the analog signal to thereby provide the multi-band predistorted analog RF signal. Preferably, the multi-band predistorted digital RF signal is a real signal, and the multi-band signal processing system comprises digital-to-analog conversion circuitry configured to convert the multi-band predistorted digital RF signal from digital to analog and output a corresponding analog signal and filtering circuitry configured to filter the analog signal output by the digital-to-analog conversion circuitry (100) to provide the multi-band predistorted analog RF signal.

[0016] Preferably, the digital upconversion system includes a separate digital upconversion subsystem for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter. If so, the separate digital upconversion subsystem for the band of the multi-band transmit signal comprises, for each band of the multi-band transmit signal, complex baseband tuning circuitry configured to perform complex tuning of the digital input signal for the band to a desired baseband tuning frequency to thereby provide a complex tuned digital signal, up-sampling circuitry configured to upsample the complex tuned digital signal to a desired up-sampling rate to thereby provide an upsampled digital signal, and image selection filtering circuitry configured to filter the upsampled digital signal to select a desired image of the complex tuned digital signal and thereby provide the digital RF signal for the band of the multi-band transmit signal.. Preferably, in this case, for each band of the multi-band transmit signal, the separate digital upconversion subsystem for the band of the multi-band transmit signal further comprises a rate change filter [0010a] Preferably, the RF digital predistortion system comprises a separate RF digital predistortion subsystem for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter. If so, for each band of the multi-band transmit signal, the separate RF digital predistortion subsystem for the band of the multi-band transmit signal is preferably configured to directly synthesize desired self, intra-band, and inter-band distortion products for the band based on the plurality of digital RF signals to thereby provide a corresponding one of the plurality of predistorted digital RF signals.

[0017] Alternatively, the multi-band transmitter preferably further includes a training system configured to train the RF-DPD system. Preferably, the training system is configured to train models utilized by the RF-DPD system to predistort the digital RF signals. Preferably, the training system includes a separate training subsystem for each band of the multi-band transmit signal.

[0018] In one embodiment, a method of operation of a multi-band transmitter includes digitally upconverting digital input signals to provide digital RF signals, where each digital input signal, and thus each digital RF signal, corresponds to a different band of a multi-band transmit signal to be transmitted by the multi-band transmitter. The method further includes digitally predistorting the digital RF signals at RF to provide predistorted digital RF signals, wherein digitally predistorting (1002) the plurality of digital RF signals comprises directly synthesizing intra-band and inter-band distortion products, and combining the predistorted digital RF signals to provide a multi-band predistorted digital RF signal.

[0019] Preferably, the method further includes processing the multi-band predistorted digital RF signal to provide a multi-band predistorted analog RF signal, and amplifying the multi-band predistorted analog RF signal to provide the multi-band transmit signal.

[0020] Preferably, the multi-band predistorted digital RF signal is a complex signal, and processing the multi-band predistorted digital RF signal includes obtaining a real part of the multi-band predistorted digital RF signal, converting the real part of the multi-band predistorted digital RF signal from digital to analog to thereby provide a corresponding analog signal, and filtering the corresponding analog signal to thereby provide the multi-band predistorted analog RF signal. Preferably, the multi-band predistorted digital RF signal is a real signal, and processing the multi-band predistorted digital RF signal includes converting the multi-band predistorted digital RF signal from digital to analog to thereby provide a corresponding analog signal and filtering the corresponding analog signal to thereby provide the multi-band predistorted analog RF signal.

[0021] Preferably, digitally upconverting the digital input signals includes separately digitally upconverting the digital input signals for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter. If so, digitally predistorting the digital RF signals preferably includes separately digitally predistorting the digital RF signals for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter. Preferably, separately digitally predistorting the plurality of digital RF signals comprises, for each band of the multi-band transmit signal, directly synthesizing desired self, intra-band, and inter-band distortion products for the band based on the plurality of digital RF signals to thereby provide a corresponding one of the plurality of predistorted digital RF signals.

[0022] Alternatively, the method preferably further includes training models utilized for digitally predistorting the digital RF signals. Preferably, training the models utilized for digitally predistorting the digital RF signals includes separately training the models for corresponding bands of the multi-band transmit signal.

[0023] Those skilled in the art will appreciate the scope of the present disclosure and realize additional aspects thereof after reading the following detailed description of the preferred embodiments in association with the accompanying drawing figures.

Brief Description of the Drawing Figures



[0024] The accompanying drawing figures incorporated in and forming a part of this specification illustrate several aspects of the disclosure, and together with the description serve to explain the principles of the disclosure.

Figure 1 illustrates a single-band transmitter utilizing conventional Baseband Digital Predistortion (BB-DPD);

Figures 2A through 2C illustrate a single-carrier signal, a single-band signal, and a multi-band signal, respectively.

Figure 3 illustrates a dual-carrier single-band transmitter utilizing conventional BB-DPD;

Figure 4 illustrates a dual-carrier dual-band transmitter utilizing conventional BB-DPD;

Figure 5 illustrates a dual-band transmitter including RF Digital Predistortion (RF-DPD) according to one embodiment of the present disclosure;

Figure 6 illustrates a more general multi-band transmitter including RF-DPD according to one embodiment of the present disclosure;

Figure 7 is a flow chart that illustrates the operation of the multi-band transmitter of Figure 6 according to one embodiment of the present disclosure; and

Figure 8 is a flow chart that illustrates the operation of the training system of Figure 6 according to one embodiment of the present disclosure.


Detailed Description



[0025] The embodiments set forth below represent the necessary information to enable those skilled in the art to practice the embodiments and illustrate the best mode of practicing the embodiments. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the disclosure and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure and the accompanying claims.

[0026] Systems and methods for Radio Frequency Digital Predistortion (RF-DPD) in a multi-band transmitter are disclosed. Before discussing embodiments of a multi-band transmitter including RF-DPD, a discussion of conventional Baseband Digital Predistortion (BB-DPD) is beneficial. In this regard, Figure 1 illustrates a transmitter 10 utilizing conventional BB-DPD. A complex-valued digital baseband signal (n) from a source 12 is predistorted by a BB-DPD actuator 14 using complex-valued tap weights {aqm} computed by a training subsystem 16 to thereby provide a predistorted digital baseband signal (n). The predistorted digital baseband signal (n) is upconverted to radio frequency and Digital-to-Analog (D/A) converted by an upconversion subsystem 18 to thereby provide a predistorted radio frequency analog signal. In one implementation, the upconversion subsystem 18 includes digital upconversion circuitry that upconverts the predistorted digital baseband signal (n) in the digital domain followed by a wideband Radio Frequency (RF) Digital-to-Analog Converter (DAC) as described in U.S. Patent Application Publication No. 2010/0098191 A1, entitled METHODS AND SYSTEMS FOR PROGRAMMABLE DIGITAL UP-CONVERSION, filed on October 20, 2008, and published on April 22, 2010. The predistorted radio frequency analog signal is amplified by a Power Amplifier (PA) 20 to thereby provide a real-valued transmit signal z(n) at the output of the PA 20.

[0027] In general, the training subsystem 16 computes the tap weights {aqm}, and thereby trains the BB-DPD actuator 14, such that the predistortion introduced by the BB-DPD actuator 14 compensates for non-linearity of the PA 20 (i.e., the predistortion is an inverse of distortion resulting from the non-linearity of the PA 20). More specifically, a downconversion and digitization subsystem 22, which can be referred to as a transmit observation receiver (TOR), receives the transmit signal z(n) from the output of the PA 20 via a coupler 24 and downconverts and digitizes the transmit signal z(n) to thereby provide a feedback signal that is representative of the transmit signal z(n). A combiner 26, or in this case a subtraction node, outputs an error signal that is equal to a difference between the digital baseband signal (n) and the feedback signal. The training subsystem 16 uses a training procedure that observes the error signal and adaptively configures the tap weights {aqm} to minimize the error signal (i.e., to make the error signal zero).

[0028] Typically, the BB-DPD actuator 14 implements a predistortion function based on a Volterra representation of the non-linear system, or its simplified "Memory Polynomial" (MP) variant, since the tap weights {aqm} may be optimized by solving a linear system of equations. Equation 1 below shows how to
compute the predistorted digital baseband signal (n) from the digital baseband input signal (n) and the set of tap weights {aqm} for the MP based actuator. Typically, the BB-DPD actuator 14 includes a set Q = {1,3,5} of non-linear terms in practice. Also, the BB-DPD actuator 14 may use M = 0 for the case where the PA 20 response exhibits a memoryless non-linear characteristic, or some suitable non-zero value of M based on the worst-cast extent of the PA 20 memory otherwise.



