This specification relates to an apparatus which stores a machine-learned model including a first neural network block including multiple convolutional neural network layers, a second neural network block including at least one dilated convolutional neural network layer, and a linear transformation block. The apparatus is configured to receive input data representing in-phase and quadrature signals of an input signal that is to be amplified by a power amplifier, to process the received input data using the first neural network block of the machine learned model to generate a first neural network block output, to process the received input data using the second neural network block of the machine learned model to generate a second neural network block output, and to combine, using the linear transformation block, the first neural network block output and the second neural network block output to generate a pre-distorted signal for amplification by the power amplifier.

A method and system for digital pre-distortion of an input signal to compensate for non-linear operation of a power amplifier. According to one aspect, some embodiments provide overlapping spline functions that are defined for two adjacent bins, where any two spline functions overlap in only one bin. Each spline function is computed as a function of one of an input signal envelope and a delayed signal envelope. According to another aspect, a tap weight evaluator includes a least mean squares, LMS, tap correlator updater configured to modulate a step size of an adaptation process to update each tap weight, the step size being modulated based on an approximate logarithm of the average power of the input to a tap weight computation.

This disclosure provides systems, methods, and apparatus, including computer programs encoded on computer-readable media, for configuring components of transmission circuitry. In one aspect, a first set of linear kernels and a first set of nonlinear kernels associated with a composite digital pre-distortion (DPD) kernel design is determined based on a first iteration of a DPD kernel analysis process. The first set of linear kernels is separated from the first set of nonlinear kernels according to a first iteration of a linear filter separation process. A final set of linear kernels and a final set of nonlinear kernels are determined based on one or more additional iterations of the DPD kernel analysis process and the linear filter separation process. A pre-DPD filter for the transmission circuitry is configured using a final set of filter coefficients derived based on the final set of linear kernels.

Supporting suppression of distortion caused by a power amplifier, PA, included in a transmitter system configured to perform digital predistortion, DPD, and feedforward, FF, linearization on multiple digital input signals relating to different frequency bands, respectively. The PA is used for power amplification in preparation for transmission by a wireless communication network and is operative with an instantaneous bandwidth, IBW. Information is obtained identifying one or more intermodulation, IM, components outside the frequency bands but within the IBW, and caused by said PA. The identified IM components are selectively processed as part of said DPD to thereby suppress formation of at least some of the identified IM components, and/or as part of the FF linearization by adding reference signals to the FF linearization, which reference signals correspond to at least some of the identified IM components.

An RF transmitter arrangement using analog pre-distortion is disclosed. The arrangement includes lower bandwidth circuitry, an analog pre-distorter, and a non-linear amplifier chain. The lower bandwidth circuitry is configured to generate an analog signal. The analog pre-distorter is configured to apply a non-linear distortion to the analog original signal based on a coupled feedback signal in order to generate an RF output signal. The non-linear amplifier chain is configured to amplify the RF output signal to generate a transmission signal relative to the analog original signal. The coupled feedback signal is derived from the transmission signal.

A Cartesian loop circuit includes a reference signal amplifier, a forward path coupled to the reference signal amplifier, a feedback path coupled to the forward path, and a controller. The forward path includes an up-mixer to up mix a forward path signal to a radio frequency signal. The feedback path includes a down-mixer to down mix a feedback signal to a frequency of a baseband reference signal inputted to the forward path. The feedback path provides the down-mixed feedback signal to the forward path. The controller is to perform power control at a low power by controlling a gain of the reference signal amplifier and is to perform power control at a high power by controlling a gain of the down-mixer. At the high power, the controller may perform power control by further controlling the gain of the up-mixer.

In a system, known digital representations are generated, and test analog signals are generated using the known digital representations. The test analog signals are transmitted using a transmitter of a transmission system. The test analog signals are received using a receiver of the transmission system and used to generate test received digital representations. The test received digital representations are cross-correlated with the known digital representations to generate a mixing matrix. The mixing matrix is inverted to generate a de-mixing matrix, which is applied to subsequent digital data to be encoded onto a signal and transmitted by the transmitter to generate pre-compensated digital data.

A Cartesian loop circuit includes a reference signal amplifier, a forward path coupled to the reference signal amplifier, a feedback path coupled to the forward path, and a controller. The forward path includes an up-mixer to up mix a forward path signal to a radio frequency signal. The feedback path includes a down-mixer to down mix a feedback signal to a frequency of a baseband reference signal inputted to the forward path. The feedback path provides the down-mixed feedback signal to the forward path. The controller is to perform power control at a low power by controlling a gain of the reference signal amplifier and is to perform power control at a high power by controlling a gain of the down-mixer. At the high power, the controller may perform power control by further controlling the gain of the up-mixer.

A distortion compensating device includes: a filter that receives input of a transmitting signal including a plurality of subcarrier signals assigned to respective frequencies and that superimposes filter coefficients on the respective subcarrier signals; a first signal converting unit that converts the subcarrier signals, on which the respective filter coefficients are superimposed, from a frequency domain into a time domain to obtain an input signal; a distortion compensating unit that superimposes a distortion compensation coefficient on the input signal to obtain an output signal; a power amplifier that amplifies and outputs the output signal; and a control unit that generates the filter coefficients according to an arithmetic equation using the subcarrier signals and a feedback signal from the power amplifier, and outputs the filter coefficients to the filter.

A radio frequency (RF) communication assembly includes an RF communication circuit and a compensator apparatus. The compensator apparatus receives an input including an I-component of a pre-compensated signal, a Q-component of the pre-compensated signal, and encoded operating conditions of the RF communication circuit. The RF communication circuit includes RF circuit components causing signal impairments. The compensator apparatus perform neural network computing on the input, and the RF communication assembly generates a compensated output signal that compensates for at least a portion of the signal impairments.