Wireless devices and systems including examples of compensating I/Q imbalance with neural networks or recurrent neural networks
View Patent ↗Examples described herein include methods, devices, and systems which compensates input data for I/Q imbalance or noise related thereto to generate compensated input data. In doing such the above compensation, during an uplink transmission time interval (TTI), a switch path is activated to provide converted input data to a receiver stage including a recurrent neural network (RNN). The RNN calculates an error representative of the noise based partly on the input signal to be transmitted and a feedback signal to generate filter coefficient data associated with the I/Q imbalance. The feedback signal is provided, after processing through the receiver, to the RNN. During an uplink TTI, the converted input data is transmitted as the RF wireless transmission via an RF antenna. During a downlink TTI, the switch path is deactivated and the receiver stage receives an additional RF wireless transmission to be processed in the receiver stage.
1. An apparatus, comprising:
a transceiver configured to generate, using a digital to analog converter (DAC), a plurality of radio frequency (RF) signals for transmission based on input signals;
a first set of processing units configured to generate a first intermediate processing result by mixing the plurality of RF signals received via the transceiver as feedback and first signals based on respective outputs of the first processing units using a first plurality of coefficients; and
aa second set of processing units configured to generate filter coefficient data by mixing the first intermediate processing result and second signals that are based on respective outputs of the respective second set of processing units using a second plurality of coefficients, wherein the filter coefficient data is applied to the input signals prior to provision to the DAC.
2. The apparatus of claim 1 , wherein the first set of processing units each include a multiplication/accumulation unit.
3. The apparatus of claim 2 , wherein the second set of processing units each include a multiplication/accumulation unit.
4. The apparatus of claim 1 , further comprising a digital filter configured to filter the input signals prior to provision to the DAC based on the filter coefficient data.
5. The apparatus of claim 4 , wherein the digital filter is further configured to provide output data as a compensated plurality of RF signals, wherein the output data is based partly on the filter coefficient data.
6. The apparatus of claim 4 , wherein a recurrent neural network (RNN) is configured to provide the filter coefficient data to the digital filter.
7. The apparatus of claim 6 , wherein the RNN is configured receive the plurality of RF signals as feedback via a receive path of the transceiver.
8. The apparatus of claim 4 , wherein the transceiver includes an RF mixer to generate the plurality of radio frequency (RF) signals for transmission.
9. The apparatus of claim 8 , wherein the digital filter is further configured to filter signals prior to the input to the DAC and prior to an input to the RF mixer, in accordance with the filter coefficient data calculated using the plurality of MAC units.
10. The apparatus of claim 1 , further comprising a switch coupled to the transceiver and configured to activate a switch path that provides the plurality of RF signals as the feedback from a transmit path of the transceiver via a receive path of the transceiver.
11. The apparatus of claim 10 , wherein the switch is further configured to receive a selection signal indicating whether the switch path is to be activated, the selection signal based at least in part on a transmission time interval (TTI) of a time-division duplex (TDD) configured radio frame.
12. The apparatus of claim 11 , wherein the selection signal is configured to indicate that the switch path is to be activated based on an uplink TTI of the TDD configured radio frame.
13. The apparatus of claim 11 , wherein the selection signal is configured to indicate that the switch path is to be deactivated based on a downlink TTI of the TDD configured radio frame.
14. The apparatus of claim 11 , wherein the switch is further configured to, responsive to the selection signal indicating the switch path is to be activated, activate the switch path that couples the transmit path to the receive path to provide the feedback.
15. The apparatus of claim 11 , wherein the switch is further configured to deactivate the switch path responsive to the selection signal indicating the switch path is to be deactivated.
16. The apparatus of claim 1 , wherein the plurality of RF signals include data indicative of at least one of a livestock status, a water usage status, an agricultural field status, a wind measurement, a power generation status, an oil flow status, an energy storage status, or a power consumption status.
17. A method, comprising:
generating, at a transceiver, a first transmission signal based on a first input signal by converting the first transmission signal using a digital to analog converter (DAC);
calculating, at a recurrent neural network (RNN), a first processing result based on the first transmission signal and delayed versions of respective outputs of a first plurality of processing units with a first plurality of coefficients;
calculating, at the RNN, output data based on the first processing results and delayed versions of respective outputs of a respective additional plurality of processing units with a second plurality of coefficients; and
filtering a second transmission signal based on the output data.
18. The method of claim 17 , further comprising providing the first transmission signal as feedback from a transmit path of a transceiver to the RNN as the input data via a receive path of the transceiver.
19. The method of claim 18 , further comprising providing the first transmission signal as the feedback after activating a switch path that couples the transmit path to the RNN.
20. The method of claim 17 , further comprising providing, from the RNN, the output data as filter coefficient data to a digital filter that at least partially compensates I/Q imbalance in transmission signals.