Ensuring consistency between training and inference for multi-cell

Abstract

Systems, methods, apparatuses, and computer program products for ensuring consistency between training and inference for multi-cell. A method can include receiving a plurality of identifiers configured for mobility data collection, each identifier of the plurality of identifiers indicating at least one DL- RS of a corresponding candidate cell, the plurality of identifiers including a first identifier identifying at least one DL- RS of a first candidate cell and a second identifier identifying at least one DL- RS of a second candidate cell, wherein a number of the plurality of identifiers is equal to a number of DL- RS resource sets; performing at least one measurement of the DL- RS indicated by each identifier of the plurality of identifiers; ranking a subset of the plurality of identifiers according to parameters of the DL- RS measurements using a set of configuration rules; and training at least one machine learning model utilizing the ranked subset of the plurality of identifiers.

Description

Technical Field

[0001] Some example embodiments may generally relate to mobile or wireless telecommunications systems, such as 3GPP Long Term Evolution (LTE), 5G Radio Access Technology (RAT), New Radio (NR) Access Technology, 6G, and / or other communication systems. For example, some example embodiments may relate to systems and / or methods for ensuring consistency between training and inference using multiple cells. Background Technology

[0002] Examples of mobile or wireless telecommunications systems can include radio frequency (RF) 5G RAT, Universal Mobile Telecommunications System (UMTS) Terrestrial Radio Access Network (UTRAN), Evolved LTE UTRAN (E-UTRAN), LTE Advanced (LTE-A), LTE-A Pro, NR access technologies and / or the MulteFire Alliance. 5G radio systems refer to next-generation (NG) radio systems and network architectures. 5G systems are typically built on 5G NR, but 5G (or NG) networks can also be built on E-UTRA radio. NR is expected to support service categories such as enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine-type communications (mMTC). NR is expected to deliver ultra-wideband, ultra-robust, low-latency connectivity and massive networking to support the Internet of Things (IoT). Next-generation radio access network (NG-RAN) refers to the radio access network (RAN) for 5G, which can provide radio access for NR, LTE, and LTE-A. It is important to note that nodes that provide radio access functionality to user equipment in 5G (e.g., similar to Node B in UTRAN or Evolved Node B (eNB) in LTE) can be called Next Generation Node B (gNB) when built on NR radio, and can be called Next Generation eNB (NG-eNB) when built on E-UTRA radio. Summary of the Invention

[0003] According to some example embodiments, a method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters measured for the downlink reference signal using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0004] According to some example embodiments, an apparatus may include components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell. The apparatus may further include components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include components for ranking a subset of the plurality of identifiers based on parameters measured from the downlink reference signal using a configured rule set. The apparatus may further include components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0005] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0006] According to some example embodiments, a computer program product can perform a method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0007] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus to receive at least a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to rank a subset of the plurality of identifiers based on parameters measured for the downlink reference signals, using at least a set of configured rules. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to train at least one machine learning model with at least the ranked subset of the plurality of identifiers.

[0008] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink of a first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell. The apparatus may further include a ranking circuitry configured to perform ranking of a subset of the plurality of identifiers based on parameters measured from the downlink reference signal using a set of configuration rules. The apparatus may further include an execution circuitry configured to perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include a training circuitry configured to perform training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0009] According to some example embodiments, a method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, and the subset of the plurality of identifiers is ranked based on parameters measured for the downlink reference signal. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0010] According to some example embodiments, an apparatus may include components for receiving a mobility data collection configuration comprising a plurality of identifiers. The apparatus may also include components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of identifiers is ranked based on parameters measured from the downlink reference signal. The apparatus may further include components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers.

[0011] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured from the downlink reference signal. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers.

[0012] According to some example embodiments, a computer program product can perform a method. The method may include receiving a mobility data collection configuration including a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured from the downlink reference signal. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers.

[0013] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus to receive at least a mobility data collection configuration including a plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to select at least a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to use at least the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0014] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform a mobility data collection configuration including a plurality of identifiers. The apparatus may also include a selection circuitry configured to perform a selection of a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked based on parameters measured from the downlink reference signal. The apparatus may further include an execution circuitry configured to perform at least one inference using the machine learning model to perform at least one measurement on the plurality of identifiers and the plurality of cells.

