Utilizing machine learning models to estimate user device spatiotemporal behavior
A device may receive a first type of data identifying measurements associated with user devices and/or base stations of a mobile radio environment, and a second type of data identifying spatiotemporal behavior associated with the user devices. The device may train a first model, with the first type of data, to generate a trained first model that yields dimensionality-reduced spatiotemporal characteristics of the first type of data, and may train a second model, with the second type of data and the dimensionality-reduced spatiotemporal characteristics, to generate a trained second model. The device may receive particular data identifying measurements associated with a user device and/or base stations, and may process the particular data, with the trained first model, to generate a dimensionality-reduced spatiotemporal characteristic of the particular data. The device may process the dimensionality-reduced spatiotemporal characteristic, with the trained second model, to predict a spatiotemporal behavior of the user device.
1. A method, comprising:
receiving, by a device, a first type of data identifying measurements associated with user devices and/or base stations of a mobile radio environment,
wherein the first type of data includes one or more of:
geographical mapping data corresponding to the mobile radio environment,
obstruction data corresponding to the mobile radio environment, or
event data identifying events corresponding to the user devices;
receiving, by the device, a second type of data identifying spatiotemporal behavior associated with the user devices of the mobile radio environment;
training, by the device, a first model, with the first type of data, to generate a trained first model that yields dimensionality-reduced spatiotemporal characteristics of the first type of data;
training, by the device, a second model, with the second type of data and the dimensionality-reduced spatiotemporal characteristics, to generate a trained second model;
receiving, by the device, particular data identifying measurements associated with a user device and/or one or more base stations of the mobile radio environment;
processing, by the device, the particular data, with the trained first model, to generate a dimensionality-reduced spatiotemporal characteristic of the particular data; and
processing, by the device, the dimensionality-reduced spatiotemporal characteristic of the particular data, with the trained second model, to predict a spatiotemporal behavior of the user device.
2. The method of claim 1 , wherein the second type of data includes spatiotemporal behavior generated by other systems and including attributes to be correlated.
3. The method of claim 1 , wherein the first model and the second model are jointly trained, or
wherein the first model is separately trained first, and the second model is separately trained after the first model is trained.
4. The method of claim 1 , wherein the spatiotemporal behavior includes at least one of:
a geolocation of the user device at a particular time,
a trajectory of the user device through space and time,
a velocity of the user device at a particular time,
an acceleration of the user device at a particular time,
an orientation of the user device at a particular time,
a height of the user device at a particular time, or
a bearing of the user device at a particular time.
5. The method of claim 1 , wherein the dimensionality-reduced spatiotemporal characteristics facilitate storage and sharing of the dimensionality-reduced spatiotemporal characteristics between network devices of the mobile radio environment.
6. The method of claim 1 , wherein the first model identifies and preserves spatiotemporal patterns in the first type of data, and
wherein the first model is one of a neural network model or an auto-encoder model.
7. A device, comprising:
one or more memories; and
one or more processors, coupled to the one or more memories, configured to:
receive a first type of data identifying measurements associated with user devices and/or base stations of a mobile radio environment,
wherein the first type of data includes one or more of:
geographical mapping data corresponding to the mobile radio environment,
obstruction data corresponding to the mobile radio environment, or
event data identifying events corresponding to the user devices;
receive a second type of data identifying spatiotemporal behavior associated with the user devices of the mobile radio environment;
train a first model, with the first type of data, to generate a trained first model that yields dimensionality-reduced spatiotemporal characteristics of the first type of data;
train a second model, with the second type of data and the dimensionality-reduced spatiotemporal characteristics, to generate a trained second model;
receive particular data identifying measurements associated with a user device and/or one or more base stations of the mobile radio environment;
process the particular data, with the trained first model, to generate a dimensionality-reduced spatiotemporal characteristic of the particular data; and
process the dimensionality-reduced spatiotemporal characteristic of the particular data, with the trained second model, to predict a spatiotemporal behavior of the user device.
8. The device of claim 7 , wherein the second type of data includes spatiotemporal behavior generated by other systems and including attributes to be correlated.