[0029] Often in practice, the digital baseband signal (n) is a complex-valued Code Division Multiple Access (CDMA) or Orthogonal Frequency Division Multiplexing (OFDM) signal on a single carrier in a single RF band, such that the real-valued transmit signal z(n) is given by

In this case, the transmitter 10 in Figure 1 is referred to as a "single-carrier" system. The transmitted radio spectrum in this case is shown in Figure 2A.

[0030] Alternately, the digital baseband input signal (n) may consist of a set of multiple carriers transmitted in a contiguous set of carriers, or a "single-band" as shown in Figure 2B. In this case, the real-valued transmit signal z(n) is given by Equation 2, where a total of S carriers are transmitted in the single-band centered at carrier frequency ωo. Within that band, the carrier frequency for the s-th carrier is given by ωo + Δz.



[0031] Finally, the digital baseband signal (n) may consist of multiple disjoint sets of contiguous carriers transmitted in non-contiguous bands, or a "multi-band" scenario as shown in Figure 2C. In this case, the real-valued transmit signal z(n) is given in Equation 3, where there are B bands, with Sb carriers transmitted in the b-th signal band.



[0032] Figure 3 is a block diagram of a transmitter 28 that implements BB-DPD for a multi-carrier, single-band system. In this example, there are two carriers. Here, two digital baseband signals from corresponding sources 30-1 and 30-2 are complex tuned to desired frequencies Δ1 and Δ2 by corresponding complex tuners 32-1 and 32-2. The complex tuned digital baseband signals are then combined by a combiner 34, or in this case a summation node, to provide a multiple-carrier, single-band digital baseband input signal (n). From this point, the transmitter 28 operates in the same manner as described above with respect to the transmitter 10. Specifically, the digital baseband input signal (n), which is complex-valued, is predistorted by a BB-DPD actuator 36 using complex-valued tap weights {aqm} computed by a training subsystem 38 to thereby provide a predistorted digital baseband signal (n). The predistorted digital baseband signal (n) is upconverted to radio frequency and D/A converted by an upconversion subsystem 40 to thereby provide a predistorted radio frequency analog signal. In one implementation, the upconversion subsystem 40 includes digital upconversion circuitry that upconverts the predistorted digital baseband signal (n) in the digital domain followed by a wideband RF DAC as described in U.S. Patent Application Publication No. 2010/0098191 A1. The predistorted radio frequency analog signal is amplified by a PA 42 to thereby provide a real-valued transmit signal z(n) at the output of the PA 42.

[0033] The training subsystem 38 computes the tap weights {aqm} and thereby trains the BB-DPD actuator 36 such that the predistortion introduced by the BB-DPD actuator 36 compensates for non-linearity of the PA 42 (i.e., the predistortion is an inverse of distortion resulting from the non-linearity of the PA 42). More specifically, a downconversion and digitization subsystem 44, which can be referred to as a TOR, receives the transmit signal z(n) from the output of the PA 42 via a coupler 46 and downconverts and digitizes the transmit signal z(n) to thereby provide a feedback signal that is representative of the transmit signal z(n). A combiner 48, or in this case a subtraction node, outputs an error signal that is equal to a difference between the digital baseband input signal (n) and the feedback signal. The training subsystem 38 uses a training procedure that observes the error signal and adaptively configures the tap weights {aqm} to minimize the error signal (i.e., to make the error signal zero). In this way, Equation 1 above applies equally to a single carrier scenario and a multi-carrier scenario and, for the multi-carrier scenario, produces the correct distortion products provided that the multi-carrier spectrum is centered properly at baseband.

[0034] The predistorted digital baseband signal (n) for the multi-carrier, single-band system contains cross-products of the form 1(n - m)|z(n - m)|q-1 and z(n - m)|ũ1(n - m)|q-1 in addition to the self-products 1(n - m)|1(n - m)q-1 and 2(n - m)|2(n - m)|q-1, as is necessary for effective predistortion. By operating on an output of the combiner 34 (i.e., by operating on the summed multiple carrier digital baseband input signal (n)), the BB-DPD actuator 36 generates these cross products automatically. Based on the non-linear products of Equation 1, the predistorted digital baseband signal (n) typically contains 3rd and 5th order distortion products. Given that the multiple carrier, single-band digital baseband input signal (n) exhibits a bandwidth W as shown in Figure 3, it follows that the BB-DPD actuator 36 must operate with a sampling rate of 5W or higher. As a result, the bandwidth of the multiple carrier, single-band digital baseband input signal (n) directly affects the computational complexity of the BB-DPD actuator 36, as well as that of the following upconversion subsystem 40 (especially for digital upconversion due to this high sampling rate).

[0035] Figure 4 is a block diagram of a transmitter 50 that implements BB-DPD for a multi-carrier, multi-band system. In this example, there are two bands, each having two carriers. Here, each single-band component of the multi-band signal must be tuned to an intermediate frequency (IF) such that a resulting combined multi-band signal lies centered at baseband. In this way, Equation 1 applies equally to a multi-band scenario and a single-band scenario. More specifically, as illustrated, two multi-carrier digital baseband signals each having two carriers are provided by corresponding sources 52-1 and 52-2. The multi-carrier digital baseband signals are complex tuned to desired IF frequencies ωc - ω1 and ω2 - ωc, respectively, by corresponding complex tuners 54-1 and 54-2. The complex-tuned multi-carrier digital baseband signals are then combined by a combiner 56, or in this case a summation node, to provide a multiple-carrier, multi-band digital baseband input signal (n). From this point, the transmitter 50 operates in the same manner as described above with respect to the transmitters 10 and 28.

[0036] Specifically, the multiple carrier, multi-band digital baseband input signal (n), which is complex-valued, is predistorted by a BB-DPD actuator 58 using complex-valued tap weights {aqm} computed by a training subsystem 60 to thereby provide a predistorted digital baseband signal (n). The predistorted digital baseband signal (n) is upconverted to radio frequency and D/A converted by an upconversion subsystem 62 to thereby provide a predistorted radio frequency analog signal. In one implementation, the upconversion subsystem 62 includes digital upconversion circuitry that upconverts the predistorted digital baseband signal (n) in the digital domain followed by a wideband RF DAC as described in U.S. Patent Application Publication No. 2010/0098191 A1. The predistorted radio frequency analog signal is amplified by a PA 64 to thereby provide a real-valued transmit signal z(n) at the output of the PA 64.

[0037] The training subsystem 60 computes the tap weights {aqm} and thereby trains the BB-DPD actuator 58 such that the predistortion introduced by the BB-DPD actuator 58 compensates for non-linearity of the PA 64 (i.e., the predistortion is an inverse of distortion resulting from the non-linearity of the PA 64). More specifically, a downconversion and digitization subsystem 66, which can be referred to as a TOR, receives the transmit signal z(n) from the output of the PA 64 via a coupler 68 and downconverts and digitizes the transmit signal z(n) to thereby provide a feedback signal that is representative of the transmit signal z(n). A combiner 70, or in this case a subtraction node, outputs an error signal that is equal to a difference between the digital baseband input signal (n) and the feedback signal. The training subsystem 60 uses a training procedure that observes the error signal and adaptively configures the tap weights {aqm} to minimize the error signal (i.e., to make the error signal zero).

[0038] The predistorted digital baseband signal (n) contains several types of distortion products including self-products of the form 1,1(n - m)|1,1(n - m)|q-1, intra-band products of the form 1,1(n - m)|1,2(n - m)|q-1, and inter-band products of the form 1,1(n - m)|2,2(n - m)|q-1. By operating on the summed multi-band signal, the BB-DPD actuator 58 generates all of these cross products automatically. Based on the non-linear products of Equation 1, the predistorted digital baseband signal (n) typically contains 3rd and 5th order distortion products. Given the multiple carrier, multi-band digital baseband input signal (n) exhibits a bandwidth WM as shown in Figure 4, it follows that the BB-DPD actuator 58 must operate with a sampling rate of 5WM or higher. Note that WMW due to the fact that each band contains multiple carriers, and there may be a large separation G in frequency between the bands.