[0015] According to some example embodiments, a method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may further include receiving the first identifier from the first candidate cell. The method may also include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0016] According to some example embodiments, an apparatus may include components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The apparatus may also include components for receiving the first identifier from the first candidate cell. The apparatus may further include components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0017] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may also include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0018] According to some example embodiments, a computer program product can perform a method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may also include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0019] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus to receive at least a mobility data collection configuration including a plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to select at least a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to use at least the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0020] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform a mobility data collection configuration including a plurality of identifiers. The apparatus may also include a selection circuitry configured to perform a selection of a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked based on parameters measured from the downlink reference signal. The apparatus may further include an execution circuitry configured to perform at least one inference using the machine learning model to perform at least one measurement on the plurality of identifiers and the plurality of cells.

[0021] According to some example embodiments, a method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0022] According to some example embodiments, an apparatus may include components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The apparatus may further include components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include components for ranking a subset of the plurality of identifiers based on parameters measured for the downlink reference signals using a configured rule set. The apparatus may further include components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0023] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0024] According to some example embodiments, a computer program product can perform a method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0025] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus to receive at least a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to perform at least one measurement on at least the downlink reference signal indicated by each of the plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to rank a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements, using at least a set of configured rules. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to train at least one machine learning model with at least the ranked subset of the plurality of identifiers.

[0026] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The apparatus may further include an execution circuitry configured to perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include a ranking circuitry configured to perform ranking of a subset of the plurality of identifiers based on parameters measured from the downlink reference signal using a set of configuration rules. The apparatus may further include a training circuitry configured to perform training of at least one machine learning model using the ranked subset of the plurality of identifiers.

[0027] According to some example embodiments, a method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0028] According to some example embodiments, an apparatus may include components for receiving a mobility data collection configuration comprising a plurality of identifiers. The apparatus may also include components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The apparatus may further include components for performing at least one inference using the machine learning model on at least one measurement of the plurality of identifiers and the plurality of cells.

[0029] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0030] According to some example embodiments, a computer program product can perform a method. The method may include receiving a mobility data collection configuration including a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0031] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus to receive at least a mobility data collection configuration including a plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to select at least a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to use at least the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0032] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform a mobility data collection configuration including a plurality of identifiers. The apparatus may also include a selection circuitry configured to perform a selection of a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. The apparatus may further include an execution circuitry configured to perform at least one inference using the machine learning model to perform at least one measurement on the plurality of identifiers and the plurality of cells.

[0033] According to some example embodiments, a method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may further include receiving the first identifier from the first candidate cell. The method may also include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0034] According to some example embodiments, an apparatus may include components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The apparatus may also include components for receiving the first identifier from the first candidate cell. The apparatus may further include components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0035] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may further include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0036] According to some example embodiments, a computer program product can perform a method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may also include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0037] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus to send a request for a first identifier configured for mobility data collection to at least a first candidate cell, wherein the first identifier identifies the first candidate cell. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to receive the first identifier from at least the first candidate cell. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to send at least a plurality of identifiers, each including the first identifier, to a user equipment, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0038] According to various example embodiments, an apparatus may include a transmitting circuitry configured to perform a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The apparatus may also include a receiving circuitry configured to perform receiving the first identifier from the first candidate cell. The apparatus may further include a transmitting circuitry configured to perform transmitting to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0039] According to some example embodiments, a method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0040] According to some example embodiments, an apparatus may include components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The apparatus may further include components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include components for ranking a subset of the plurality of identifiers based on parameters measured for the downlink reference signals using a configured rule set. The apparatus may further include components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0041] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0042] According to some example embodiments, a computer program product can perform a method. The method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The method may further include ranking a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements using a configured rule set. The method may further include training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0043] According to some example embodiments, an apparatus may include at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus to receive at least a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to perform at least one measurement on at least the downlink reference signal indicated by each of the plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to rank a subset of the plurality of identifiers based on parameters of the downlink reference signal measurements, using at least a set of configured rules. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to train at least one machine learning model with at least the ranked subset of the plurality of identifiers.

[0044] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The apparatus may further include an execution circuitry configured to perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers. The apparatus may further include a ranking circuitry configured to perform ranking of a subset of the plurality of identifiers based on parameters measured from the downlink reference signal using a set of configuration rules. The apparatus may further include a training circuitry configured to perform training of at least one machine learning model using the ranked subset of the plurality of identifiers.

