A digital predistortion (DPD) system includes an input configured to receive an input signal. In some examples, a first signal path configured to generate a first signal based on the input signal. In some examples, an error model provider configured to generate an error model signal modeled after a gate bias error voltage associated with the DPD system. In some examples, a first combiner configured to combine the first signal and the error model signal to generate a first intermediate signal, and the DPD system generates an output signal based at least on the first intermediate signal.

Fast envelope tracking systems are provided herein. In certain embodiments, an envelope tracking system for a power amplifier includes a switching regulator and a differential error amplifier configured to operate in combination with one another to generate a power amplifier supply voltage for the power amplifier based on an envelope of a radio frequency (RF) signal amplified by the power amplifier. The envelope tracking system further includes a differential envelope amplifier configured to amplify a differential envelope signal to generate a single-ended envelope signal that changes in relation to the envelope of the RF signal. Additionally, the differential error amplifier generates an output current operable to adjust a voltage level of the power amplifier supply voltage based on comparing the single-ended envelope signal to a reference signal.

The disclosure presents a method and an amplifier system for minimizing variation in an acoustical signal caused by variation in gain of an amplifier, comprising a battery for providing a supply voltage to the amplifier, a digital signal processor for providing the acoustical signal to the amplifier, a controller unit receiving an enablement signal when the supply voltage is in an offset mode, and based on the enablement signal requesting a measured voltage during a time period, and a first analog-to-digital converter configured for measuring the supply voltage to the amplifier when receiving the request from the controller unit or the first analog-to-digital converter is configured for measuring the supply voltage to the amplifier continuously, and where variations in the measured voltage relates to variations in the supply voltage during the time period. Furthermore, the controller unit is configured to predict offset modes (i.e. changes) in the supply voltage based on the enablement signals and a fitting of the measured voltages, and wherein the controller unit is configured to generate a compensating signal based on the fitting and transmit the compensating signal to the digital signal processor, the digital signal processor is then configured to minimize variation in the acoustical signal at the output of the amplifier by compensating the variation in gain of the amplifier based on the compensating signal.

The present disclosure provides a method of linearization for a non-linear system, comprising: a group of steps comprising: inputting an input signal to the non-linear system; obtaining an output signal from the non-linear system in response to the input signal being input to the non-linear system; obtaining a desired signal for the non-linear system; comparing the output signal with the desired signal, to determine whether the input signal is a target signal for the non-linear system to achieve a desired performance; and in response to determining that the input signal is not the target signal for the non-linear system to achieve the desired performance, generating an updated signal in frequency domain by applying a coefficient to eliminate a difference between the output signal and the desired signal, for updating the input signal to the non-linear system. The present disclosure also provides a corresponding device, computer programs, and computer-readable storage.

The invention relates to a device for linearising a power amplifier by employing digital predistortion, comprising: a digital predistortion module, configured to infer a polar domain predistortion to be applied to a signal, and comprising a first neural network and a second neural network respectively configured to correct amplitude and phase distortion produced by the amplifier; an optimisation module of each of said neural networks configured to implement meta-learning, using: a meta-initialisation providing a prior initialisation of the initial weights of each of said neural networks; a meta-matching of the initial weights into optimal weights of each of said neural networks.

Systems, devices, and methods related to hybrid basis function, neural network-based digital predistortion (DPD) are provided. An example apparatus for a radio frequency (RF) transceiver includes a digital predistortion (DPD) actuator to receive an input signal associated with a nonlinear component of the RF transceiver and output a predistorted signal. The DPD actuator includes a basis-function-based actuator to perform a first DPD operation using a set of basis functions associated with a first nonlinear characteristic of the nonlinear component. The DPD actuator further includes a neural network-based actuator to perform a second DPD operation using a first neural network associated with a second nonlinear characteristic of the nonlinear component. The predistorted signal is based on a first output signal of the basis-function-based actuator and a second output signal of the neural network-based actuator.

Various communication devices may benefit from the appropriate use of modeling techniques. For example, devices that include components that may be driven into non-linear ranges of operation may benefit from low complexity non-linear modelling techniques. Such devices may be used, for example, in wireless communication systems. A method can include obtaining a sample of a signal representative of power consumed by a device while the device is operating in a non-linear range while being driven according to a driving signal. The method can also include computing a correction to the driving signal based on the sample. The correction can be calculated based on a plurality of non-overlapped non-linear sections corresponding to a response of the device. The method can further include applying the correction to adjust the driving signal. The correction can be configured to adjust the power to a desired value of power.

Envelope tracking systems with modeling for power amplifier supply voltage filtering are provided herein. In certain embodiments, an envelope tracking system includes a supply voltage filter, a power amplifier that receives a power amplifier supply voltage through the supply voltage filter, and an envelope tracker that generates the power amplifier supply voltage. The power amplifier provides amplification to a radio frequency (RF) signal that is generated based on digital signal data, and the envelope tracker generates the power amplifier supply voltage based on an envelope signal corresponding to an envelope of the RF signal. The envelope tracking system further includes digital modeling circuitry that models the supply voltage filter and operates to digitally compensate the digital signal data for effects of the supply voltage filter, such as distortion.

A predistortion circuit for a wireless transmitter includes a signal input configured to receive a baseband signal. Further, the predistortion circuit includes a predistorter configured to generate a predistorted baseband signal using the baseband signal and a select of one of a first predistorter configuration and a second predistorter configuration.

Various examples are directed to systems and methods for digital predistortion (DPD). A linear digital predistortion (DPD) circuit may be programmed to generate a pre-distorted signal linear component based at least in part on a complex baseband signal. A nonlinear DPD circuit may be programmed to generate a pre-distorted signal nonlinear component based at least in part on the complex baseband signal. A mixer circuit programmed to generate a pre-distorted signal based at least in part on the pre-distorted signal linear component and the pre-distorted signal nonlinear component.