[0039] Applying conventional BB-DPD concepts to multi-band systems produces radio architectures with computational complexity that is prohibitive to implement. This complexity occurs due to the large distortion bandwidths that result with multi-band signals of interest and the wide separations in frequency that may exist between bands of interest. More specifically, multi-band signals exhibit a large bandwidth WM requiring distortion bandwidths on the order of 5WM or higher. Assuming that the upconversion subsystem 62 is performed by digital upconversion circuitry followed by a RF DAC as described in U.S. Patent Application Publication No. 2010/0098191 A1, it follows that the minimum sampling rate of all data path operations in the digital upconversion circuitry must also be 5WM or higher if the BB-DPD actuator 58 is implemented at baseband prior to digital upconversion. Certain variants of BB-DPD can be conceived to reduce this impact. For example, architectures in which the 1st, 3rd, and 5th order distortion products of Equation 1 are upconverted separately and then summed at RF would require lower minimum sampling rates (at least for the 1st and 3rd order cases). However, this solution would entail building at least three parallel digital upconversion data paths, where one of the data paths still requires the 5WM or higher sampling rate.

[0040] In addition, multi-band signals may exhibit a large separation in frequency between bands as (shown as G in Figure 4). So, even though the actual signal bandwidth of the occupied carriers may be modest, a large separation G results in a large overall bandwidth WM and, therefore, a large sampling rate of 5WM or higher is required for the digital upconversion. Again, alternative variants of BB-DPD can be conceived to reduce this impact. For systems with a large separation G, it is possible to form the required distortion self-products and intra-band products for each band in isolation directly from a separate BB-DPD actuator driven by the input signal for that band. The system must then provision additional BB-DPD actuators to produce the inter-band distortion products for the multi-band scenario. This solution helps to reduce the complexity of the digital upconversion circuitry since each branch of the digital upconversion circuitry requires a sampling rate of 5W (the worst-case bandwidth of a single-band) instead of 5WM. However, the solution requires additional parallel Digital Radio Frequency (DRF) upconversion data paths to position the inter-band distortion products at the correct carrier frequency. For a multi-band scenario with several disjoint bands, the complexity of these additional digital upconversion paths quickly becomes prohibitive.

[0041] Systems and methods are disclosed herein for addressing the complexity burden of BB-DPD for multi-band signals. In particular, in one embodiment, a transmitter implements Radio Frequency Digital Predistortion (RF-DPD). In this embodiment, DPD is applied directly to a digital RF signal after digital upconversion (referred to herein as DRF upconversion) before D/A conversion. With predistortion performed at RF, the minimal sampling rate required for DRF upconversion is substantially reduced as compared to the multi-carrier, multi-band BB-DPD scenario discussed above. In addition, there is no need for additional DRF upconversion data paths to upconvert inter-band distortion products as is needed with the multi-carrier, multi-band BB-DPD scenario discussed above. In addition, desired intra-band and inter-band distortion products can be computed explicitly by separate RF-DPD actuators for each band of the multi-band signal.

[0042] In this regard, Figure 5 illustrates a transmitter 72 that implements RF-DPD according to one embodiment of the present disclosure. In this embodiment, the transmitter 72 is a multi-carrier, dual-band system (e.g., a transmitter for a CDMA or Long Term Evolution (LTE) base station). The two bands of the dual-band system are referred to as a lower band (BAND L) and an upper band (BAND U). As illustrated, complex-valued digital baseband input signals for the two bands are generated by corresponding sources 74-1 and 74-2. Bandwidth allocations in each band may be identical or may be different (i.e., a symmetric versus an asymmetric case). The digital baseband input signals are digitally upconverted by a digital upconversion system 76 to thereby provide complex-valued digital RF signals L(n) and U(n). The digital RF signals L(n) and U(n) are predistorted, at digital RF, by a RF-DPD system 78 to thereby provide corresponding complex-valued predistorted digital RF signals L(n) and U(n). The predistorted digital RF signals L(n) and U(n) are summed by a combiner 80, and a resulting predistorted multi-band digital RF signal is processed by a multi-band signal processing system 82 to provide a predistorted multi-band analog RF signal, which is amplified by a PA 84 to provide a multi-band transmit signal z(n).

[0043] More specifically, the digital upconversion system 76 includes a DRF upconversion subsystem 86-1 for the lower band (referred to as DRF-L) and a DRF upconversion subsystem 86-2 for the upper band (referred to as DRF-U). The DRF upconversion subsystem 86-1 for the lower band includes a rate change filter 88-1, a complex baseband (BB) tuner 90-1, an upsampler 92-1, and an image selection filter 94-1. In this embodiment, the rate change filter 88-1 converts a sampling rate of the digital baseband input signal from the source 74-1 to fs/N prior to complex baseband tuning, where fs is a sampling rate of a DAC 100 of the multi-band signal processing system 82 (discussed in greater detail below) and N is the up-sampling rate of the upsampler 92-1. Alternatively, a sampling rate of the digital baseband input signal from the source 74-1 for the lower band is fS/N in which case the rate change filter 88-1 is not needed. Note that if rate conversion is needed, the rate conversion may alternatively be performed elsewhere in the DRF upconversion subsystem 86-1 prior to the DAC 100 (e.g., subsequent to the complex baseband tuner 90-1, subsequent to the upsampler 92-1, subsequent to the image selection filter 94-1, or prior to the DAC 100).

[0044] The complex baseband tuner 90-1 performs a complex tuning of the complex digital baseband input signal for the lower band. More specifically, the complex baseband tuner 90-1 tunes the digital baseband input signal for the lower band to a baseband tuning frequency (fL) to thereby produce a complex tuned digital signal. In one embodiment, the baseband tuning frequency (fL) is programmable or otherwise selectable within a range of -fS/2N and fS/2N, where fS is an effective sampling rate of the DAC 100 and N is an up-sampling rate of the upsampler 92-1. The complex baseband tuner 90-1 is preferably utilized to provide fine tuning to achieve the desired carrier frequency for the corresponding frequency band of the multi-band transmit signal z(n) output by the PA 84. A particular example of a complex baseband tuner is described in U.S. Patent Application Publication No. 2009/0316838 A1 entitled CORDIC BASED COMPLEX TUNER WITH EXACT FREQEUNCY RESOLUTION, filed on June 23, 2008 and published on December 24, 2009.

[0045] The upsampler 92-1 upsamples the complex tuned digital signal output by the complex baseband tuner 90-1 at an up-sampling rate N, where N ≥ 2, to produce an upsampled digital signal having a sampling rate of fS. In the frequency domain, the upsampled digital signal includes N images of the complex tuned digital signal output by the complex baseband tuner 90-1 equally spaced apart in the frequency range of 0 to fS. The upsampler 92-1 is preferably utilized to provide coarse tuning to achieve the desired carrier frequency for the lower band of the multi-band signal. The image selection filter 94-1 filters the upsampled digital signal to select a desired one of the images of the complex tuned digital signal and thereby provide the digital RF signal for the lower band.
More specifically, the image selection filter 94-1 is preferably programmable via one or more parameters (e.g., filter coefficients) such that a passband of the image selection filter 94-1 is centered at a desired filter tuning frequency. The filter tuning frequency is selected such that the desired image of the complex tuned digital signal falls within the passband of the image selection filter 94-1. In one embodiment, the center frequency of the digital RF signal is equal to the desired center frequency for the low band in the multi-band transmit signal z(n). However, the center frequency of the digital RF signal is not limited thereto. For instance, the center frequency of the digital RF signal may be such that the desired center frequency of the low band in the multi-band transmit signal z(n) falls in the second or higher Nyquist zone of the DAC 100.

[0046] In the same manner, the DRF upconversion subsystem 86-2 includes a rate change filter 88-2, a complex baseband tuner 90-2, an upsampler 92-2, and an image selection filter 94-2. In this embodiment, the rate change filter 88-2 converts a sampling rate of the digital baseband input signal from the source 74-2 to fS/N prior to complex baseband tuning, where fS is a sampling rate of the DAC 100 and N is the up-sampling rate of the upsampler 92-2. Alternatively, a sampling rate of the digital baseband input signal from the source 74-2 for the higher band is fS/N in which case the rate change filter 88-2 is not needed. Note that if rate conversion is needed, the rate conversion may alternatively be performed elsewhere in the DRF upconversion subsystem 86-2 prior to the DAC 100 (e.g., subsequent to the complex baseband tuner 90-2, subsequent to the upsampler 92-2, subsequent to the image selection filter 94-2, or prior to the DAC 100).

[0047] The complex baseband tuner 90-2 performs a complex tuning of the digital baseband input signal for the upper band. More specifically, the complex baseband tuner 90-2 tunes the digital baseband input signal for the upper band to a baseband tuning frequency (fU) to thereby produce a complex tuned digital signal. In one embodiment, the baseband tuning frequency (fU) is programmable or otherwise selectable within a range of -fS/2N and fS/2N, where fS is an effective sampling rate of the DAC 100 and N is an up-sampling rate of the upsampler 92-2. The complex baseband tuner 90-2 is preferably utilized to provide fine tuning to achieve the desired carrier frequency for the corresponding frequency band of the multi-band transmit signal z(n) output by the PA 84.