[0045] According to some example embodiments, a method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0046] According to some example embodiments, an apparatus may include components for receiving a mobility data collection configuration comprising a plurality of identifiers. The apparatus may also include components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The apparatus may further include components for performing at least one inference using the machine learning model on at least one measurement of the plurality of identifiers and the plurality of cells.

[0047] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0048] According to some example embodiments, a computer program product can perform a method. The method may include receiving a mobility data collection configuration that includes a plurality of identifiers. The method may further include selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The method may further include using the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0049] According to certain example embodiments, an apparatus may include at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus to receive at least a mobility data collection configuration including a plurality of identifiers. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to select at least a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to use at least the machine learning model to perform at least one inference of the plurality of identifiers and at least one measurement of the plurality of cells.

[0050] According to various example embodiments, an apparatus may include a receiving circuitry configured to perform a mobility data collection configuration including a plurality of identifiers. The apparatus may also include a selection circuitry configured to perform a selection of a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. The apparatus may further include an execution circuitry configured to perform at least one inference using the machine learning model to perform at least one measurement on the plurality of identifiers and the plurality of cells.

[0051] According to some example embodiments, a method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may further include receiving the first identifier from the first candidate cell. The method may also include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for the first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0052] According to some example embodiments, an apparatus may include components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The apparatus may also include components for receiving the first identifier from the first candidate cell. The method may further include components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0053] According to various example embodiments, a non-transitory computer-readable medium may include program instructions that, when executed by a device, cause the device to perform at least one method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may also include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0054] According to some example embodiments, a computer program product can perform a method. The method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The method may also include receiving the first identifier from the first candidate cell. The method may further include sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0055] According to certain example embodiments, an apparatus may include at least one processor and at least one memory storing instructions, which, when executed by the at least one processor, cause the apparatus to send a request for a first identifier configured for mobility data collection to at least a first candidate cell, wherein the first identifier identifies the first candidate cell. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to receive the first identifier from at least the first candidate cell. The at least one memory and the instructions, when executed by the at least one processor, may also cause the apparatus to send at least a plurality of identifiers, each including the first identifier, to a user equipment, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0056] According to various example embodiments, an apparatus may include a transmitting circuitry configured to perform a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell. The apparatus may also include a receiving circuitry configured to perform a receiving circuitry to receive the first identifier from the first candidate cell. The apparatus may further include a transmitting circuitry configured to perform a transmitting circuitry to a user equipment that includes at least a plurality of identifiers comprising the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to a candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets. Attached Figure Description

[0057] To properly understand the exemplary embodiments, reference should be made to the accompanying drawings, in which:

[0058] Figure 1 An example of a measurement model is illustrated;

[0059] Figure 2 The illustration shows an example of ranking associated identifiers;

[0060] Figures 3A to 3B An example of a signaling diagram according to certain embodiments is illustrated;

[0061] Figure 4 An example flowchart illustrating a method according to some example embodiments is shown;

[0062] Figure 5 Examples of flowcharts illustrating methods according to various exemplary embodiments are shown;

[0063] Figure 6 An example flowchart illustrating a method according to certain example embodiments is shown;

[0064] Figure 7 An example flowchart illustrating a method according to some example embodiments is shown;

[0065] Figure 8 Examples of flowcharts illustrating methods according to various exemplary embodiments are shown;

[0066] Figure 9 An example flowchart illustrating a method according to certain example embodiments is shown;

[0067] Figure 10 Examples of flowcharts illustrating methods according to various exemplary embodiments are shown;

[0068] Figure 11 Examples of flowcharts illustrating methods according to various exemplary embodiments are shown;

[0069] Figure 12 Examples of flowcharts illustrating methods according to various exemplary embodiments are shown;

[0070] Figure 13 Examples of various network devices according to some exemplary embodiments are illustrated; and

[0071] Figure 14 The illustration shows examples of 5G network and system architectures according to certain example embodiments. Detailed Implementation

[0072] It will be readily understood that, as generally described and illustrated in the accompanying drawings, components of some example embodiments can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of some example embodiments of systems, methods, apparatuses, and computer program products for ensuring consistency between training and inference using multiple cells is not intended to limit the scope of certain example embodiments, but rather represents selected example embodiments.

