US-12618930-B2 - Performing a handover procedure

US12618930B2US 12618930 B2US12618930 B2US 12618930B2US-12618930-B2

Abstract

In a method in a user equipment, UE, in a communications network, of determining whether to perform a handover procedure from a first network node to a second network node, a location of the UE is provided as input to a model stored on the UE, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on the location of the UE. A prediction of conditions on the second network node at the provided location of the UE is provided by the model. The received predicted conditions are then used to determine whether to perform a handover procedure.

Inventors

Henrik Rydén
Martin ISAKSSON
Vijaya Yajnanarayana
Sakib bin Redhwan
Roman ZHOHOV
Maksym GIRNYK
Abdulrahman Alabbasi

Assignees

TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

Dates

Publication Date: 20260505
Application Date: 20191128

Claims (17)

1 . A method in a user equipment (UE) for determining whether to perform a handover procedure from a first network node to a second network node, the method comprising: the UE obtaining the current location of the UE; the UE providing the current location of the UE as input to a model stored on the UE, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on the input current location of the UE; the UE obtaining from the model at least a first prediction of a first condition on the second network node at the provided current location of the UE; and the UE using the first prediction of the first condition to determine whether to perform a handover procedure to the second network node.
2 . The method of claim 1 , wherein the model was trained using a federated machine learning process.
3 . The method of claim 1 , wherein the step of using the received predicted conditions to determine whether to perform a handover comprises: sending a second message comprising the received predicted conditions to a node in the communications network for use by the node in determining whether the UE should perform a handover procedure to the second network node.
4 . The method of claim 3 , further comprising comparing the received predicted conditions on the second network node to a reporting condition and wherein the step of sending the second message comprises sending the second message if the received predicted conditions on the second network node satisfy the reporting condition.
5 . The method of claim 4 , wherein the reporting condition comprises one of: a threshold and wherein the reporting condition is satisfied if the received predicted conditions change from being below the threshold to above the threshold, or if the received predicted conditions change from being above the threshold to below the threshold; a threshold signal strength on the second network node and wherein the reporting condition is satisfied if the received predicted conditions indicate that the signal strength on the second network node is above the threshold signal strength; and/or a threshold load on the UE and wherein the reporting condition is satisfied if the load on the UE is above the threshold load.
6 . The method of claim 1 , further comprising performing a handover procedure to the second network node based on the received prediction of the conditions on the second network node at the provided location of the UE.
7 . The method of claim 1 , wherein the first prediction of the first condition is a prediction of a strength of a signal to be transmitted by the second network node.
8 . A method in a node of a communications network of determining whether to instruct a user equipment (UE) to perform a handover procedure from a first network node to a second network node, the method comprising: sending a first message to a UE, the first message instructing the UE to create a local copy of a model, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on the current location of the UE; receiving from the UE a second message comprising a first prediction of a first condition on the second network node at the current location of the UE, the first prediction of the first condition having been obtained using the model; and determining whether the UE should handover to the second network node, based on the received first prediction of the first condition.
9 . The method of claim 8 , further comprising sending a third message to the UE instructing the UE to handover to the second network node.
10 . The method of claim 8 , further comprising: sending a fourth message to the UE, instructing the UE to perform further training on the model according to a federated machine learning procedure.
11 . The method of claim 8 , wherein the step of sending a first message to the UE comprises sending a configuration of the model to the UE; and wherein the method further comprises: determining one or more network layers of the communications network over which to send the configuration of the model, based on a size of the model and/or a latency requirement for the model transfer.
12 . The method of claim 11 , wherein the step of determining one or more network layers comprises: determining to send the configuration to the UE using all available network layers if the model is greater than a first threshold model size and/or if the latency required for the model transfer is less than a first threshold latency.
13 . The method of claim 11 , wherein the step of determining one or more network layers comprises: determining to send the configuration to the UE using a physical, PHY, network layer if the model is less than a second threshold model size and/or if the latency required for the model transfer is less than a second threshold latency.
14 . The method of claim 11 , wherein the step of determining one or more network layers comprises: determining to send the configuration to the UE using an application network layer if the model is less than a third threshold model size and/or if the latency required for the model transfer is greater than a third threshold latency.
15 . A user equipment (UE), the UE being configured to determine whether to perform a handover procedure from a first network node to a second network, wherein the UE comprises: a memory comprising instruction data representing a set of instructions; and a processor configured to communicate with the memory and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the processor to: obtain the current location of the UE; provide the current location of the UE as input to a model stored on the UE, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on the input current location of the UE; obtain from the model at least a first prediction of a first condition on the second network node at the provided current location of the UE; and use the first prediction of the first condition to determine whether to perform a handover procedure to the second network node.
16 . A node in a communications network the node being configured to determine whether to instruct a user equipment (UE) to perform a handover procedure from a first network node to a second network node, the node comprising: a memory comprising instruction data representing a set of instructions; and a processor configured to communicate with the memory and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the processor to: send a first message to a UE, the first message instructing the UE to create a local copy of a model, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on location of the UE; receive from the UE a second message comprising a first prediction of a first condition on the second network node at the current location of the UE, the first prediction of the first condition having been obtained using the model; and determine whether the UE should handover to the second network node, based on the received first prediction of the first condition.
17 . A computer program product comprising a computer readable medium, the computer readable medium having computer readable code embodied therein, the computer readable code being configured such that, on execution by a suitable computer or processor, the computer or processor is caused to perform the method of claim 1 .

