Enhanced DPD method to correct gain flatness for wireless communications systems
In a dynamic signal traffic scenario, a narrowband to wideband transition in a DPD system results in a tilt in the output spectrum until the next DPD adaptation cycle occurs. To address this problem, regularization term is applied with a weighing factor when performing DPD coefficient estimation and adaptation. The regularization term can be obtained from in a variety of ways: using pre-stored waveforms, through factory or in-situ calibration, or through an adaptive or opportunistic update by observing the system. Application of the regularization term improves the spectrum flatness for a narrow to wideband signal transition, and does not require transmitting additional calibration tones to correct the gain flatness.
1. A transmitter with gain flatness correction when transitioning from narrowband operation to wideband operation, comprising:
an input;
a digital predistortion function to predistort the input and generate a predistorted input;
an amplifier to receive the predistorted input, amplify the predistorted input, and generate an output;
data store to store (1) one or more captures of the predistorted input and the output of the amplifier and (2) regularization data, wherein the regularization data is pre-loaded on the data store; and
adaptation logic coupled to the data store to determine coefficients for the digital predistortion function.
2. The transmitter of claim 1 , wherein the regularization data is captured while the amplifier is processing a wideband signal.
3. The transmitter of claim 2 , wherein the wideband signal processed by the amplifier is part of normal data traffic.
4. The transmitter of claim 1 , wherein the adaptation logic is further to:
compute a joint matrix based on (1) captures of the predistorted input and the output of the amplifier and (2) regularization data, and
determine the coefficients based on the joint matrix.
5. The transmitter of claim 1 , wherein the transmitter is to switch from narrowband operation to wideband operation.
6. The transmitter of claim 1 , wherein the transmitter is a part of a massive input massive output radio.
7. One or more non-transitory computer-readable media having instructions encoded thereon, the instructions being executable by one or more processors to perform the following:
triggering one or more captures of a predistorted input signal and an output signal of an amplifier;
receiving the one or more captures;
receiving regularization data corresponding to a wideband input signal and a wideband output signal;
forming a cost function based on the one or more captures and the regularization data; and
determining coefficients that minimizes the cost function, wherein the coefficients are applied to predistort an input signal to the amplifier.
8. The one or more non-transitory computer-readable media of claim 7 , wherein the regularization data is pre-loaded on the one or more non-transitory computer-readable media.
9. The one or more non-transitory computer-readable media of claim 7 , further comprising:
triggering the wideband input signal to be transmitted as the input signal of the amplifier; and
triggering the regularization data to be captured while the wideband input signal is transmitted by the amplifier.
10. The one or more non-transitory computer-readable media of claim 7 , further comprising:
detecting that that normal data traffic being transmitted as the input signal of the amplifier represents a wideband signal; and
triggering the regularization data to be captured while the normal data traffic is transmitted by the amplifier.
11. A method for digital predistortion for an amplifier, the method comprising:
receiving one or more captures of a predistorted input signal and an output signal of the amplifier;
retrieving regularization data corresponding to a wideband input signal and a wideband output signal of the amplifier; and
determining coefficients for predistorting an input signal to the amplifier based on the one or more captures and the regularization data.
12. The method of claim 11 , wherein the regularization data is pre-stored data representing input and output signals received and output respectively by the amplifier during wideband operation of the amplifier.
13. The method of claim 11 , wherein the regularization data is pre-stored data representing wideband input and output signals received and output respectively by the amplifier as if the amplifier was linear.
14. The method of claim 11 , further comprising:
transmitting a wideband signal through the amplifier while the amplifier is not transmitting normal data traffic; and
storing, as the regularization data, a capture of the wideband signal and the output signal of the amplifier while the amplifier is transmitting the wideband signal.
15. The method of claim 11 , further comprising:
providing normal data traffic as input to the amplifier;
detecting that the normal data traffic is or is close to being a wideband signal; and
storing, as the regularization data, a capture of the wideband signal and the output signal of the amplifier while the amplifier is transmitting the wideband signal.
16. The method of claim 11 , further comprising:
retrieving one or more capture covariance matrices formed from the one or more captures; and
retrieving a regularization covariance matrix formed from the regularization data.
17. The method of claim 16 , further comprising:
combining the one or more capture covariance matrices and the regularization covariance matrix into a joint matrix.
18. The method of claim 16 , further comprising:
combining the one or more capture covariance matrices and the regularization covariance matrix into a joint matrix based on a weighing factor.
19. The method of claim 11 , further comprising:
predistorting input data to the amplifier based on the determined coefficients.