[0073] 3GPP Rel-19 includes beam management (BM) based on artificial intelligence (AI) / machine learning (ML), comprising both spatial beam prediction (i.e., BM-Case 1) and temporal beam prediction (i.e., BM-Case 2). For example, spatial beam prediction (i.e., BM-Case 1) can predict the optimal transmit (Tx) / receive (Rx) beam for different spatial locations, while temporal beam prediction (i.e., BM-Case 2) can predict the beam most likely to be used for future moments (e.g., beam prediction in the spatial domain (i.e., BM-Case 1)).

[0074] 3GPP may include a general AI / ML framework for single-side AI / ML models. The signaling and protocol aspects of Lifecycle Management (LCM) may enable (if appropriate) selection, activation, deactivation, switching, and rollback of functionalities and models, including signaling related to identification. Furthermore, the general AI / ML framework may include any signaling / (multiple) mechanisms necessary for LCM to facilitate model training, inference, performance monitoring, and / or data collection for both UE-side and NW-side models. Additionally, the general AI / ML framework may include signaling mechanisms applicable to functionalities / models.

[0075] The BM can include downlink (DL) Tx beam prediction for both the UE-side model and the NW-side model, including: spatial DL Tx beam prediction for beam set A based on measurements from beam set B (i.e., BM-Case 1), and temporal DL Tx beam prediction for beam set A based on historical measurements from beam set B (i.e., BM-Case 2). The BM can also specify necessary signaling / (multiple) mechanisms to facilitate LCM operation (if any) specific to the BM use case. This also ensures consistency between training and inference regarding additional NW-side conditions (if identified) for inference by the UE.

[0076] 3GPP can also use association identifiers (IDs) (e.g., 3GPP association IDs) for consistency of NW-side additional conditions across training and inference for UE-side models of BM-Case 1 and BM-Case 2. For example, in cases where NW-side additional conditions may at least affect UE assumptions about beams in set A / set B, such consistency can be achieved using an association ID that indicates what can be assumed across training and inference UEs using the same association ID. Association IDs can be used in the UE-side model within the BM.

[0077] For multiple UE-side models developed at the UE (e.g., trained, updated), in order to collect data, the network can signal multiple data collection-related configurations and any multiple association IDs. Association IDs can be associated with additional conditions on the NW side.

[0078] 3GPP Rel-19 AI / ML includes several case studies for predicting Radio Resource Management (RRM) measurements. In Case 1, multiple Layer 1 (L1) beam-level measurements can be predicted based on (multiple) actual L1 beam-level measurements, and then Layer 3 (L3) cell-level measurements can be generated. In Case 2, multiple L3 cell-level measurements can be predicted based on (multiple) actual L3 cell-level measurements. In Case 3, multiple L3 cell-level measurements can be predicted based on (multiple) actual L1 beam-level measurements. (The text continues with further examples.) Figure 1 As illustrated, filtering can occur at two distinct levels: at the physical layer (i.e., L1) to derive beam quality, and then at the Radio Resource Control (RRC) level (i.e., L3) to derive cell quality from multiple beams. In Cases 1 and 3, beam-level measurements (i.e., L1 Reference Signal Received Power (RSRP)) of the Synchronization Signal Block (SSB) beamset in Frequency Range (FR) 1 can be input into the AI / ML model, while for Case 2, the L3-RSRP of the SSB beams can be used as input for the AI / ML model.

[0079] For Case 1, the input to the AI / ML model can be a set of L1-RSRPs from SSB beam measurements, which may include spatial beam measurement reduction and / or historical beam measurement reduction (e.g., using spatial L1-RSRP values ​​or historical L1-RSRP values ​​for 32 of the 64 SSB beams). The output of the prediction method (e.g., random forest) may include the top K L1-RSRPs of the SSB beams. The L3-RSRP can then be derived based on the predicted L1-RSRPs of the SSB beams. In Case 2, the input to the AI / ML model can be the L3-RSRPs of the beam set to predict the L3-RSRPs of the top K beams. In Case 3, the input to the AI / ML model can be a set of L1-RSRPs from SSB beam measurements, which may include spatial beam measurement reduction / historical beam measurement reduction (e.g., using spatial L1-RSRP values ​​or historical L1-RSRP values ​​for the smallest SSB beam set out of the 64 SSB beams). The output of a prediction method (e.g., random forest) can include the first K L3-RSRPs of the SSB beam.

