Unsupervised location estimation and mapping based on multipath measurements
Certain aspects of the present disclosure provide methods, apparatus, and systems for predicting a location of a device in a spatial environment using a machine learning model. An example method generally includes measuring a plurality of signals received from a network entity at a device. A channel state information (CSI) measurement is generated from the measured plurality of signals. Generally, the CSI measurement includes a multipath component. Positions of one or more anchors in a spatial environment are identified based on a machine learning model trained to identify the positions of the one or more anchors based on the CSI measurement. A location of the device is estimated based on the identified positions of the one or more anchors.
1 . A processor-implemented method, comprising:
measuring a plurality of signals received from a network entity at a device;
generating a channel state information (CSI) measurement from the measured plurality of signals, the CSI measurement including a multipath component;
identifying positions of one or more anchors in a spatial environment using a machine learning model trained to identify the positions of the one or more anchors based on the CSI measurement;
estimating a location of the device based on the identified positions of the one or more anchors;
reporting, to the network entity, the estimated location of the device;
receiving, from the network entity, interference management parameters for subsequent communications with the network entity; and
communicating with the network entity based on the interference management parameters.
2 . The method of claim 1 , further comprising generating a map of the spatial environment based on the identified positions of the one or more anchors and the estimated location of the device.
3 . The method of claim 1 , further comprising selecting beamforming parameters for communications with the network entity based on the identified positions of the one or more anchors and the estimated location of the device.
4 . The method of claim 1 , wherein identifying the positions of the one or more anchors in the spatial environment comprises:
extracting, from the CSI measurement, a set of timing information, each entry in the set of timing information being associated with one of a plurality of multipath components in the CSI measurement; and
inputting the extracted set of timing information into the machine learning model to identify the positions of the one or more anchors.
5 . The method of claim 4 , further comprising identifying the location of the device in the spatial environment based on the set of timing information and a deep set model.
6 . The method of claim 4 , further comprising estimating a number of virtual anchors of the one or more anchors in the spatial environment based on the extracted set of timing information, wherein the number of the virtual anchors is initialized in a decoder.
7 . A processor-implemented method, comprising:
receiving a data set comprising channel state information (CSI) measurements;
extracting a data set of timing information from the CSI measurements; and
training a machine learning model to predict, based on the data set of the timing information:
a location of a device in a spatial environment; and
a location of each virtual anchor of one or more virtual anchors in the spatial environment;
wherein the machine learning model is configured to associate a number of timing information samples with a number of anchor positions.
8 . The method of claim 7 , further comprising deploying the trained machine learning model to a wireless communications device.
9 . The method of claim 7 , wherein training the machine learning model comprises training the machine learning model using unsupervised learning techniques.
10 . The method of claim 7 , wherein extracting the data set of the timing information comprises, for each respective CSI measurement in the data set comprising the CSI measurements, extracting one or more of time-of-flight or time-difference-of-arrival measurements from one or more multipath components of the respective CSI measurement.
11 . The method of claim 7 , wherein:
the machine learning model is implemented by an encoder-decoder neural network comprising an encoder and a decoder,
the encoder is trained to encode timing information into data representing the location of the device, and
the decoder is trained to decode the location of each virtual anchor based on the timing information associated with the location of the device.
12 . The method of claim 7 , wherein the machine learning model is trained to minimize a loss function based on a difference between actual timing information and predicted timing information for the one or more virtual anchors identified by the machine learning model.
13 . The method of claim 12 , wherein the loss function comprises a first term for timing information associated with a line-of-sight measurement and a second term for timing information associated with one or more non-line-of-sight measurements.
14 . The method of claim 12 , wherein the loss function comprises a first term for angle-of-arrival information associated with a line-of-sight measurement and a second term for angle-of-arrival information associated with one or more non-line-of-sight measurements.
15 . The method of claim 7 , wherein extracting the data set of the timing information from the CSI measurements comprises extracting the timing information based on super-resolution signal processing.
16 . A processing system, comprising:
a memory having computer-executable instructions stored thereon; and
a processor configured to execute the computer-executable instructions in order to cause the processing system to:
measure a plurality of signals received from a network entity at a device;
generate a channel state information (CSI) measurement from the measured plurality of signals, the CSI measurement including a multipath component;
identify positions of one or more anchors in a spatial environment using a machine learning model trained to identify the positions of the one or more anchors based on the CSI measurement;
estimate a location of the device based on the identified positions of the one or more anchors;
report, to the network entity, the estimated location of the device;
receive, from the network entity, interference management parameters for subsequent communications with the network entity; and
communicate with the network entity based on the interference management parameters.
17 . The processing system of claim 16 , wherein the processor is further configured to cause the processing system to generate a map of the spatial environment based on the identified positions of the one or more anchors and the estimated location of the device.
18 . The processing system of claim 16 , wherein the processor is further configured to cause the processing system to select beamforming parameters for communications with the network entity based on the identified positions of the one or more anchors and the estimated location of the device.
19 . The processing system of claim 16 , wherein in order to identify the positions of the one or more anchors in the spatial environment, the processor is configured to cause the processing system to:
extract, from the CSI measurement, a set of timing information, each entry in the set of timing information being associated with one of a plurality of multipath components in the CSI measurement; and
input the extracted set of timing information into the machine learning model to identify the positions of the one or more anchors.
20 . The processing system of claim 19 , wherein the processor is configured to cause the processing system to identify the location of the device in the spatial environment based on the set of timing information and a deep set model.
21 . The processing system of claim 19 , wherein the processor is configured to cause the processing system to estimate a number of virtual anchors of the one or more anchors in the spatial environment based on the extracted set of timing information and wherein the number of the virtual anchors is initialized in a decoder.
22 . A processing system, comprising:
a memory having executable instructions stored thereon; and
a processor configured to execute the executable instructions in order to cause the processing system to:
receive a data set comprising channel state information (CSI) measurements;
extract a data set of timing information from the CSI measurements; and
train a machine learning model to predict, based on the data set of the timing information:
a location of a device in a spatial environment; and
a location of each virtual anchor of one or more virtual anchors in the spatial environment;
wherein the machine learning model is configured to associate a number of timing information samples with a number of anchor positions.
23 . The processing system of claim 22 , wherein in order to extract the data set of the timing information, the processor is configured to cause the processing system to extract, for each respective CSI measurement in the data set comprising the CSI measurements, one or more of time-of-flight or time-difference-of-arrival measurements from one or more multipath components of the respective CSI measurement.
24 . The processing system of claim 22 , wherein:
the machine learning model is implemented by an encoder-decoder neural network comprising an encoder and a decoder,
the encoder is trained to encode timing information into data representing the location of the device, and
the decoder is trained to decode the location of each virtual anchor based on the timing information associated with the location of the device.
25 . The processing system of claim 22 , wherein the machine learning model is trained to minimize a loss function based on a difference between actual timing information and predicted timing information for the one or more virtual anchors identified by the machine learning model.
26 . The processing system of claim 22 , wherein extracting the data set of the timing information from the CSI measurements comprises extracting the timing information based on super-resolution signal processing.