Location prediction using hierarchical classification
Some embodiments of the present disclosure provide a neural network location classifier that is designed and trained in accordance with a hierarchical architecture, thereby producing a hierarchical neural network location classifier. Further embodiments relate to obtaining, through use of the hierarchical neural network location classifier, an inferred hierarchical label for a user equipment location. The inferred hierarchical label may then be decoded to obtain a location.
1. A single-transmission-point user equipment (UE) locating method, the method comprising:
receiving a reference signal;
obtaining, by processing the reference signal, a normalized input tensor;
obtaining, through use of a hierarchical neural network location classifier, an inferred hierarchical label for a location corresponding to the normalized input tensor;
obtaining a predicted location by decoding the inferred hierarchical label; and
transmitting an indication of the predicted location.
2. The method of claim 1 wherein the decoding comprises employing an inverse of a hierarchical labelling function, where the hierarchical labelling function is configured to receive, as input, a particular location and produce, as output, a particular hierarchical label to associate with the particular location.
3. The method of claim 2 further comprising:
receiving map information for an environment for the UE; and
carrying out ray-tracing on the map information to, thereby, generate a plurality of label vectors, with an individual label vector associated with each point among a plurality of points in the environment.
4. The method of claim 3 further comprising configuring the hierarchical labelling function based upon applying a hierarchical clustering algorithm to the plurality of label vectors.
5. The method of claim 3 further comprising:
applying a dimensionality reduction tool to the plurality of label vectors to, thereby, generate a plurality of reduced-dimensionality label vectors; and
configuring the hierarchical labelling function based upon applying a hierarchical clustering algorithm to the plurality of reduced-dimensionality label vectors.
6. The method of claim 5 wherein the dimensionality reduction tool comprises an auto-encoder.
7. The method of claim 2 further comprising configuring the hierarchical labelling function based upon global positioning system information associated with an un-normalized input tensor.
8. The method of claim 2 further comprising configuring the hierarchical labelling function based upon vision sensing information associated with an un-normalized input tensor.
9. The method of claim 2 further comprising configuring the hierarchical labelling function based upon map information associated with an un-normalized input tensor.
10. The method of claim 1 wherein the inferred hierarchical label includes a plurality of label parts and each label part among the plurality of label parts is encoded using one-hot encoding.
11. An apparatus for location management, the apparatus comprising:
a memory storing instructions; and
a processor caused, by the instructions, to:
receive a reference signal;
obtain, by processing the reference signal, a normalized input tensor;
obtain, through use of a hierarchical neural network location classifier, an inferred hierarchical label for a location corresponding to the normalized input tensor;
obtain a predicted location by decoding the inferred hierarchical label; and
transmit an indication of the predicted location.
12. The apparatus of claim 11 wherein the processor is further caused to obtain the predicted location by employing an inverse of a hierarchical labelling function, where the hierarchical labelling function is configured to receive, as input, a particular location and produce, as output, a particular hierarchical label to associate with the particular location.
13. The apparatus of claim 12 wherein the processor is further caused to:
receive map information for an environment for the UE; and
carry out ray-tracing on the map information to, thereby, generate a plurality of label vectors, with an individual label vector associated with each point among a plurality of points in the environment.
14. The apparatus of claim 13 wherein the processor is further caused to configure the hierarchical labelling function based upon applying a hierarchical clustering algorithm to the plurality of label vectors.
15. The apparatus of claim 13 wherein the processor is further caused to:
apply a dimensionality reduction tool to the plurality of label vectors to, thereby, generate a plurality of reduced-dimensionality label vectors; and
configure the hierarchical labelling function based upon applying a hierarchical clustering algorithm to the plurality of reduced-dimensionality label vectors.
16. The apparatus of claim 15 wherein the dimensionality reduction tool comprises an auto-encoder.
17. The apparatus of claim 12 wherein the processor is further caused to configure the hierarchical labelling function based upon global positioning system information associated with an un-normalized input tensor.
18. The apparatus of claim 12 wherein the processor is further caused to configure the hierarchical labelling function based upon vision sensing information associated with an un-normalized input tensor.
19. The apparatus of claim 12 wherein the processor is further caused to configure the hierarchical labelling function based upon map information associated with an un-normalized input tensor.
20. The apparatus of claim 12 wherein the inferred hierarchical label includes a plurality of label parts and each label part among the plurality of label parts is encoded using one-hot encoding.
21. A computer-readable medium storing instructions for location management, the instructions, when executed by a processor, causing the processor to:
receive a reference signal;
obtain, by processing the reference signal, a normalized input tensor;
obtain, through use of a hierarchical neural network location classifier, an inferred hierarchical label for a location corresponding to the normalized input tensor;
obtain a predicted location by decoding the inferred hierarchical label; and
transmit an indication of the predicted location.
22. The computer-readable medium of claim 21 wherein the instructions further cause the processor to obtain the predicted location by employing an inverse of a hierarchical labelling function, where the hierarchical labelling function is configured to receive, as input, a particular location and produce, as output, a particular hierarchical label to associate with the particular location.
23. The computer-readable medium of claim 22 wherein the instructions further cause the processor to:
receive map information for an environment for the UE; and
carry out ray-tracing on the map information to, thereby, generate a plurality of label vectors, with an individual label vector associated with each point among a plurality of points in the environment.
24. The computer-readable medium of claim 22 wherein the instructions further cause the processor to configure the hierarchical labelling function based upon map information associated with an un-normalized input tensor.
25. The computer-readable medium of claim 21 wherein the inferred hierarchical label includes a plurality of label parts and each label part among the plurality of label parts is encoded using one-hot encoding.