[0080] For UEs that support ML models to generate cell-level results (for multiple cells) based on predicted beamforming results (from multiple cells), the cell-level results can select the optimal cell for mobility purposes. AI / ML models can predict lost beamforming results (including the optimal beam) based on measurements of a subset of reference signal (RS) resources. As discussed above, in Case 1, multiple L1 beamforming measurements can be predicted based on (multiple) actual L1 beamforming measurements, and then L3 cell-level measurements can be generated.

[0081] In the UE-side model for Case 1, the UE can be configured to measure a subset of SSB resources from candidate cells (e.g., cells 1 to 5), and the UE can predict L1 beam-level measurements (e.g., the first K predicted beams for each candidate cell) for the entire SSB resource set of the candidate cells (e.g., cells 1 to 5). Based on the predicted L1 beam-level measurements, the UE can generate cell-level results (for multiple cells) based on the predicted beam results (from multiple cells), where the cell-level results can be selected as the best cell for mobility purposes. The AI / ML model can predict lost beam results (including the best beam) based on the measurements of the RS resource subset.

[0082] If the model used by the UE is associated with the NW-side appendage conditions of the cell (e.g., cell 1 to cell 5), the model may perform poorly when the NW-side appendage conditions used by the cell change over time (different NW-side appendages in the training and inference phases) because the model training and inference phases are inconsistent with the NW-side appendage conditions.

[0083] Association IDs can be indicators provided during the training data collection and inference phases to ensure consistency between training and inference. However, these association IDs may be cell-specific and / or may only be used for beam prediction within a cell. Furthermore, for AI / ML applications focused on mobility, association IDs may not be reused to ensure consistency between training and inference.

[0084] Some of the example embodiments described herein may have various benefits and / or advantages to overcome the disadvantages described above. For example, some example embodiments may improve the consistency between ML training and inference. Therefore, some example embodiments discussed below relate to improvements in computer-related techniques.

[0085] Figures 3A to 3B An example of signaling diagram 300 is illustrated, depicting techniques for ensuring consistency between training and inference using multiple cells. According to certain example embodiments, such as... Figure 13 As illustrated, NE 330 to NE 350 can be similar to NE 1310, and UE 320 can be similar to UE 1320. As an example, NE 330 can be the serving cell, while NE 340 and NE 350 can be neighboring cells. Although only one serving cell (i.e., NE 330) and two neighboring cells (i.e., NE 340 and NE 350) are shown, additional candidate cells and / or serving cells can be used to define a sequence of associated IDs.

[0086] Additionally, UE 320 can support AI / ML-based mobility via a UE-side model, where UE 320 is configured to measure (e.g., L1-RRP or L3-RSRP) all or a subset of DL RSs from multiple candidate cells, and apply AI / ML models to predict cell-level or beam-level measurements to facilitate mobility decisions. DL RSs (e.g., SSBs) associated with multiple candidate cells can be configured by UE 320 to enable data collection.

[0087] In operation 301, NE 330 and neighboring cell 340 can exchange measurement coordination configurations on association IDs. For example, NE 330 can send a request to neighboring cell 340 for an association ID for the DL RS TX applicable to the candidate cell, and NE 330 can receive a response with an association ID for the DL RS Tx applicable to the candidate cell. Such requests and responses may occur when communication is made regarding the DL RS configuration of the candidate cell for mobility purposes.

[0088] Similar to operation 301, in operation 302, NE 330 and neighboring cell 350 can exchange measurement coordination configurations on association IDs. For example, NE 330 can send a request to neighboring cell 350 for an association ID applicable to the candidate cell's DL RS Tx, and NE 330 can receive a response with an association ID applicable to the candidate cell's DL RS Tx. This request and response may occur when communication is made regarding the DL RS configuration of the candidate cell for mobility purposes.

[0089] In various example embodiments, after data is determined at NE 330 and after UE 320 collects and reports that data, NE 330 can assign the order of association IDs to the dataset collected at NE 330, instead of sending the association ID sequence along with the data collection configuration. The UE can then download the dataset and the association ID order to train its model.