[0048] Again, a particular example of a complex baseband tuner is described in U.S. Patent Application Publication No. 2009/0316838 A1.

[0049] The upsampler 92-2 upsamples the complex tuned digital signal output by the complex baseband tuner 90-2 at an up-sampling rate N, where N ≥ 2, to produce an upsampled digital signal having a sampling rate of fS. In the frequency domain, the upsampled digital signal includes N images of the complex tuned digital signal output by the complex baseband tuner 90-2 equally spaced apart in the frequency range of 0 to fS. The upsampler 92-2 is preferably utilized to provide coarse tuning to achieve the desired carrier frequency for the upper band of the multi-band signal. The image selection filter 94-2 filters the upsampled digital signal to select a desired one of the images of the complex tuned digital signal and thereby provide the digital RF signal for the upper band. More specifically, the image selection filter 94-2 is preferably programmable via one or more parameters (e.g., filter coefficients) such that a passband of the image selection filter 94-2 is centered at a desired filter tuning frequency. The filter tuning frequency is selected such that the desired image of the complex tuned digital signal falls within the passband of the image selection filter 94-2. In one embodiment, the center frequency of the digital RF signal is equal to the desired center frequency for the upper band in the multi-band transmit signal z(n). However, the center frequency of the digital RF signal is not limited thereto. For instance, the center frequency of the digital RF signal may be such that the desired center frequency of the upper band in the multi-band transmit signal z(n) falls in the second or higher Nyquist zone of the DAC 100.

[0050] The RF-DPD system 78 includes separate RF-DPD actuators 96-1 and 96-2 for the lower and upper bands, respectively. Using complex-valued weights {aqm} for the lower band (not shown), the RF-DPD actuator 96-1 predistorts the digital RF signal for the lower band by directly synthesizing self
and intra-band distortion products based on the digital RF signal for the lower band and inter-band distortion products (or distortion cross-products) based on both the digital RF signal for the lower band output by the DRF upconversion subsystem 86-1 and the digital RF signal for the upper band output by the DRF upconversion subsystem 86-2. In one preferred embodiment, the self, intra-band, and inter-band distortion products generate the predistorted digital RF signal for the lower band in such a manner as to introduce predistortion that is substantially an inverse of distortion caused by non-linearity of the PA 84 with respect to the lower band.

[0051] Equation 4 below is a mathematical representation of the operation of the RF-DPD actuator 96-1 according to one exemplary embodiment of the present disclosure. In general, the RF-DPD actuator 96-1 generates distortion products as desired on system requirements and a frequency plan for the transmitter 72 as well as tradeoffs between which PA distortions are to be mitigated by RF-DPD versus analog filtering.

where:
  • The {a1,m} taps produce 1st order terms centered at the center frequency of the lower band (fL) to equalize the non-linear PA 84.
  • The [a3,m} taps produce the 3rd order self and intra-band distortion products centered at the center frequency of the lower band (fL).
  • The {a5,m} taps produce the 5th order self and intra-band distortion products centered at the center frequency of the lower band (fL).
  • The {b3,m} taps produce the 3rd order inter-band distortion products centered at the center frequency of the lower band (fL).
  • The {b5,m} and {c5,m} taps produce the 5th order inter-band distortion products centered at the center frequency of the lower band (fL).
  • The {d3,m} taps produce the 3rd order inter-band distortion products centered at the lower intermodulation frequency between the two bands (i.e., 2fL - fU).
  • The {d5,m} and {e5,m} taps produce the 5th order inter-band distortion products centered at the lower intermodulation frequency between the two bands (i.e., 2fL - fU).
The distortion products shown in Equation 4 represent only a particular example. Additional inter-band products centered at alternative intermodulation frequencies such as, for example, 3fL - 2fU or others can easily be generated and utilized by the RF-DPD actuator 96-1 using the same approach.

[0052] In a similar manner, using complex-valued tap weights {aqm} for the upper band (not shown), the RF-DPD actuator 96-2 predistorts the digital RF signal for the upper band by directly synthesizing self and intra-band distortion products based on the digital RF signal for the upper band and inter-band distortion products (or distortion cross-products) based on both the digital RF signal for the upper band output by the DRF upconversion subsystem 86-2 and the digital RF signal for the lower band output by the DRF upconversion subsystem 86-1. In one preferred embodiment, the self, intra-band, and inter-band distortion products generate the predistorted digital RF signal for the upper band in such a manner as to introduce predistortion that is substantially an inverse of distortion caused by non-linearity of the PA 84 with respect to the upper band. In one embodiment, the RF-DPD actuator 96-2 for the upper band generates distortion products of the same form as those shown in Equation 4 but with the role of the "L" and "U" signals interchanged (i.e., replace LU
throughout in Equation 4 and assume an independent set of complex-valued tap weights are used).

[0053] The predistorted digital RF signals for the lower and upper bands output by the RF-DPD actuators 96-1 and 96-2, respectively, are summed by the combiner 80 to provide the predistorted multi-band digital RF signal. In the illustrated embodiment, a real-valued predistorted multi-band digital RF signal is obtained by a REAL{•} function 98. The real-valued predistorted multi-band digital RF signal is then D/A converted by the DAC 100 and optionally filtered by an RF filter 102. The resulting predistorted multi-band analog signal is provided to the PA 84 for amplification.

[0054] Before proceeding, it should be noted that the REAL{•} function 98 may alternatively be implemented within the RF-DPD system 78 prior to the combiner 80 and the DAC 100. Still further, the need for the including the REAL{•} function 98 may be removed by computing only the real parts of the predistorted digital RF signals in the RF-DPD actuators 96-1 and 96-2 (rather than computing both the real and imaginary parts of the predistorted digital RF signals). Doing so reduces the computational complexity of the RF-DPD system 78 by a factor of 2 and removes the need for the REAL{•} function 98.

[0055] The RF-DPD system 78, and in particular the tap weights {aqm} of the RF-DPD actuators 96-1 and 96-2, are trained by a training system 104. In general, the training system 104 trains the RF-DPD system 78, and in particular the tap weights {aqm} of the DPD actuators 96-1 and 96-2, using a closed-loop training algorithm that employs a transmit, or feedback, observation receiver. In Figure 5, the training system 104 is based on an "indirect learning" principle where the (delayed) target signals are taken from the RF-DPD actuators 96-1 and 96-2. Alternatively, a "direct" solution may be used where the target signals are taken from the outputs of the DRF upconversion subsystems 86-1 and 86-2. In either case, the tap weights {aqm} may be computed using any suitable form of a least squares algorithm or stochastic gradient descent algorithm. For instance, the training system 104 may utilize a least squares algorithm since the predistorted digital RF signals (e.g., those generated according to Equation 4) are linear functions of the corresponding tap weights {aqm}.

[0056] More specifically, in the example illustrated in Figure 5, an Analog-to-Digital Converter (ADC) 106 is coupled to the output of the PA 84 via a coupler 108. The ADC 106 outputs a multi-band digital RF signal that corresponds to the multi-band transmit signal output by the PA 84. The training system 104 includes bandpass filters 110-1 and 110-2 centered at the lower band and the upper band, respectively. For the lower band, the multi-band digital RF signal is filtered by the bandpass filter 110-1 and the resulting filtered digital RF signal for the lower band is translated to baseband by a complex tuner 112-1 and down-sampled by a factor M by a decimation filter 114-1 to provide a digital baseband feedback signal for the lower band. Likewise, for the upper band, the multiband digital RF signal is filtered by the bandpass filter 110-2 and the resulting filtered digital RF signal for the upper band is translated to baseband by a complex tuner 112-2 and then down-sampled by a factor M by a decimation filter 114-2 to provide a digital baseband feedback signal for the upper band. The digital baseband feedback signals for the lower and upper bands are provided to baseband training subsystems 116-1 and 116-2, as illustrated. Thus, there are separate baseband training subsystems 116-1 and 116-2 for the different frequency bands. Note, however, that each baseband training subsystem 116-1 and 116-2 utilizes both of the digital baseband feedback signal for the lower band and the digital baseband feedback signal for the upper band.

