Similarity learning for crowd-sourced positioning
Aspects presented herein may enhance the accuracy and/or latency of UE positioning based on crowd-sourcing, where a network entity may compute a position estimate of a UE based on neighbor-cell scan data from the UE and one or more reference UEs. In one aspect, a network entity receives a first set of measurements associated with at least one cell from a UE. The network entity performs a position estimation of the UE based on at least one of the first set of measurements associated with the at least one cell, a second set of measurements for each of a set of reference UEs, or a location of each of the set of reference UEs via an ML model, where the UE and the set of reference UEs include at least one common cell.
1. An apparatus for wireless communication at a network entity, comprising:
memory; and
at least one processor coupled to the memory and, based at least in part on information stored in the memory, the at least one processor is configured to:
receive, from a user equipment (UE), a first set of measurements associated with at least one cell; and
perform a position estimation of the UE based on at least one of the first set of measurements associated with the at least one cell, a second set of measurements for each of a set of reference UEs, or a location of each of the set of reference UEs via at least one machine learning (ML) model, wherein the UE and the set of reference UEs include at least one common cell, wherein the at least one ML model is capable of identifying the at least one common cell between the UE and the set of reference UEs, and wherein an output of the at least one ML model includes a similarity score between the UE and the set of reference UEs, wherein the similarity score is on a per-cell basis.
2. The apparatus of claim 1 , wherein the position estimation of the UE is further based on a location of the at least one cell.
3. The apparatus of claim 1 , wherein the first set of measurements or the second set of measurements include at least one of:
a reference signal received power (RSRP) for the at least one cell,
a coverage area for the at least one cell,
a channel response for the at least one cell,
device characteristics of the UE or each of the set of reference UEs,
the location of each of the set of reference UEs, or
a position uncertainty for the location of each of the set of reference UEs.
4. The apparatus of claim 3 , wherein the channel response includes at least one of a timing advance (TA), a reference signal received quality (RSRQ), a received signal strength indicator (RSSI), a reference signal signal to noise ratio (RSSNR), or a channel quality indicator (CQI).
5. The apparatus of claim 1 , wherein a set of outputs from the at least one ML model is fused based on a weighted sum associated with the set of reference UEs, the at least one cell, the at least one common cell, or a combination thereof.
6. The apparatus of claim 5 , wherein the weighted sum is based on a number of the set of reference UEs, a number of cells reported by each of the set of reference UEs, a number of cells reported by the UE, a number of the at least one common cell.
7. The apparatus of claim 1 , wherein the at least one processor is further configured to:
receive an indication of criteria associated with the UE from the UE; and
select the at least one ML model and a type of fusion for the at least one ML model for performing the position estimation of the UE based on the indication.
8. The apparatus of claim 1 , wherein the at least one ML model includes a first ML model and a second ML model, the first ML model being configured to infer a first set of weights for the first set of measurements associated with the at least one cell and the second ML model being configured to infer a second set of weights for the second set of measurements for each of the set of reference UEs.
9. The apparatus of claim 8 , wherein the position estimation of the UE is based on fusing outputs of the first ML model based on the first set of weights and the second ML model based on the second set of weights.
10. The apparatus of claim 1 , wherein the at least one common cell is a single common cell, and wherein the ML model identifies the at least one common cell based on an identifier (ID) of the single common cell.
11. The apparatus of claim 1 , wherein the at least one common cell is multiple common cells, and wherein the ML model identifies the at least one common cell based on an identifier (ID) of the multiple common cells.
12. The apparatus of claim 1 , wherein the at least one processor is further configured to:
train the at least one ML model using one or more measurements from the set of reference UEs for the at least one common cell.
13. The apparatus of claim 1 , wherein the at least one cell for the UE is a serving cell or a non-serving cell across multiple frequency bands.
14. The apparatus of claim 1 , wherein the at least one processor is further configured to:
transmit a configuration for the UE to configure the UE to measure the at least one cell; and
receive the first set of measurements associated with the at least one cell in response to the configuration.
15. The apparatus of claim 1 , wherein the at least one common cell is associated with at least one feature including one or more of:
a measurement for the at least one common cell,
a UE type,
an environmental condition,
a cell characteristic associated with the at least one common cell, or
a combination thereof.