[0090] In operation 303, NE 330 may send a mobility data collection configuration, including the associated ID order, to UE 320 (e.g., in...). SSB-ConfigMobility (in Chinese). The UE 320 can be configured with a sequence of associated IDs that are associated with multiple DL RS (e.g., SSB) resource sets corresponding to multiple candidate cells in the data collection configuration, wherein each DL RS resource set can be associated with a candidate cell.

[0091] In various example embodiments, the associated ID can be found in information elements (IEs) (e.g., MeasObjectNR Configured in SSB-ConfigMobility Part of a field. (IE) MeasObjectNR Information applicable to intra-frequency / inter-frequency measurements and / or intra-frequency / inter-frequency measurements of Synchronization Signals (SS) / Physical Block Channels (PBCH) blocks (multiple) can be specified. IE SSB-ConfigMobility It can be configured as: The AssociateIDs-mobility-Rel20 ENUMERATED {n1,n2,…nM} structure represents the associated IDs of neighboring cell 1, neighboring cell 2,…, and neighboring cell M, respectively. The SSB-associatedIDs-mob-Rel20 ENUMERATED {m1,m2,…mM} structure represents the number of SSB beams corresponding to the associated IDs n1, n2,…nM, respectively.

[0092] In some example embodiments, the size of the sequence can be equal to the number of multiple DL RS resource sets, wherein the associated IDs in the sequence are also mapped to each DL RS resource set or each candidate cell. Therefore, in the IE discussed above... MeasObjectNR In the example, AssociateIDs-mob-Rel20 can be equal to SSB-associatedIDs-mob-Rel20.

[0093] In some example embodiments, the size of the sequence may differ from the number of multiple DL RS resource sets. For example, the mobility data collection configuration received in operation 303 may include an instruction to UE 320 on how UE 320 should map the associated IDs in the sequence to each DL RS resource set or each candidate cell. Therefore, in the IE discussed above MeasObjectNR In the example, AssociateIDs-mob-Rel20 may not be equal to SSB-associatedIDs-mob-Rel20.

[0094] In operation 304, when UE 320 determines the sequence of associated IDs (e.g., using a list of SSB resource sets associated with multiple candidate cells), if the associated IDs are provided as (Associated ID #1, Associated ID #3, Associated ID #6, Associated ID #8, Associated ID #2) for 5 SSB resource sets (from 5 candidate cells), UE 320 can consider this as an ordering of the associated IDs. As an example, UE 320 can determine the order of associated IDs used for data collection (e.g., for a list of SSB resource sets associated with multiple candidate cells); the order of associated IDs used for data collection may have already been included in the mobility data collection configuration received in operation 303. For example, when defining the order, from the lowest cell ID to the highest cell ID can be considered.

[0095] In operation 305, neighboring cell 340 can send DL RS Tx to UE 320. Similarly, in operation 306, neighboring cell 350 can send DL RS Tx to UE 320. Furthermore, in operation 307, NE 330 can send DLRS Tx to UE 330.

[0096] In operation 308, UE 320 may collect data based on DL RS measurements from (multiple) candidate cells. The collected data samples may be associated with the order of the associated IDs. Therefore, the dataset collected by UE 320 may be associated with the order of the associated IDs.

[0097] In various example embodiments, assuming there are three cells, the UE 320 can use association IDs (#3, #5, #8) as the association ID sequence for cells (#1, #2, #3). When SSB measurements from cells #1, #2, and #3 are considered sequentially in the model input, the UE dataset considered for the model input can be linked to the order of association IDs #3, #5, and #8.

[0098] In operation 309, UE 320 can train at least one ML model for the collected dataset, wherein the order of the associated IDs can be stored as metadata.

[0099] Figure 2 An example of a training model (model-X) is depicted, which can be linked sequentially with associated IDs (#N1, #N2), and the order of cell IDs (cell #1 and cell #2) is shown. For example, when UE 320 receives a sequence of associated IDs (e.g., #N2, #N1 for two SSB resource sets), UE 320 can consider the order of associated IDs from the lowest cell ID to the highest cell ID.