[0057] In a similar manner, the predistorted digital RF signal output by the RF-DPD actuator 96-1 is translated, or tuned, to baseband by a complex tuner 118-1, down-sampled by a factor M by a decimation filter 120-1, and delayed by a delay 122-1 to provide a delayed baseband equivalent signal for the predistorted digital RF signal output by the RF-DPD actuator 96-1 for the lower band. Likewise, the predistorted digital RF signal output by the RF-DPD actuator 96-2 is translated, or tuned, to baseband by a complex tuner 118-2, down-sampled by a factor M by a decimation filter 120-2, and delayed by a delay 122-2 to provide a delayed baseband equivalent signal for the predistorted digital RF signal output by the RF-DPD actuator 96-2 for the upper band.

[0058] The baseband training subsystem 116-1 directly synthesizes a baseband equivalent version of the same self, intra-band, and inter-band distortion products generated by the RF-DPD actuator 96-1 based on the digital baseband feedback signal for the lower band and the digital baseband feedback signal for the upper band to thereby output a baseband feedback signal for the lower band. Ideally, the baseband feedback signal for the lower band output of the baseband training subsystem 116-1 is equal to the delayed baseband equivalent signal for the predistorted digital RF signal output by the RF-DPD actuator 96-1 for the lower band. Likewise, the baseband training subsystem 116-2 directly synthesizes a baseband equivalent version of the same self, intra-band, and inter-band distortion products generated by the RF-DPD actuator 96-2 based on the digital baseband feedback signal for the upper band and the digital baseband feedback signal for the upper band to thereby output a baseband feedback signal for the upper band. Ideally, the baseband feedback signal for the upper band output of the baseband training subsystem 116-2 is equal to the delayed baseband equivalent signal for the predistorted digital RF signal output by the RF-DPD actuator 96-2 for the lower band.

[0059] In operation, a subtractor 124-1 outputs an error signal for the lower band that is equal to a difference between the delayed baseband equivalent signal for the predistorted digital RF signal for the lower band and the baseband feedback signal for the lower band. Based on the error signal from the subtractor 124-1, the baseband training subsystem 116-1 utilizes a suitable training algorithm (e.g., least squares) to adjust tap weights utilized by the baseband training subsystem 116-1 to generate the desired distortion products to minimize the error signal. At some desired point, the tap weights of the baseband training subsystem 116-1 that have been adaptively configured based on the error signal are provided to the RF-DPD actuator 96-1 and thereafter used as the tap weights
{aqm} of the RF-DPD actuator 96-1. This process is repeated as desired to adaptively configure the RF-DPD actuator 96-1.

[0060] In the same manner, a subtractor 124-2 outputs an error signal for the upper band that is equal to a difference between the delayed baseband equivalent signal for the predistorted digital RF signal for the upper band and the baseband feedback signal for the upper band. Based on the error signal from the subtractor 124-2, the baseband training subsystem 116-2 utilizes a suitable training algorithm (e.g., least squares) to adjust tap weights utilized by the baseband training subsystem 116-2 to generate the desired distortion products to minimize the error signal. At some desired point, the tap weights of the baseband training subsystem 116-2 that have been adaptively configured based on the error signal are provided to the RF-DPD actuator 96-2 and thereafter used as the tap weights {aqm} of the RF-DPD actuator 96-2. This process is repeated as desired to adaptively configure the RF-DPD actuator 96-2.

[0061] Note that the training procedure performed by the training system 104 computes the two sets of complex-valued tap weights {aqm} for the RF-DPD actuators 96-1 and 96-2, respectively, independently from one common dual-band input signal obtained from a TOR. The fact that the training procedure for each band may be performed independently provides the following computation advantage. Typically the computational complexity of the least squares procedure is O(N3) for N total tap weights. With B separate bands and assuming NB tap weights per band, it follows the training procedure has a complexity of

whereas a general least squares procedure would be O((B - NB)3).

[0062] Figure 6 illustrates an embodiment of the transmitter 72 of Figure 5 that has been generalized for any multi-band signal having any number (M) of bands (M≥2) whereas the embodiment of Figure 5 is for a dual-band scenario. Thus, the digital upconversion system 76 includes separate DRF upconversion subsystems 86-1 through 86-M for the M bands, the RF-DPD system 78 includes separate RF-DPD actuators 96-1 through 96-M for the M bands, and the training system 104 includes separate bandpass filters 110-1 through 110-M, complex tuners 112-1 through 112-M, decimation filters 114-1 through 114-M, baseband training subsystems 116-1 through 116-M, complex tuners 118-1 through 118-M, decimation filters 120-1 through 120-M, delays 122-1 through 122-M, and subtractors 124-1 through 124-M for the M bands. Otherwise, the details of the transmitter 72 are the same as described above. As such, the details are not repeated.

[0063] Figure 7 is a flow chart that illustrates the operation of the dual-band embodiment of the transmitter 72 of Figure 5 and, more generally, the operation of the multi-band embodiment of the transmitter 72 of Figure 6 to generate and transmit the multi-band transmit signal according to one embodiment of the present disclosure. First, the digital upconversion system 76, and more specifically the DRF upconversion subsystems 86-1 through 86-M, digitally upconverts the digital baseband input signals for corresponding bands to thereby provide the digital RF signals (step 1000). Next, the RF-DPD system 78, and in particular the RF-DPD actuators 96-1 through 96-M, digitally predistorts the digital RF signals to provide the corresponding predistorted digital RF signals (step 1002). The combiner 80 then combines, or sums, the predistorted digital RF signals to provide the predistorted multi-band digital RF signal (step 1004). The multi-band signal processing system 82 then processes the predistorted multi-band digital RF signal to provide the corresponding predistorted multi-band analog RF signal (step 1006). Lastly, the PA 84 amplifies the predistorted multi-band analog RF signal to provide the multi-band transmit signal (step 1008).

[0064] Figure 8 is a flow chart that illustrates the operation of transmitter 72 of Figures 5 and 6 to train the RF-DPD system 78 according to one embodiment of the present disclosure. In this embodiment, the ADC 106 receives the multi-band transmit signal from the output of the PA 84 via the coupler 108 (step 2000). The ADC 106, the bandpass filters 110-1 through 110-M, the complex tuners 112-1 through 112-M, and the decimation filters 114-1 through 114-M then process the multi-band transmit signal to provide a separate single-band digital baseband feedback signal for each band of the multi-band transmit signal (step 2002). Lastly, the training subsystems 112-1 through 112-M train the RF-DPD system 78, and in particular the tap weights {aqm} of the corresponding RF-DPD actuators 96-1 through 96-M, based on the digital baseband feedback signals for the different bands, as discussed above (step 2004).

[0065] While the systems and methods for implementing RF-DPD discussed herein are not limited to any particular advantages, some exemplary advantages are provided below. Note, however, that the concepts disclosed and claimed herein are not limited to any of these advantages. First, for multi-band signals with several disjoint bands, moving the DPD function from baseband to RF reduces the sampling rate requirements of the DRF upconversion data paths (i.e., the data paths through the DRF upconversion subsystems 86-1 through 86-M) because the DRF upconversion data paths operate on undistorted digital baseband signals, which have a 3x, 5x, or more reduction in bandwidth as compared to the predistorted digital baseband signals when BB-DPD is used. This reduction in sampling rate leads to a significant reduction in computational complexity.

[0066] Second, for multi-band signals with several disjoint bands or large frequency separation between those bands, moving the DPD function from baseband to RF results in fewer DRF upconversion data paths (i.e., fewer DRF upconversion subsystems 86) since low-bandwidth DRF data paths may be allocated to (undistorted) signal bands rather than to (distorted) products of signals originating from multiple bands. The distinct number of distortion products between bands grows rapidly and quickly exceeds the number of bands, as that number of bands increases. In other words, for RF-DPD, no additional DRF upconversion data paths are needed to upconvert inter-band distortion products, as is needed with BB-DPD.

[0067] Third, the RF-DPD actuators 96-1 through 96-M separately and explicitly synthesize the desired intra-band and inter-band distortion products. One advantage to maintaining separate RF-DPD actuators 96-1 through 96-M is that each complex-valued tap weight may incorporate an independent rotation to cancel an unknown carrier phase inherent in the corresponding distortion product at RF. The tap weight adaptation performed by the corresponding training subsystems 112-1 through 112-M solves for these rotations implicitly.

[0068] Fourth, as compared to BB-DPD, RF-DPD provides greater flexibility in selecting specific distortion products for digital predistortion versus using analog filtering since the number of DRF data paths is generally independent of the suite of selected distortion products. In essence, the RF-DPD architecture provides better scalability for multi-band signals. Lastly, a least squares training procedure for RF-DPD has less complexity for multi-band signals than for an equivalent solution implemented at baseband since the tap weights for each band may be optimized independently. Finally, it should be noted that although the complexity of the RF-DPD actuators 96-1 through 96-M is greater than corresponding BB-DPD actuators (due to the higher sampling rate of the former), the savings achieved in the digital upconversion system 76 through fewer branches and lower sampling rates in each branch makes RF-DPD preferable, especially for multi-band signals with several disjoint bands or with large separations between bands.

