PREDISTORTION WITH SECTIONED BASIS FUNCTIONS

• US 20090302940A1
• Filed: 03/12/2009
• Published: 12/10/2009
• Est. Priority Date: 06/04/2008
• Status: Active Grant

• Current Assignee: Apple Inc.
• Original Assignee: Nortel Networks Limited (Nortel Networks Corporation)
• Inventors: MORRIS, BRADLEY JOHN, FULLER, ARTHUR THOMAS GERALD
• Granted Patent: US 8,369,447 B2
• Time in Patent Office: Days
• US Class Current: 330/149
• CPC Class Codes: H03F 1/3247 using feedback acting on predistortion circuits H03F1/3264 takes precedence

Abstract

A predistortion actuator is provided. The predistortion actuator includes a plurality of branches, each of which implements a basis function that acts on a digital input signal. For at least one of the branches, the respective basis function is a sectioned basis function, where each section of the sectioned basis function corresponds to a respective section of a range of at least one input signal characteristic associated with the digital input signal, such as a value range of a magnitude, a temporal characteristic, or a hybrid of the two. A power amplifier system including the predistortion actuator is also provided. Utilizing sectioned basis functions can potentially reduce the hardware resources necessary to realize the predistortion actuator relative to conventional global basis functions. In addition, signal conditioning during coefficient training can potentially be used to reduce the dynamic range of coefficients associated with each sectioned basis function.

Claims

1. A predistortion actuator for predistorting a digital input signal, the predistortion actuator comprising:

a plurality of branches that each implement a respective basis function, wherein for each of at least one of the plurality of branches, the respective basis function is a sectioned basis function having a respective distinct definition for each section of a respective plurality of sections, the respective plurality of sections defining a range of at least one input signal characteristic associated with the digital input signal.

2. The predistortion actuator of claim 1, wherein the at least one of the plurality of branches comprises all of the plurality of branches.

3. The predistortion actuator of claim 2, wherein the respective plurality of sections is the same for each of the plurality of branches.

4. The predistortion actuator of claim 1, wherein the at least one input signal characteristic comprises magnitude of the digital input signal, such that each sectioned basis function has a respective distinct definition for each section of a respective plurality of sections defining a magnitude range of the digital input signal.

5. The predistortion actuator of claim 1, wherein the at least one input signal characteristic comprises a temporal characteristic of the digital input signal, such that each sectioned basis function has a respective distinct definition for each section of a respective plurality of sections defining a time period relative to some time marker associated with the digital input signal.

6. The predistortion actuator of claim 1, wherein the at least one input signal characteristic comprises a temporal characteristic of the digital input signal and magnitude of the digital input signal, such that each sectioned basis function has a respective distinct definition for each section of a plurality of sections defining both:

a) a time period relative to some time marker associated with the digital input signal; and

b) a magnitude range of the digital input signal.

7. The predistortion actuator of claim 1, wherein each sectioned basis function comprises a normalization factor that is held constant across each of the respective plurality of sections.

8. The predistortion actuator of claim 1, wherein each sectioned basis function comprises a normalization factor that varies on at least one of:

a per section basis; and

a per branch basis.

9. The predistortion actuator of claim 1, wherein each sectioned basis function is implemented according to:

ƒ

_k(x(n),φ

)=x(n)(normEnv(x(n),φ

)^k(upperSecVal(φ

)−

lowerSecVal(φ

))+lowerSecVal(φ

))sFact(k,φ

),

where

$\mathrm{normEnv}<br><br>\left(x<br><br>\left(n\right),\mathrm{\phi <br><br>}\right)=\frac{<br><br>x<br><br>\left(n\right)<br><br>-\mathrm{lower}<br><br>\mathrm{Sec}<br><br>\mathrm{Val}<br><br>\left(\mathrm{\phi <br><br>}\right)}{\mathrm{upper}<br><br>\mathrm{Sec}<br><br>\mathrm{Val}<br><br>\left(\mathrm{\phi <br><br>}\right)-\mathrm{lower}<br><br>\mathrm{Sec}<br><br>\mathrm{Val}<br><br>\left(\mathrm{\phi <br><br>}\right)},$

where k is a branch number, x(n) is the input signal, φ

is a section number that is determined in accordance with

lowerSecVal(φ

)≦

|x(n)|<

upperSecVal(φ

),

where lowerSecVal and upperSecVal are vectors that define the lower and upper boundaries of each section, respectively, and sFact(k,φ

