Adaptive demodulation reference signal (DMRS)
A method of wireless communication by a user equipment (UE) indicates, to a base station, a training state of a machine learning model for a given channel condition, and a request for a change in demodulation reference signal (DMRS) transmissions. The UE also receives DMRS transmissions in accordance with the training state for the given channel condition. The UE performs online training of the machine learning model with the DMRS transmissions. A UE may also request, from a base station, a specific number of demodulation reference signal (DMRS) symbols for a slot, and receive DMRS transmissions in response to the request to estimate a raw channel.
1. A method of wireless communication by a user equipment (UE), comprising:
indicating, to a base station, a training state of a machine learning model for a given channel condition, and a request for a change in demodulation reference signal (DMRS) transmissions;
receiving DMRS transmissions in accordance with the training state for the given channel condition; and
performing online training of the machine learning model with the DMRS transmissions.
2. The method of claim 1 , in which the training state indicates a decoding quality is above a threshold value, and the receiving comprises receiving a light DMRS pattern.
3. The method of claim 1 , in which the training state indicates a decoding quality is above a threshold value, and the receiving comprises receiving fewer DMRS symbols in a slot.
4. The method of claim 1 , in which the training state indicates a decoding quality is below a threshold value, and the receiving comprises receiving a heavy DMRS pattern.
5. The method of claim 1 , in which the training state indicates a decoding quality is below a threshold value, and the receiving comprises receiving additional DMRS symbols in a slot.
6. The method of claim 1 , in which the training state is based on a bit error rate associated with decoding by the machine learning model.
7. The method of claim 1 , in which the indicating occurs via a physical uplink control channel.
8. The method of claim 1 , further comprising indicating, to the base station, online training as a reason for indicating the training state.
9. The method of claim 1 , in which a pattern of the received DMRS differs from a legacy DMRS pattern.
10. A method of wireless communication by a user equipment (UE), comprising:
requesting, from a base station, a specific number of demodulation reference signal (DMRS) symbols for a slot; and
receiving DMRS transmissions across a plurality of beams, in accordance with the requesting to estimate a raw, non-beamformed channel including the plurality of beams.
11. The method of claim 10 , in which requesting comprises requesting a specific number of repeated DMRS symbols.
12. The method of claim 10 , in which requesting comprises requesting a specific number of newly defined DMRS symbols.
13. The method of claim 10 , in which the specific number of DMRS symbols comprises three symbols or four symbols.
14. The method of claim 13 , in which the three symbols or four symbols comprise three consecutive symbols or four consecutive symbols.
15. The method of claim 10 , further comprising performing a beam sweep across a specific number of beams corresponding to the specific number of DMRS symbols to obtain an estimate of the raw, non-beamformed channel.
16. The method of claim 10 , further comprising requesting a rank corresponding to the specific number of DMRS symbols.
17. An apparatus for wireless communication by a user equipment (UE), comprising:
a processor,
memory coupled with the processor; and
instructions stored in the memory and operable, when executed by the processor, to cause the apparatus:
to indicate, to a base station, a training state of a machine learning model for a given channel condition, and a request for a change in demodulation reference signal (DMRS) transmissions;
to receive DMRS transmissions in accordance with the training state for the given channel condition; and
to perform online training of the machine learning model with the DMRS transmissions.
18. The apparatus of claim 17 , in which the training state indicates a decoding quality is above a threshold value, and in which the processor causes the apparatus to receive a light DMRS pattern.
19. The apparatus of claim 17 , in which the training state indicates a decoding quality is above a threshold value, and in which the processor causes the apparatus to receive fewer DMRS symbols in a slot.
20. The apparatus of claim 17 , in which the training state indicates a decoding quality is below a threshold value, and in which the processor causes the apparatus to receive a heavy DMRS pattern.
21. The apparatus of claim 17 , in which the training state indicates a decoding quality is below a threshold value, and in which the processor causes the apparatus to receive additional DMRS symbols in a slot.
22. The apparatus of claim 17 , in which the training state is based on a bit error rate associated with decoding by the machine learning model.
23. The apparatus of claim 17 , in which the processor causes the apparatus to indicate via a physical uplink control channel.
24. The apparatus of claim 17 , in which the processor causes the apparatus to indicate, to the base station, online training as a reason for indicating the training state.