[0069] The following acronyms are used throughout this disclosure.
• ADC Analog-to-Digital Converter
• BB Baseband
• BB-DPD Baseband Digital Predistortion
• CDMA Code Division Multiple Access
• D/A Digital-to-Analog
• DAC Digital-to-Analog Converter
• DRF Digital Radio Frequency
• DSP Digital Signal Processing
• IF Intermediate Frequency
• LTE Long Term Evolution
• MHz Megahertz
• MP Memory Polynomial
• OFDM Orthogonal Frequency Duplex Multiplexing
• PA Power Amplifier
• RF Radio Frequency
• RF-APD Radio Frequency Analog Predistortion
• RF-DPD Radio Frequency Digital Predistortion
• TOR Transmit Observation Receiver


[0070] Those skilled in the art will recognize improvements and modifications to the preferred embodiments of the present disclosure. All such improvements and modifications are considered within the scope of the claims that follow.


Claims

1. A multi-band transmitter (72), comprising:

a digital upconversion system (76) configured to digitally upconvert a plurality of digital input signals to provide a plurality of digital radio frequency, RF, signals (ũ1(n), ũL(n), ũU(n), ũM(n)), each digital input signal of the plurality of digital input signals and thus each digital RF signal of the plurality of digital RF signals corresponding to a different band of a plurality of bands of a multi-band transmit signal (z(n)) to be transmitted by the multi-band transmitter (72);

a RF digital predistortion system (78) configured to:

- digitally predistort the plurality of digital RF signals to provide a plurality of predistorted digital RF signals (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)), and

- directly synthesize intra-band and inter-band distortion products based on the plurality of digital radio frequency signals providing a corresponding one of the plurality of predistorted digital RF signals and combiner circuitry (80) configured to combine the plurality of predistorted digital RF signals to provide a multi-band predistorted digital RF signal.


 
2. The multi-band transmitter (72) of claim 1 further comprising:

a multi-band signal processing system (82) configured to process the multi-band predistorted digital RF signal to provide a multi-band predistorted analog RF signal; and

power amplifier circuitry (84) configured to amplify the multi-band predistorted analog RF signal to provide the multi-band transmit signal.


 
3. The multi-band transmitter (72) of claim 2 wherein the multi-band predistorted digital RF signal is a complex signal, and the multi-band signal processing system (82) comprises:

circuitry (98) configured to receive the multi-band predistorted digital RF signal and output a real part of the multi-band predistorted digital RF signal;

digital-to-analog conversion circuitry (100) configured to convert the real part of the multi-band predistorted digital RF signal from digital to analog and output a corresponding analog signal; and

filtering circuitry (102) configured to filter the analog signal output by the digital-to-analog conversion circuitry (100) to thereby provide the multi-band predistorted analog RF signal; or

wherein the multi-band predistorted digital RF signal is a real signal, and the multi-band signal processing system (82) comprises:

digital-to-analog conversion circuitry (100) configured to convert the multi-band predistorted digital RF signal from digital to analog and output a corresponding analog signal; and

filtering circuitry (102) configured to filter the analog signal output by the digital-to-analog conversion circuitry (100) to thereby provide the multi-band predistorted analog RF signal.


 
4. The multi-band transmitter (72) of claim 1 wherein the digital upconversion system (76) comprises a separate digital upconversion subsystem (86-1 through 86-M) for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter (72).
 
5. The multi-band transmitter (72) of claim 4 wherein, for each band of the multi-band transmit signal, the separate digital upconversion subsystem (86-1 through 86-M) for the band of the multi-band transmit signal comprises:

complex baseband tuning circuitry (90-1 through 90-M) configured to perform complex tuning of the digital input signal for the band to a desired baseband tuning frequency to thereby provide a complex tuned digital signal;

up-sampling circuitry (92-1 through 92-M) configured to upsample the complex tuned digital signal to a desired up-sampling rate to thereby provide an upsampled digital signal; and

image selection filtering circuitry (94-1 through 94-M) configured to filter the upsampled digital signal to select a desired image of the complex tuned digital signal and thereby provide the digital RF signal for the band of the multi-band transmit signal.


 
6. The multi-band transmitter (72) of claim 5 wherein, for each band of the multi-band transmit signal, the separate digital upconversion subsystem (86-1 through 86-M) for the band of the multi-band transmit signal further comprises a rate change filter (88-1 through 88-M).
 
7. The multi-band transmitter (72) of claim 4 wherein the RF digital predistortion system (78) comprises a separate RF digital predistortion subsystem (96-1, 96-2, 96-M) for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter (72).
 
8. The multi-band transmitter (72) of claim 7 wherein, for each band of the multi-band transmit signal, the separate RF digital predistortion subsystem (96-1, 96-2, 96-M) for the band of the multi-band transmit signal is configured to:
directly synthesize desired self, intra-band, and inter-band distortion products for the band based on the plurality of digital RF signals to thereby provide a corresponding one of the plurality of predistorted digital RF signals.
 
9. The multi-band transmitter (72) of claim 7 further comprising a training system (104) configured to train the RF digital predistortion system (78).
 
10. The multi-band transmitter (72) of claim 9 wherein the training system (104) comprises a separate training subsystem (116-1 through 116-M) for each band of the multi-band transmit signal.
 
11. A method of operation of a multi-band transmitter (72), comprising:

digitally upconverting (1000) a plurality of digital input signals to provide a plurality of digital radio frequency, RF, signals (ũ1(n), ũL(n), ũU(n), ũM(n)), each digital input signal of the plurality of digital input signals and thus each digital RF signal of the plurality of digital RF signals corresponding to a different band of a plurality of bands of a multi-band transmit signal (z(n)) to be transmitted by the multi-band transmitter (72);

digitally predistorting (1002) the plurality of digital RF signals to provide a plurality of predistorted digital RF signals (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)), wherein digitally predistorting (1002) the plurality of digital RF signals comprises directly synthesizing intra-band and inter-band distortion products; and

combining (1004) the plurality of predistorted digital RF signals to provide a multi-band predistorted digital RF signal.


 
12. The method of claim 11 further comprising:

processing (1006) the multi-band predistorted digital RF signal to provide a multi-band predistorted analog RF signal; and

amplifying (1008) the multi-band predistorted analog RF signal to provide the multi-band transmit signal.


 
13. The method of claim 11 wherein digitally upconverting (1000) the plurality of digital input signals comprises separately digitally upconverting the plurality of digital input signals for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter (72).
 
14. The method of claim 13 wherein digitally predistorting the plurality of digital RF signals comprises separately digitally predistorting the plurality of digital RF signals for each band of the multi-band transmit signal to be transmitted by the multi-band transmitter (72).
 
15. The method of claim 14 wherein:

- separately digitally predistorting the plurality of digital RF signals comprises, for each band of the multi-band transmit signal, directly synthesizing desired self, intra-band, and inter-band distortion products for the band based on the plurality of digital RF signals to thereby provide a corresponding one of the plurality of predistorted digital RF signals; or

- the method further comprises training models utilized for digitally predistorting the plurality of digital RF signals.


 


Ansprüche

1. Mehrbandsender (72), umfassend:

ein digitales Aufwärtswandlungssystem (76), das konfiguriert ist, um eine Vielzahl von digitalen Eingangssignalen digital aufwärtszuwandeln, um eine Vielzahl von digitalen Hochfrequenzsignalen, HF-Signalen, (ũ1(n), ũL(n), ũU(n), ũM(n)) bereitzustellen, wobei jedes digitale Eingangssignal von der Vielzahl von digitalen Eingangssignalen und somit jedes digitale HF-Signal von der Vielzahl von digitalen HF-Signalen einem anderen Band von einer Vielzahl von Bändern eines durch den Mehrbandsender (72) zu übertragenden Mehrbandsendesignals (z(n)) entspricht;

ein digitales HF-Vorverzerrungssystem (78), das konfiguriert ist zum:

- digitalen Vorverzerren der Vielzahl von digitalen HF-Signalen, um eine Vielzahl von vorverzerrten digitalen HF-Signalen (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)) bereitzustellen, und

- direkten Synthetisieren von Intraband- und Interband-Verzerrungsprodukten basierend auf der Vielzahl von digitalen Hochfrequenzsignalen, die ein entsprechendes von der Vielzahl von vorverzerrten digitalen HF-Signalen bereitstellen und Kombiniererschaltungen (80), die konfiguriert sind, um die Vielzahl von vorverzerrten digitalen HF-Signalen zu kombinieren, um ein mehrbandiges vorverzerrtes digitales HF-Signal bereitzustellen.