9. The device of claim 7 , wherein the one or more processors, to train the first model and the second model, are configured to:
jointly train the first model and the second model, or
train the second model after the first model is trained.
10. The device of claim 7 , wherein the spatiotemporal behavior includes at least one of:
a geolocation of the user device at a particular time,
a trajectory of the user device through space and time,
a velocity of the user device at a particular time,
an acceleration of the user device at a particular time,
an orientation of the user device at a particular time,
a height of the user device at a particular time, or
a bearing of the user device at a particular time.
11. The device of claim 7 , wherein the dimensionality-reduced spatiotemporal characteristics facilitate storage and sharing of the dimensionality-reduced spatiotemporal characteristics between network devices of the mobile radio environment.
12. The device of claim 7 , wherein the first model identifies and preserves spatiotemporal patterns in the first type of data, and
wherein the first model is one of a neural network model or an auto-encoder model.
13. A non-transitory computer-readable medium storing a set of instructions, the set of instructions comprising:
one or more instructions that, when executed by one or more processors of a device, cause the device to:
receive a first type of data identifying measurements associated with user devices and/or base stations of a mobile radio environment,
wherein the first type of data includes one or more of:
geographical mapping data corresponding to the mobile radio environment,
obstruction data corresponding to the mobile radio environment, or event data identifying events corresponding to the user devices;
receive a second type of data identifying spatiotemporal behavior associated with the user devices of the mobile radio environment;
train a first model, with the first type of data, to generate a trained first model that yields dimensionality-reduced spatiotemporal characteristics of the first type of data;
train a second model, with the second type of data and the dimensionality-reduced spatiotemporal characteristics, to generate a trained second model;
receive particular data identifying measurements associated with a user device and/or one or more base stations of the mobile radio environment;
process the particular data, with the trained first model, to generate a dimensionality-reduced spatiotemporal characteristic of the particular data; and
process the dimensionality-reduced spatiotemporal characteristic of the particular data, with the trained second model, to predict a spatiotemporal behavior of the user device.
14. The non-transitory computer-readable medium of claim 13 , wherein the second type of data includes spatiotemporal behavior generated by other systems and including attributes to be correlated.
15. The non-transitory computer-readable medium of claim 13 , wherein, to cause the device to train the first model and the second model, the one or more instructions cause the device to:
jointly train the first model and the second model, or
train the second model after the first model is trained.
16. The non-transitory computer-readable medium of claim 13 , wherein the spatiotemporal behavior includes at least one of:
a geolocation of the user device at a particular time,
a trajectory of the user device through space and time,
a velocity of the user device at a particular time,
an acceleration of the user device at a particular time,
an orientation of the user device at a particular time,
a height of the user device at a particular time, or
a bearing of the user device at a particular time.
17. The non-transitory computer-readable medium of claim 13 , wherein the dimensionality-reduced spatiotemporal characteristics facilitate storage and sharing of the dimensionality-reduced spatiotemporal characteristics between network devices of the mobile radio environment.
18. The non-transitory computer-readable medium of claim 13 , wherein the one or more instructions further cause the device to:
perform one or more actions based on the spatiotemporal behavior,
wherein, to cause the device to perform the one or more actions, the one or more instructions cause the device to at least one of:
optimize handover decisions for the user device,
update one or more of the first model or the second model, or
provide data representing the spatiotemporal behavior to the base station to cause one or more parameters to be adjusted.
19. The method of claim 1 , further comprising:
performing one or more actions based on the spatiotemporal behavior,
wherein the one or more actions comprise at least one of:
optimizing handover decisions for the user device,
updating one or more of the first model or the second model, or
providing data representing the spatiotemporal behavior to the base station to cause one or more parameters to be adjusted.
20. The device of claim 7 , wherein the one or more processors are further configured to:
perform one or more actions based on the spatiotemporal behavior,
wherein the one or more processors, to perform the one or more actions, are configured to at least one of:
optimize handover decisions for the user device,
update one or more of the first model or the second model, or
provide data representing the spatiotemporal behavior to the base station to cause one or more parameters to be adjusted.