Managing a wireless device that is operable to connect to a communication network
Methods are provided for managing a wireless device that is operable to connect to a communication network comprising a Radio Access Network (RAN). One method, performed by a node of the communication network, comprises: selecting, on the basis at least of information about one or more of a behavior of the wireless device and a radio environment experienced by the wireless device, a machine-learning (ML) model from a plurality of candidate ML models to be downloaded to the wireless device, the candidate models providing outputs on the basis of which respective RAN operations performed by the wireless device are configured; and causing transmission of the selected ML model to the wireless device for execution by the wireless device.
1 . A method for managing a wireless device that is operable to connect to a communication network, wherein the communication network comprises a Radio Access Network, RAN, the method, performed by a node of the communication network, wherein the node of the communication network comprises a RAN node of the communication network, a core network node of the communication network, or a server connected to the communication network, the method comprising:
selecting, by the node of the communication network, on a basis at least of information about one or more of a behavior of the wireless device with respect to the communication network and a radio environment experienced by the wireless device in the communication network, a machine-learning, ML, model from a plurality of candidate ML models to be downloaded to the wireless device, the candidate models providing outputs on the basis of which respective RAN operations performed by the wireless device are configured; and
causing, by the node of the communication network, transmission of the selected ML model to the wireless device for execution by the wireless device.
2 . The method according to claim 1 , wherein the behavior of the wireless device comprises a mobility of the wireless device.
3 . The method according to claim 2 , wherein the mobility of the wireless device is a predicted mobility of the wireless device.
4 . The method according to claim 3 , wherein the selected ML model is adapted for use locally in one or more cells falling within the predicted mobility.
5 . The method according to claim 4 , further comprising receiving, from one or more RAN nodes serving the one or more cells falling within the predicted mobility, information relating to local ML models for use in the one or more cells.
6 . The method according to claim 5 , wherein the information relating to local ML models for use in the one or more cells comprises training data specific to the one or more cells, and wherein the selected ML model is trained using the training data.
7 . The method according to claim 5 , wherein the information relating to local ML models for use in the one or more cells comprises one or more local ML models for the one or more cells, and wherein the selected ML model transmitted to the wireless device is based on the one or more local ML models.
8 . The method according to claim 2 , wherein the mobility of the wireless device comprises one or more of: a series of locations of the wireless device; and a velocity of the wireless device.
9 . The method according to claim 1 , wherein the behavior of the wireless device comprises a pattern of radio traffic between the wireless device and the communication network.
10 . The method according to claim 9 , wherein the pattern of radio traffic comprises a predicted pattern of radio traffic between the wireless device and the communication network.
11 . The method according to claim 1 , wherein the RAN operations comprise one or more of: selecting a secondary carrier for transmissions between the wireless device and the communication network; mitigating a predicted drop in radio signal quality or radio signal strength; and compressing reporting of channel state information.
12 . The method according to claim 1 , wherein the radio environment experienced by the wireless device comprises one or more of: a measured signal strength or signal quality of transmissions between the wireless device and a serving RAN node; a measured signal strength or signal quality of transmissions by one or more non-serving RAN nodes; a number of detected non-serving RAN nodes; a level of interference experienced by the wireless device as a result of transmissions by one or more interfering nodes of the communication network.
13 . A node of a communication network comprising a Radio Access Network, RAN, wherein the node is for managing a wireless device that is operable to connect to the communication network, and wherein the node comprises processing circuitry configured to cause the node to: select, on the basis at least of information about one or more of a behavior of the wireless device and a radio environment experienced by the wireless device, a machine-learning, ML, model from a plurality of candidate ML models to be downloaded to the wireless device, the candidate models providing outputs on the basis of which respective RAN operations performed by the wireless device are configured; and
cause transmission of the selected ML model to the wireless device for execution by the wireless device.
14 . A method for managing a wireless device that is operable to connect to a communication network, wherein the communication network comprises a Radio Access Network, RAN, the method, performed by a RAN node of the communication network, comprising:
receiving, from one or more further RAN nodes of the communication network, information relating to local ML models for use in cells served by the one or more further RAN nodes; and
causing transmission, to the wireless device, of a ML model for execution by the wireless device,
wherein the ML model is configured for use in the one or more cells, based on the information received from the one or more further RAN nodes, and
wherein the ML model provides an output on the basis of which a RAN operation performed by the wireless device is configured.
15 . The method according to claim 14 , wherein the information relating to local ML models for use in the one or more cells comprises training data specific to the one or more cells, and wherein the ML model transmitted to the wireless device is trained using the training data, or wherein the information relating to local ML models for use in the one or more cells comprises one or more local ML models for the one or more cells, and wherein the ML model transmitted to the wireless device is based on the one or more local ML models.
16 . The method according to claim 14 , further comprising requesting, from the one or more further RAN nodes of the communication network, information relating to local ML models for one or more RAN operations.
17 . The method according to claim 14 , wherein the cells served by one or more further RAN nodes fall on a predicted mobility path of the wireless device.
18 . A Radio Access Network, RAN, node of a communication network comprising a RAN, wherein the RAN node is for managing a wireless device that is operable to connect to the communication network, and wherein the node comprises processing circuitry configured to cause the RAN node to:
receive, from one or more further RAN nodes of the communication network, information relating to local machine-learning, ML, models for use in cells served by the one or more further RAN nodes; and
cause transmission, to the wireless device, of a ML model for execution by the wireless device,
wherein the ML model is configured for use in the one or more cells, based on the information received from the one or more further RAN nodes, and
wherein the ML model provides an output on the basis of which a RAN operation performed by the wireless device is configured.