[0100] As Figure 2 In another example, UE 320 can be configured with an associated ID sequence that is associated with DL RS resource sets corresponding to candidate cell #1 and candidate cell #2. UE 320 can collect datasets based on the order of the associated IDs, where UE 320 receives associated IDs (e.g., #N1, #N2) as an associated ID sequence for cell #1 and cell #2. UE 320 can input SSB measurements from cell #1 and cell #2 into the ML model, thereby providing UE 320 with a mapping to the order of associated IDs #N1 and #N2.

[0101] Beginning with operation 310, some example embodiments may include an inference phase, wherein DL RS (e.g., SSB) associated with multiple candidate cells can be sent to UE 320 to enable the inference operation. For example, operations 310 to 313 may be similar to operations 301 to 304. In operations 310 to 311, when association IDs received from multiple candidate cells matching the association ID sequence are sent to UE 320, NE 330 may perform the inference operation as follows.

[0102] In some example embodiments, NE 330 may send a request to neighboring cell 340 and / or neighboring cell 350 to use the association ID for the candidate cell's DL RS Tx, and NE 330 may receive an indication from neighboring cell 340 and / or neighboring cell 350 as to whether it is feasible. NE 330 may determine the likelihood of supporting one or more known orders of association IDs (from operation 301 and / or operation 302) via multiple candidate cells.

[0103] When multiple candidate cells confirm that they can support the associated ID sequence for UE 320 in operation 301, UE 320 may consider performing an inference operation in operation 318, as discussed below.

[0104] In operation 314, UE 320 may be configured with an associated ID sequence that is associated with multiple DL RS (e.g., SSB) resource sets corresponding to multiple candidate cells in the inference configuration, wherein each DL RS resource set may be associated with a candidate cell. UE 320 may select an ML model based at least on the order of the associated IDs provided in the inference configuration.

[0105] Operations 315 to 317 can be similar to operations 305 to 307 discussed above.

[0106] In operation 318, UE 320 can perform inference operations based on the selected ML model, where it is assumed that the training phase and the inference phase are consistent with respect to beam properties, power level properties, and other additional conditions on the NW side.

[0107] In some example implementations, SSBs associated with multiple candidate cells can be configured in the order of their association IDs. UE 320 can assume similar properties (for beam ordering, beam size, and other details on the physical beam) for beams defined in the order of their association IDs.

[0108] Figure 4 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13 An example of a flowchart of method 400 performed by the illustrated UE 1320.

[0109] In step 401, the method may include receiving a plurality of identifiers configured for mobility data collection. Each of the plurality of identifiers indicates at least one downlink reference signal for a corresponding candidate cell. The plurality of identifiers includes a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell and the second identifier identifying at least one downlink reference signal for a second candidate cell.

[0110] In step 402, the method may further include performing at least one measurement on a downlink reference signal indicated by each of a plurality of identifiers.

[0111] In step 403, the method may further include using a configuration rule set to rank a subset of multiple identifiers based on parameters measured by the downlink reference signal.

[0112] In step 404, the method may further include training at least one machine learning model using a ranked subset of multiple identifiers.

[0113] In some example embodiments, at least one measurement performed may include the Channel State Information Reference Signal (CSI-RS), the Synchronization Signal Reference Signal Received Power (SS-RSRP), and the Synchronization Signal Received Power (SSB-RP).

[0114] In some example implementations, multiple identifiers may be associated with the serving cell configuration.

[0115] In various example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0116] Figure 5 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13 An example of a flowchart of method 500 performed by the illustrated UE 1320.

[0117] In step 501, the method may include receiving a mobility data collection configuration, which includes a plurality of identifiers.

[0118] In step 502, the method may further include selecting a machine learning model trained on a ranked subset of a plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for the corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured from the downlink reference signal.

[0119] In step 503, the method may further include using a machine learning model to perform at least one inference on at least one of the following: multiple identifiers and at least one measurement of multiple cells.

[0120] In some example embodiments, the method may further include determining the order of identifiers for a list of Synchronization Signal / Physical Broadcast Channel Block (SSB) resource sets associated with multiple candidate cells.

[0121] In some example embodiments, the method may also include receiving downlink reference signal transmissions.

[0122] In various example implementations, multiple identifiers may be associated with the serving cell configuration.

[0123] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0124] Figure 6 The illustrations depict various example embodiments that can be generated by an NE (such as...) Figure 13 An example of a flowchart of method 600 performed by NE 1310 is shown.