) is a normalization factor defined for each actuator branch k of the plurality of branches and section φ

of the plurality of sections independently.

10. The predistortion actuator of claim 9, wherein for each sectioned basis function, the normalization factor sFact(k,φ

) is constant across the respective plurality of sections.

11. The predistortion actuator of claim 10, wherein the normalization factor sFact(k,φ

) is determined according to;

$\mathrm{sFact}<br><br>\left(k,\mathrm{\phi <br><br>}\right)=\frac{1}{\max <br><br>\left(\mathrm{upper}<br><br>\mathrm{Sec}<br><br>\mathrm{Val}<br><br>\left(\mathrm{\phi <br><br>}\right)\right)},$

where max(upperSecVal(φ

) is the maximum upper section value of the vector of upper section values upperSecVal(φ

) for all branches k and sections φ

.

12. The predistortion actuator of claim 9, wherein the normalization factor sFact(k,φ

) is determined according to;

$\mathrm{sFact}<br><br>\left(k,\mathrm{\phi <br><br>}\right)=\frac{1}{\mathrm{upper}<br><br>\mathrm{Sec}<br><br>\mathrm{Val}<br><br>\left(\mathrm{\phi <br><br>}\right)}.$

13. The predistortion actuator of claim 1, wherein for at least one of the sectioned basis functions, the section boundaries of the respective plurality of sections are adaptively adjusted.

14. The predistortion actuator of claim 1, wherein each branch of the plurality of branches comprises a Finite Impulse Response (FIR) filter arranged to respectively filter an output of the respective basis function of the respective branch.

15. The predistortion actuator of claim 14, further comprising:

a plurality of adders arranged to collectively add outputs of the FIR filters to generate a predistorted digital signal.

16. The predistortion actuator of claim 14, wherein operation of each FIR filter is based on a respective set of filter coefficients, the predistortion actuator further comprising a distinct set of coefficients for each of the plurality of sections for each of the FIR filters.

17. The predistortion actuator of claim 16, further comprising a section determiner configured to:

for each sample period of the digital input signal, determine a current section of the respective plurality of sections based on a current value of the at least one input signal characteristic of the digital input signal, and select a respective one of the distinct sets of coefficients for the respective FIR filter of each of the plurality of branches based on the current section.

18. The predistortion actuator of claim 1, wherein for each sectioned basis function:

a respective set of coefficients for each section of the plurality of sections is determined using a coefficient training algorithm comprising training signal conditioning.

19. A power amplifier system, comprising:

a predistortion actuator according to claim 1 configured to predistort a baseband digital input signal to generate a predistorted baseband digital signal;

an upconversion function configured to generate a predistorted radio frequency (RF) signal corresponding to the predistorted baseband digital signal;

a predistortion actuator model configured to determine predistortion coefficients for each of the branches of the predistortion actuator, wherein the predistortion actuator predistorts the baseband digital input signal based on the predistortion coefficients;

a power amplifier configured to amplify the predistorted RF signal to generate an RF output signal; and

a downconversion function configured to generate a baseband digital output signal corresponding to the RF output signal,

wherein the predistortion actuator model determines the predistortion coefficients for each of the branches of the predistortion actuator based on the baseband digital output signal and the predistorted baseband digital signal.

20. The power amplifier system of claim 19, wherein each sectioned basis function comprises a normalization factor that varies on at least one of:

a per section basis; and

a per branch basis.