16. The apparatus of claim 1 , further comprising at least one of a transceiver or an antenna coupled to the at least one processor, wherein to receive the first set of measurements, the at least one processor is configured to receive the first set of measurements via at least one of the transceiver or the antenna.
17. A method of wireless communication at a network entity, comprising:
receiving, from a user equipment (UE), a first set of measurements associated with at least one cell; and
performing a position estimation of the UE based on at least one of the first set of measurements associated with the at least one cell, a second set of measurements for each of a set of reference UEs, or a location of each of the set of reference UEs via at least one machine learning (ML) model, wherein the UE and the set of reference UEs include at least one common cell, wherein the at least one ML model is capable of identifying the at least one common cell between the UE and the set of reference UEs, and wherein an output of the at least one ML model includes a similarity score between the UE and the set of reference UEs, wherein the similarity score is on a per-cell basis.
18. The method of claim 17 , wherein the position estimation of the UE is further based on a location of the at least one cell.
19. The method of claim 17 , wherein the first set of measurements or the second set of measurements include at least one of:
a reference signal received power (RSRP) for the at least one cell,
a coverage area for the at least one cell,
a channel response for the at least one cell,
device characteristics of the UE or each of the set of reference UEs,
the location of each of the set of reference UEs, or
a position uncertainty for the location of each of the set of reference UEs.
20. The method of claim 19 , wherein the channel response includes at least one of a timing advance (TA), a reference signal received quality (RSRQ), a received signal strength indicator (RSSI), a reference signal signal to noise ratio (RSSNR), or a channel quality indicator (CQI).
21. The method of claim 17 , wherein a set of outputs from the at least one ML model is fused based on a weighted sum associated with the set of reference UEs, the at least one cell, the at least one common cell, or a combination thereof.
22. The method of claim 21 , wherein the weighted sum is based on a number of the set of reference UEs, a number of cells reported by each of the set of reference UEs, a number of cells reported by the UE, a number of the at least one common cell.
23. The method of claim 17 , further comprising:
receiving an indication of criteria associated with the UE from the UE; and
selecting the at least one ML model and a type of fusion for the at least one ML model for performing the position estimation of the UE based on the indication.
24. The method of claim 17 , wherein the at least one ML model includes a first ML model and a second ML model, the first ML model being configured to infer a first set of weights for the first set of measurements associated with the at least one cell and the second ML model being configured to infer a second set of weights for the second set of measurements for each of the set of reference UEs.
25. The method of claim 24 , wherein the position estimation of the UE is based on fusing outputs of the first ML model based on the first set of weights and the second ML model based on the second set of weights.
26. The method of claim 25 , wherein the at least one common cell is a single common cell or multiple common cells, and wherein the at least one ML model identifies the at least one common cell based on an identifier (ID) of the single common cell or the multiple common cells.
27. The method of claim 17 , further comprising:
training the at least one ML model using one or more measurements from the set of reference UEs for the at least one common cell.
28. The method of claim 17 , further comprising:
transmitting a configuration for the UE to configure the UE to measure the at least one cell; and
receiving the first set of measurements associated with the at least one cell in response to the configuration.
29. An apparatus for wireless communication at a network entity, comprising:
means for receiving, from a user equipment (UE), a first set of measurements associated with at least one cell; and
means for performing a position estimation of the UE based on at least one of the first set of measurements associated with the at least one cell, a second set of measurements for each of a set of reference UEs, or a location of each of the set of reference UEs via at least one machine learning (ML) model, wherein the UE and the set of reference UEs include at least one common cell, wherein the at least one ML model is capable of identifying the at least one common cell between the UE and the set of reference UEs, and wherein an output of the at least one ML model includes a similarity score between the UE and the set of reference UEs, wherein the similarity score is on a per-cell basis.
30. A non-transitory computer-readable medium storing computer executable code at a network entity, the code when executed by a processor causes the processor to:
receive, from a user equipment (UE), a first set of measurements associated with at least one cell; and
perform a position estimation of the UE based on at least one of the first set of measurements associated with the at least one cell, a second set of measurements for each of a set of reference UEs, or a location of each of the set of reference UEs via at least one machine learning (ML) model, wherein the UE and the set of reference UEs include at least one common cell, wherein the at least one ML model is capable of identifying the at least one common cell between the UE and the set of reference UEs, and wherein an output of the at least one ML model includes a similarity score between the UE and the set of reference UEs, wherein the similarity score is on a per-cell basis.