Radio resource control procedures for machine learning
An example method, apparatus, and computer-readable storage medium are provided for radio resource control (RRC) procedures for machine learning (ML). In an example implementation, the method may include receiving, by a user equipment (UE), machine learning (ML) configuration from a network node; collecting, by the user equipment (UE), machine learning (ML) data based at least on the machine learning (ML) configuration received from the network node, the machine learning (ML) data being collected from one or more layers of the user equipment (UE) in a coordinated manner; and transmitting, by the user equipment (UE), the collected machine learning (ML) data to the network node. In another example implementation, the method may include transmitting, by a network node, machine learning (ML) configuration to a user equipment (UE); and receiving, by the network node, machine learning (ML) data from the user equipment (UE), the machine learning (ML) data received in response to the machine learning (ML) configuration transmitted to the user equipment (UE).
1. A user equipment apparatus (UE) comprising:
at least one processor; and
at least one memory including instructions, when executed by the at least one processor, cause the UE to:
receive a machine learning (ML) configuration from a network node, the machine learning (ML) configuration causing the UE to collect data in a structure being used for machine learning (ML) at the network node, wherein the network node is a gNB, wherein the machine learning (ML) configuration includes a listing of cells the UE may use for configuring as secondary cells for dual connectivity and/or carrier aggregation;
collect machine learning (ML) data based at least on the machine learning (ML) configuration received from the network node, the machine learning (ML) data being collected from a packet data convergence protocol (PDCP) layer a radio link control (RLC) layer, a media access control (MAC) layer, and a physical (PHY) layer of the UE in a coordinated manner, wherein the machine learning (ML) data includes an error metric from all of the PDCP, the RLC, the MAC, and the PHY layers of the UE;
coordinate the collecting of the machine learning (ML) data at the UE, the collecting based at least on a machine learning (ML) command in the machine learning (ML) configuration, wherein the machine learning (ML) configuration includes a machine learning (ML) command that indicates the machine learning (ML) data to be collected in the coordinated manner at the user equipment (UE), and
wherein the UE coordinates the collection of machine learning (ML) data from all of the PDCP, the RLC, the MAC, and the PHY layers at the UE via direct interfaces from the RRC layer to the PDCP, the RLC, the MAC, and the PHY layers, wherein the UE coordinates machine learning (ML) data collection based at least on the machine learning (ML) command in the machine learning (ML) configuration;
receive a UECapabilityEnquiry message from the network node;
transmit a UECapability Information message to the network node, the UECapability Information message is transmitted to the network node in response to the receiving of the UECapabilityEnquiry message from the network node, wherein the machine learning (ML) configuration is received from the network node based at least on machine learning (ML) capabilities of the UE indicated to the network node in the UECapability Information message;
cause the network node to determine if raw machine learning (ML) data or averaged machine learning (ML) data is required based on the machine learning (ML) configuration;
transmit the collected machine learning (ML) data to the network node; and
cause the network node to perform an action based on the transmitted machine learning (ML) data, wherein the action includes changing values of parameters at the UE for network optimization.