 
2. Mehrbandsender (72) nach Anspruch 1, ferner umfassend:

ein Mehrbandsignalverarbeitungssystem (82), das konfiguriert ist, um das mehrbandige vorverzerrte digitale HF-Signal zu verarbeiten, um ein mehrbandiges vorverzerrtes analoges HF-Signal bereitzustellen; und

Leistungsverstärkerschaltungen (84), die konfiguriert sind, um das mehrbandige vorverzerrte analoge HF-Signal zu verstärken, um das mehrbandige Sendesignal bereitzustellen.


 
3. Mehrbandsender (72) nach Anspruch 2, wobei das mehrbandige vorverzerrte digitale HF-Signal ein komplexes Signal ist und das Mehrbandsignalverarbeitungssystem (82) Folgendes umfasst:

Schaltungen (98), die konfiguriert sind, um das mehrbandige vorverzerrte digitale HF-Signal zu empfangen und einen reellen Teil des mehrbandigen vorverzerrten digitalen HF-Signals auszugeben;

Digital-Analog-Umwandlungsschaltungen (100), die konfiguriert sind, um den reellen Teil des mehrbandigen vorverzerrten digitalen HF-Signals von digital in analog umzuwandeln und ein entsprechendes analoges Signal auszugeben; und

Filterschaltungen (102), die konfiguriert sind, um das von den Digital-Analog-Umwandlungsschaltungen (100) ausgegebene analoge Signal zu filtern, um dadurch das mehrbandige vorverzerrte analoge HF-Signal bereitzustellen; oder

wobei das mehrbandige vorverzerrte digitale HF-Signal ein reelles Signal ist und das Mehrbandsignalverarbeitungssystem (82) Folgendes umfasst:

Digital-Analog-Umwandlungsschaltungen (100), die konfiguriert sind, um das mehrbandige vorverzerrte digitale HF-Signal von digital in analog umzuwandeln und ein entsprechendes analoges Signal auszugeben; und

Filterschaltungen (102), die konfiguriert sind, um das von den Digital-Analog-Umwandlungsschaltungen (100) ausgegebene analoge Signal zu filtern, um dadurch das mehrbandige vorverzerrte analoge HF-Signal bereitzustellen.


 
4. Mehrbandsender (72) nach Anspruch 1, wobei das digitale Aufwärtswandlungssystem (76) für jedes Band des Mehrbandsendesignals, das durch den Mehrbandsender (72) zu übertragen ist, ein separates digitales Aufwärtswandlungssubsystem (86-1 bis 86-M) umfasst.
 
5. Mehrbandsender (72) nach Anspruch 4, wobei, für jedes Band des Mehrbandsendesignals, das separate digitale Aufwärtswandlungssubsystem (86-1 bis 86-M) für das Band des Mehrbandsendesignals Folgendes umfasst:

komplexe Basisbandabstimmschaltungen (90-1 bis 90-M), die konfiguriert sind, um eine komplexe Abstimmung des digitalen Eingangssignals für das Band auf eine gewünschte Basisbandabstimmfrequenz durchzuführen, um dadurch ein komplexes abgestimmtes digitales Signal bereitzustellen;

Abtastratenerhöhungsschaltungen (92-1 bis 92-M) die konfiguriert sind, um eine Erhöhung der Abtastrate des komplexen abgestimmten digitalen Signals auf eine gewünschte erhöhte Abtastrate vorzunehmen, um dadurch ein digitales Signal mit erhöhter Abtastrate bereitzustellen; und

Bildauswahlfilterschaltungen (94-1 bis 94-M), die konfiguriert sind, um das digitale Signal mit erhöhter Abtastrate zu filtern, um ein gewünschtes Bild des komplexen abgestimmten digitalen Signals auszuwählen und dadurch das digitale HF-Signal für das Band des Mehrbandsendesignals bereitzustellen.


 
6. Mehrbandsender (72) nach Anspruch 5, wobei, für jedes Band des Mehrbandsendesignals, das separate digitale Aufwärtswandlungssubsystem (86-1 bis 86-M) für das Band des Mehrbandsendesignals ferner einen Ratenänderungsfilter (88-1 bis 88-M) umfasst.
 
7. Mehrbandsender (72) nach Anspruch 4, wobei das digitale HF-Vorverzerrungssystem (78) für jedes Band des Mehrbandsendesignals, das durch den Mehrbandsender (72) zu übertragen ist, ein separates digitales HF-Vorverzerrungssubsystem (96-1, 96-2, 96-M) umfasst.
 
8. Mehrbandsender (72) nach Anspruch 7, wobei, für jedes Band des Mehrbandsendesignals, das separate digitale HF-Vorverzerrungssubsystem (96-1, 96-2, 96-M) für das Band des Mehrbandsendesignals konfiguriert ist zum:
direkten Synthetisieren von gewünschten Eigen-, Intraband- und Interband-Verzerrungsprodukten für das Band basierend auf der Vielzahl von digitalen HF-Signalen, um dadurch ein entsprechendes von der Vielzahl von vorverzerrten digitalen HF-Signalen bereitzustellen.
 
9. Mehrbandsender (72) nach Anspruch 7, ferner umfassend ein Trainingssystem (104), das konfiguriert ist, um das digitale HF-Vorverzerrungssystem (78) zu trainieren.
 
10. Mehrbandsender (72) nach Anspruch 9, wobei das Trainingssystem (104) ein separates Trainingssubsystem (116-1 bis 116-M) für jedes Band des Mehrbandsendesignals umfasst.
 
11. Verfahren zum Betrieb eines Mehrbandsenders (72), umfassend:

digitale Aufwärtswandlung (1000) einer Vielzahl von digitalen Eingangssignalen, um eine Vielzahl von digitalen Hochfrequenzsignalen, HF-Signalen, (ũ1(n), ũL(n), ũU(n), ũM(n)) bereitzustellen, wobei jedes digitale Eingangssignal von der Vielzahl von digitalen Eingangssignalen und somit jedes digitale HF-Signal von der Vielzahl von digitalen HF-Signalen einem anderen Band von einer Vielzahl von Bändern eines durch den Mehrbandsender (72) zu übertragenden Mehrbandsendesignals (z(n)) entspricht;

digitales Vorverzerren (1002) der Vielzahl von digitalen HF-Signalen, um eine Vielzahl von vorverzerrten digitalen HF-Signalen (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)) bereitzustellen, wobei das digitale Vorverzerren (1002) der Vielzahl von digitalen HF-Signalen das direkte Synthetisieren von Intraband- und Interband-Verzerrungsprodukten umfasst; und

Kombinieren (1004) der Vielzahl von vorverzerrten digitalen HF-Signalen, um ein mehrbandiges vorverzerrtes digitales HF-Signal bereitzustellen.


 
12. Verfahren nach Anspruch 11, ferner umfassend:

Verarbeiten (1006) des mehrbandigen vorverzerrten digitalen HF-Signals, um ein mehrbandiges vorverzerrtes analoges HF-Signal bereitzustellen; und

Verstärken (1008) des mehrbandigen vorverzerrten analogen HF-Signals, um das mehrbandige Sendesignal bereitzustellen.


 
13. Verfahren nach Anspruch 11, wobei das digitale Aufwärtswandeln (1000) der Vielzahl von digitalen Eingangssignalen das separate digitale Aufwärtswandeln der Vielzahl von digitalen Eingangssignalen für jedes Band des Mehrbandsendesignals umfasst, das durch den Mehrbandsender (72) zu übertragen ist.
 
14. Verfahren nach Anspruch 13, wobei das digitale Vorverzerren der Vielzahl von digitalen HF-Signalen das separate digitale Vorverzerren der Vielzahl von digitalen HF-Signalen für jedes Band des Mehrbandsendesignals umfasst, das durch den Mehrbandsender (72) zu übertragen ist.
 
15. Verfahren nach Anspruch 14, wobei:

- das separate digitale Vorverzerren der Vielzahl von digitalen HF-Signalen für jedes Band des Mehrbandsendesignals das direkte Synthetisieren gewünschter Eigen-, Intraband- und Interband-Verzerrungsprodukte für das Band basierend auf der Vielzahl von digitalen HF-Signalen umfasst, um dadurch ein entsprechendes von der Vielzahl von vorverzerrten digitalen HF-Signalen bereitzustellen; oder

- das Verfahren ferner Trainingsmodelle umfasst, die zum digitalen Vorverzerren der Vielzahl von digitalen HF-Signalen verwendet werden.


