Encoding and decoding of information for wireless transmission using multi-antenna transceivers
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training and deploying machine-learned communication over multi-input-multi-output (MIMO) channels. One of the methods includes: determining a transmitter and a receiver, at least one of which implements a machine-learning network; determining a MIMO channel model; determining first information; using the transmitter to process the first information and generate first RF signals representing inputs to the MIMO channel model; determining second RF signals representing outputs of the MIMO channel model, each second RF signal representing aggregated reception of the first RF signals altered by transmission through the MIMO channel model; using the receiver to process the second RF signals and generate second information as a reconstruction of the first information; calculating a measure of distance between the second and first information; and updating the machine-learning network based on the measure of distance between the second and first information.
1. A method performed by at least one processor for communication using multiple receive antennas over a wireless communication channel, the method comprising:
receiving, at a receiver device associated with the multiple receive antennas, a plurality of radio frequency (RF) signals based on transmission by a source device over the wireless communication channel;
estimating channel state information (CSI) corresponding to the wireless communication channel based on processing the plurality of RF signals
using at least one machine-learning network that is trained to estimate CSI and is associated with the receiver device, wherein the CSI indicates at least one of
a state of the wireless communication channel,
spatial information corresponding to the multiple receive antennas, or
scheduling information corresponding to multiple users of the wireless communication channel; and
providing, by the receiver device, the estimated CSI to the source device for processing further RF signals for transmission over the wireless communication channel.
2. The method of claim 1 , wherein the source device encodes first information in the plurality of RF signals, and wherein estimating the CSI comprises:
generating second information based on decoding the plurality of RF signals at the receiver device, the second information corresponding to the first information modified by one or more channel effects during transmission over the wireless communication channel; and
estimating the CSI based on processing the second information using the at least one machine-learning network.
3. The method of claim 2 , wherein generating the second information based on decoding the plurality of RF signals at the receiver device comprises decoding the plurality of RF signals using at least a second machine learning network at the receiver device to generate the second information.
4. The method of claim 3 , wherein the second machine learning network comprises at least one of
a linear layer for regression of reconstructed information in decoding the plurality of RF signals,
a sigmoid activation layer for one of probability regression or slicing of the plurality of RF signals, or
a combination of sigmoid expressions configured to compute a pseudo-likelihood or pseudo-probability of one or more of a discrete message, discrete portion of a message, or one or more bits.
5. The method of claim 3 , wherein the source device encodes the first information in the plurality of RF signals using at least a third machine learning network at the source device.
6. The method of claim 1 , wherein receiving the plurality of RF signals at the receiver device comprises receiving the plurality of RF signals using a plurality of first antennas of the multiple receive antennas at the receiver device, and wherein
the plurality of RF signals are transmitted by the source device using a plurality of second antennas at the source device.
7. The method of claim 6 , wherein the wireless communication channel comprises a multi-input-multi-output (MIMO) wireless communication channel.
8. The method of claim 1 , wherein the receiver device comprises a base station and the source device comprises a mobile device.
9. The method of claim 1 , wherein the receiver device comprises a mobile device and the source device comprises a base station.
10. The method of claim 1 , wherein the plurality of RF signals comprises a QAM waveform.
11. The method of claim 1 , wherein the wireless communication channel comprises a cellular communication channel and wherein the plurality of RF signals comprises at least one of LTE, 5G, or 6G cellular signals.
12. An apparatus for communication using multiple receive antennas over a wireless communication channel, the apparatus comprising:
one or more processors; and
computer-readable media storing instructions that, when executed, are configured to cause the one or more processors to perform operations comprising:
receiving, at a receiver device associated with the multiple receive antennas, a plurality of radio frequency (RF) signals based on transmission by a source device over the wireless communication channel;
estimating channel state information (CSI) corresponding to the wireless communication channel based on processing the plurality of RF signals using at least one machine-learning network that is trained to estimate CSI and is associated with the receiver device, wherein the CSI indicates at least one of
a state of the wireless communication channel,
spatial information corresponding to the multiple receive antennas, or
scheduling information corresponding to multiple users of the wireless communication channel; and
providing, by the receiver device, the estimated CSI to the source device for processing further RF signals for transmission over the wireless communication channel.
13. The apparatus of claim 12 , wherein the source device encodes first information in the plurality of RF signals, and wherein estimating the CSI comprises:
generating second information based on decoding the plurality of RF signals at the receiver device, the second information corresponding to the first information modified by one or more channel effects during transmission over the wireless communication channel; and
estimating the CSI based on processing the second information using the at least one machine-learning network.
14. The apparatus of claim 13 , wherein generating the second information based on decoding the plurality of RF signals at the receiver device comprises decoding the plurality of RF signals using at least a second machine learning network at the receiver device to generate the second information.
15. The apparatus of claim 14 , wherein the second machine learning network comprises at least one of
a linear layer for regression of reconstructed information in decoding the plurality of RF signals,
a sigmoid activation layer for one of probability regression or slicing of the plurality of RF signals, or
a combination of sigmoid expressions configured to compute a pseudo-likelihood or pseudo-probability of one or more of a discrete message, discrete portion of a message, or one or more bits.
16. The apparatus of claim 14 , wherein the source device encodes the first information in the plurality of RF signals using at least a third machine learning network at the source device.
17. The apparatus of claim 12 , wherein receiving the plurality of RF signals at the receiver device comprises receiving the plurality of RF signals using a plurality of first antennas of the multiple receive antennas at the receiver device, and wherein
the plurality of RF signals are transmitted by the source device using a plurality of second antennas at the source device.
18. The apparatus of claim 17 , wherein the wireless communication channel comprises a multi-input-multi-output (MIMO) wireless communication channel.
19. The apparatus of claim 12 , wherein the receiver device comprises a base station and the source device comprises a mobile device.
20. The apparatus of claim 12 , wherein the receiver device comprises a mobile device and the source device comprises a base station.
21. The apparatus of claim 12 , wherein the plurality of RF signals comprises a QAM waveform.
22. The apparatus of claim 12 , wherein the wireless communication channel comprises a cellular communication channel and wherein the plurality of RF signals comprises at least one of LTE, 5G, or 6G cellular signals.