Description

CROSS REFERENCE TO RELATED APPLICATION(S) This application is a 35 U.S.C. § 371 National Stage of International Patent Application No. PCT/SE2019/051205, filed Nov. 28, 2019. TECHNICAL FIELD This disclosure relates to methods, nodes and systems in a communications network. More particularly but non-exclusively, the disclosure relates to handover procedures in a communications network. BACKGROUND Disclosures herein relate to communications networks such radio access networks. More specifically, disclosures herein relate to handover from a first node to a second node in a communications network. The number of subscribers (e.g. user equipment and devices) is ever increasing on communications networks. Furthermore, the performance requirements of said subscribers are also increasing. As such, there is a need for better coordination in communication networks to ensure good resource usage. If, for instance, certain nodes of the network are over-crowded, serving many users, while other nodes are underutilized, this can lead to unutilized network capacity since the crowded node, which has only a portion of the total network capacity, may limit its users when they could otherwise have been handed over to an underutilized node. It is thus an object of the disclosures herein to provide improved handover of a user equipment between different nodes in a communications network. SUMMARY Handover typically involves a user equipment (UE) having to perform inter-frequency measurements to determine an appropriate node to connect to. This can require the UE to reconfigure its receive chain to a different frequency carrier from that of the source cell, unless the UE can support multiple receive chains that can measure simultaneously on several frequency carriers. In the former case, the UE is not able to communicate with the source cell whilst making measurements on the new frequency carrier. Furthermore, both scenarios require increased UE battery consumption. With the deployment of mmWave driven by 5G, the number of carriers that a UE can be served by is increasing and finding the best carrier for a UE can thus require a substantial number of inter-frequency measurements to be made by the UE. One possible solution to this problem comprises using machine learning models to predict conditions on one or more possible nodes to which the UE could handover to, based on measurements of conditions associated with one or more other nodes (e.g. target carrier prediction). However, the use of target carrier prediction in this manner still requires frequent measurement of source carrier information, and in general, the target carrier predictions increase in accuracy with increased source carrier information, incentivizing measuring conditions on as many frequencies as possible. Source carrier information can comprise measurements on neighboring nodes, thus also requiring a large measurement overhead for the UE. The use of other input parameters, such as UE location, come with complications as, due to the sensitive type of location data, locations are not commonly reported to the network except for in emergency situations, or in specific commercial use cases, and can thus not be explored for building network decision functions that requires or can benefit from frequent location information such as intra/inter-freq. handover. It is an object of the embodiments herein to provide improved handover procedures that require fewer inter-frequency measurements to be made and thus consume less battery power. According to a first aspect herein there is a method in a user equipment, UE, in a communications network of determining whether to perform a handover procedure from a first network node to a second network node. The method comprises providing a location of the UE as input to a model stored on the UE, the model having been trained using a machine learning process to predict conditions on the second network node in the communications network based on the location of the UE, receiving from the model a prediction of conditions on the second network node at the provided location of the UE, and using the received predicted conditions to determine whether to perform a handover procedure to the second network node. In some embodiments the model was trained using a federated machine learning process. Predicting conditions on the second network node based on the location of the UE requires the UE to make fewer inter-frequency measurements, saving battery power whilst still providing a reliable prediction of conditions on the second network node. Furthermore, the use of a model trained using a Federated machine learning process ensures that the UE location is not transmitted across the network in order to train or use the model for the prediction. This ensures data privacy of UE location data. According to a second aspect there is a method in a node of a communications network of determining whether to instruct a user equipment, UE to perform a handover