 


Revendications

1. Émetteur multibande (72), comprenant :

un système de conversion ascendante numérique (76) configuré pour convertir numériquement de manière ascendante une pluralité de signaux d'entrée numériques pour fournir une pluralité de signaux de radiofréquence, RF, numériques (ũ1(n), ũL(n), ũU(n), ũM(n)), chaque signal d'entrée numérique de la pluralité de signaux d'entrée numériques et donc chaque signal RF numérique de la pluralité de signaux RF numériques correspondant à une bande différente d'une pluralité de bandes d'un signal de transmission multibande (z(n)) à transmettre par l'émetteur multibande (72) ;

un système de prédistorsion numérique RF (78) configuré pour :

- appliquer une prédistorsion numérique à la pluralité de signaux RF numériques pour fournir une pluralité de signaux RF numériques prédistordus (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)), et

- synthétiser directement des produits de distorsion intrabande et interbande sur la base de la pluralité de signaux radiofréquence numériques fournissant l'un correspondant parmi la pluralité de signaux RF numériques prédistordus et

un circuit combinateur (80) configuré pour combiner la pluralité de signaux RF numériques prédistordus pour fournir un signal RF numérique prédistordu multibande.


 
2. Émetteur multibande (72) selon la revendication 1 comprenant en outre :

un système de traitement de signal multibande (82) configuré pour traiter le signal RF numérique prédistordu multibande pour fournir un signal RF analogique prédistordu multibande ; et

un circuit amplificateur de puissance (84) configuré pour amplifier le signal RF analogique prédistordu multibande pour fournir le signal de transmission multibande.


 
3. Émetteur multibande (72) selon la revendication 2 dans lequel le signal RF numérique prédistordu multibande est un signal complexe, et le système de traitement de signal multibande (82) comprend :

un circuit (98) configuré pour recevoir le signal RF numérique prédistordu multibande et délivrer en sortie une partie réelle du signal RF numérique prédistordu multibande ;

un circuit de conversion numérique vers analogique (100) configuré pour convertir la partie réelle du signal RF numérique prédistordu multibande de numérique vers analogique et délivrer en sortie un signal analogique correspondant ; et

un circuit de filtrage (102) configuré pour filtrer le signal analogique délivré en sortie par le circuit de conversion numérique vers analogique (100) pour fournir de ce fait le signal RF analogique prédistordu multibande ; ou

dans lequel le signal RF numérique prédistordu multibande est un signal réel, et le système de traitement de signal multibande (82) comprend :

un circuit de conversion numérique vers analogique (100) configuré pour convertir le signal RF numérique prédistordu multibande de numérique vers analogique et délivrer en sortie un signal analogique correspondant ; et

un circuit de filtrage (102) configuré pour filtrer le signal analogique délivré en sortie par le circuit de conversion numérique vers analogique (100) pour fournir de ce fait le signal RF analogique prédistordu multibande.


 
4. Émetteur multibande (72) selon la revendication 1 dans lequel le système de conversion ascendante numérique (76) comprend un sous-système indépendant de conversion ascendante numérique (86-1 à 86-M) pour chaque bande du signal de transmission multibande à transmettre par l'émetteur multibande (72).
 
5. Émetteur multibande (72) selon la revendication 4 dans lequel, pour chaque bande du signal de transmission multibande, le sous-système indépendant de conversion ascendante numérique (86-1 à 86-M) pour la bande du signal de transmission multibande comprend :

un circuit de syntonisation de bande de base complexe (90-1 à 90-M) configuré pour mettre en oeuvre une syntonisation complexe du signal d'entrée numérique pour la bande à une fréquence de syntonisation de bande de base souhaitée pour fournir de ce fait un signal numérique syntonisé complexe ;

un circuit de suréchantillonnage (92-1 à 92-M) configuré pour suréchantillonner le signal numérique syntonisé complexe à un taux de suréchantillonnage souhaité pour fournir de ce fait un signal numérique suréchantillonné ; et

un circuit de filtrage de sélection d'image (94-1 à 94-M) configuré pour filtrer le signal numérique suréchantillonné pour sélectionner une image souhaitée du signal numérique syntonisé complexe et fournir de ce fait le signal RF numérique pour la bande du signal de transmission multibande.


 
6. Émetteur multibande (72) selon la revendication 5 dans lequel, pour chaque bande du signal de transmission multibande, le sous-système indépendant de conversion ascendante numérique (86-1 à 86-M) pour la bande du signal de transmission multibande comprend en outre un filtre de changement de taux (88-1 à 88-M).
 
7. Émetteur multibande (72) selon la revendication 4 dans lequel le système de prédistorsion numérique RF (78) comprend un sous-système indépendant de prédistorsion numérique RF (96-1, 96-2, 96- M) pour chaque bande du signal de transmission multibande à transmettre par l'émetteur multibande (72).
 
8. Émetteur multibande (72) selon la revendication 7 dans lequel, pour chaque bande du signal de transmission multibande, le sous-système indépendant de prédistorsion numérique RF (96-1, 96- 2, 96-M) pour la bande du signal de transmission multibande est configuré pour :
synthétiser directement des produits souhaités d'auto-distorsion intrabande et interbande pour la bande sur la base de la pluralité de signaux RF numériques pour fournir de ce fait l'un correspondant parmi la pluralité de signaux RF numériques prédistordus.
 
9. Émetteur multibande (72) selon la revendication 7 comprenant en outre un système d'apprentissage (104) configuré pour entraîner le système de prédistorsion numérique RF (78).
 
10. Émetteur multibande (72) selon la revendication 9 dans lequel le système d'apprentissage (104) comprend un sous-système indépendant d'apprentissage (116-1 à 116-M) pour chaque bande du signal de transmission multibande.
 
11. Procédé de fonctionnement d'un émetteur multibande (72), comprenant :

la conversion numérique ascendante (1000) d'une pluralité de signaux d'entrée numériques pour fournir une pluralité de signaux de radiofréquence, RF, numériques (ũ1(n), ũL(n), ũU(n), ũM(n)), chaque signal d'entrée numérique de la pluralité de signaux d'entrée numériques et donc chaque signal RF numérique de la pluralité de signaux RF numériques correspondant à une bande différente d'une pluralité de bandes d'un signal de transmission multibande (z(n)) à transmettre par l'émetteur multibande (72) ;

la prédistorsion numérique (1002) de la pluralité de signaux RF numériques pour fournir une pluralité de signaux RF numériques prédistordus (ŷ1(n), ŷL(n), ŷU(n), ŷM(n)), dans lequel la prédistorsion numérique (1002) de la pluralité de signaux RF numériques comprend la synthèse directe de produits de distorsion intrabande et interbande ; et

la combinaison (1004) de la pluralité de signaux RF numériques prédistordus pour fournir un signal RF numérique prédistordu multibande.


 
12. Procédé selon la revendication 11, comprenant en outre :

le traitement (1006) du signal RF numérique prédistordu multibande pour fournir un signal RF analogique prédistordu multibande ; et

l'amplification (1008) du signal RF analogique prédistordu multibande pour fournir le signal de transmission multibande.


 
13. Procédé selon la revendication 11 dans lequel la conversion numérique ascendante (1000) de la pluralité de signaux d'entrée numériques comprend la conversion numérique ascendante séparée de la pluralité de signaux d'entrée numériques pour chaque bande du signal de transmission multibande à transmettre par l'émetteur multibande (72).
 
14. Procédé selon la revendication 13 dans lequel la prédistorsion numérique de la pluralité de signaux RF numériques comprend l'application séparée d'une prédistorsion numérique à la pluralité de signaux RF numériques pour chaque bande du signal de transmission multibande à transmettre par l'émetteur multibande (72).
 
15. Procédé selon la revendication 14, dans lequel :

- l'application séparée d'une prédistorsion numérique à la pluralité de signaux RF numériques comprend, pour chaque bande du signal de transmission multibande, la synthèse directe de produits souhaités d'auto-distorsion intrabande et interbande pour la bande sur la base de la pluralité de signaux RF numériques pour fournir de ce fait l'un correspondant parmi la pluralité de signaux RF numériques prédistordus ; ou

- le procédé comprend en outre des modèles d'apprentissage utilisés pour appliquer une prédistorsion numérique à la pluralité de signaux RF numériques.


 




Drawing



























REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description