25. The apparatus of claim 17 , in which a pattern of the received DMRS differs from a legacy DMRS pattern.
26. An apparatus for wireless communication by a user equipment (UE), comprising:
a processor,
memory coupled with the processor; and
instructions stored in the memory and operable, when executed by the processor, to cause the apparatus:
to request, from a base station, a specific number of demodulation reference signal (DMRS) symbols for a slot; and
to receive DMRS transmissions across a plurality of beams, in accordance with the requesting to estimate a raw, non-beamformed channel including the plurality of beams.
27. The apparatus of claim 26 , in which the processor causes the apparatus to request a specific number of repeated DMRS symbols.
28. The apparatus of claim 26 , in which the processor causes the apparatus to request a specific number of newly defined DMRS symbols.
29. The apparatus of claim 26 , in which the specific number of DMRS symbols comprises three symbols or four symbols.
30. The apparatus of claim 29 , in which the three symbols or four symbols comprise three consecutive symbols or four consecutive symbols.
31. The apparatus of claim 26 , in which the processor causes the apparatus to perform a beam sweep across a specific number of beams corresponding to the specific number of DMRS symbols to obtain an estimate of the raw, non-beamformed channel.
32. The apparatus of claim 26 , in which the processor causes the apparatus to request a rank corresponding to the specific number of DMRS symbols.
33. A user equipment (UE) for wireless communications, comprising:
means for indicating, to a base station, a training state of a machine learning model for a given channel condition, and a request for a change in demodulation reference signal (DMRS) transmissions;
means for receiving DMRS transmissions in accordance with the training state for the given channel condition; and
means for performing online training of the machine learning model with the DMRS transmissions.
34. The UE of claim 33 , in which the training state indicates a decoding quality is above a threshold value, and the receiving means comprises means for receiving a light DMRS pattern.
35. The UE of claim 33 , in which the training state indicates a decoding quality is above a threshold value, and the receiving means comprises means for receiving fewer DMRS symbols in a slot.
36. The UE of claim 33 , in which the training state indicates a decoding quality is below a threshold value, and the receiving means comprises means for receiving a heavy DMRS pattern.
37. The UE of claim 33 , in which the training state indicates a decoding quality is below a threshold value, and the receiving means comprises means for receiving additional DMRS symbols in a slot.
38. The UE of claim 33 , in which the training state is based on a bit error rate associated with decoding by the machine learning model.
39. The UE of claim 33 , in which the means for indicating operates via a physical uplink control channel.
40. The UE of claim 33 , further comprising means for indicating, to the base station, online training as a reason for indicating the training state.
41. The UE of claim 33 , in which a pattern of the received DMRS differs from a legacy DMRS pattern.
42. A user equipment (UE) for wireless communications, comprising:
means for requesting, from a base station, a specific number of demodulation reference signal (DMRS) symbols for a slot; and
means for receiving DMRS transmissions across a plurality of beams, in accordance with the requesting to estimate a raw, non-beamformed channel including the plurality of beams.
43. The UE of claim 42 , in which the requesting means comprises means for requesting a specific number of repeated DMRS symbols.
44. The UE of claim 42 , in which the requesting means comprises means for requesting a specific number of newly defined DMRS symbols.
45. The UE of claim 42 , in which the specific number of DMRS symbols comprises three symbols or four symbols.
46. The UE of claim 45 , in which the three symbols or four symbols comprise three consecutive symbols or four consecutive symbols.
47. The UE of claim 42 , further comprising means for performing a beam sweep across a specific number of beams corresponding to the specific number of DMRS symbols to obtain an estimate of the raw, non-beamformed channel.
48. The UE of claim 42 , further comprising means for requesting a rank corresponding to the specific number of DMRS symbols.
49. A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a user equipment (UE) and comprising:
program code to indicate, to a base station, a training state of a machine learning model for a given channel condition, and a request for a change in demodulation reference signal (DMRS) transmissions;
program code to receive DMRS transmissions in accordance with the training state for the given channel condition; and
program code to perform online training of the machine learning model with the DMRS transmissions.
50. A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a user equipment (UE) and comprising:
program code to request, from a base station, a specific number of demodulation reference signal (DMRS) symbols for a slot; and
program code to receive DMRS transmissions across a plurality of beams, in accordance with the requesting to estimate a raw, non-beamformed channel including the plurality of beams.