[0125] In step 601, the method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell.

[0126] In step 602, the method may further include receiving a first identifier from a first candidate cell.

[0127] In step 603, the method may further include sending to the user equipment a plurality of identifiers including at least a first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell, and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0128] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0129] In some example embodiments, the method may further include sending downlink reference signal transmissions to the user equipment.

[0130] Figure 7 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13 An example of a flowchart of method 700 performed by the illustrated UE 1320.

[0131] In step 701, the method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0132] In step 702, the method may further include performing at least one measurement on a downlink reference signal indicated by each of a plurality of identifiers.

[0133] In step 703, the method may further include using a configuration rule set to rank a subset of multiple identifiers based on parameters measured by the downlink reference signal.

[0134] In step 704, the method may further include training at least one machine learning model using a ranked subset of multiple identifiers.

[0135] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0136] In some example implementations, multiple identifiers may be associated with the serving cell configuration.

[0137] In various example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0138] Figure 8 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13 An example of a flowchart of method 800 performed by the illustrated UE 1320.

[0139] In step 801, the method may include receiving a mobility data collection configuration, which includes a plurality of identifiers.

[0140] In step 802, the method may further include selecting a machine learning model trained on a ranked subset of a plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0141] In step 803, the method may further include using a machine learning model to perform at least one inference on at least one of the following: multiple identifiers and at least one measurement of multiple cells.

[0142] In some example embodiments, the method may further include determining the order of identifiers for a list of Synchronization Signal / Physical Broadcast Channel Block (SSB) resource sets associated with multiple candidate cells.

[0143] In some example embodiments, the method may also include receiving downlink reference signal transmissions.

[0144] In various example implementations, multiple identifiers may be associated with the serving cell configuration.

[0145] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0146] Figure 9 The illustrations depict various example embodiments that can be generated by an NE (such as...) Figure 13 An example of a flowchart of method 900 performed by NE 1310 is shown.

[0147] In step 901, the method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell.

[0148] In step 902, the method may further include receiving a first identifier from a first candidate cell.

[0149] In step 903, the method may further include sending to the user equipment a plurality of identifiers including at least a first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0150] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0151] In some example embodiments, the method may further include sending downlink reference signal transmissions to the user equipment.

[0152] Figure 10 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13An example of a flowchart of method 1000 performed by the illustrated UE 1320.

[0153] In step 1001, the method may include receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0154] In step 1002, the method may further include performing at least one measurement on a downlink reference signal indicated by each of a plurality of identifiers.

[0155] In step 1003, the method may further include using a configuration rule set to rank a subset of multiple identifiers based on parameters measured by the downlink reference signal.

[0156] In step 1004, the method may further include training at least one machine learning model using a ranked subset of multiple identifiers.

[0157] In some example embodiments, at least one measurement performed may include the Channel State Information Reference Signal (CSI-RS), the Synchronization Signal Reference Signal Received Power (SS-RSRP), and the Synchronization Signal Received Power (SSB-RP).

[0158] In some example implementations, multiple identifiers may be associated with the serving cell configuration.

[0159] In various example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0160] Figure 11 The illustrations depict various example embodiments that can be generated by a UE (such as...) Figure 13 An example of a flowchart of method 1100 performed by the illustrated UE 1320.

[0161] In step 1101, the method may include receiving a mobility data collection configuration, which includes a plurality of identifiers.

[0162] In step 1102, the method may further include selecting a machine learning model trained on a ranked subset of a plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signals, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0163] In step 1103, the method may further include using a machine learning model to perform at least one inference on at least one of the following: multiple identifiers and at least one measurement of multiple cells.

[0164] In some example embodiments, the method may further include determining the order of identifiers for a list of Synchronization Signal / Physical Broadcast Channel Block (SSB) resource sets associated with multiple candidate cells.

[0165] In some example embodiments, the method may also include receiving downlink reference signal transmissions.

[0166] In various example implementations, multiple identifiers may be associated with the serving cell configuration.

[0167] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0168] Figure 12 The illustrations depict various example embodiments that can be generated by an NE (such as...) Figure 13 An example of a flowchart of method 1200 performed by NE 1310 (illustrated).

[0169] In step 1201, the method may include sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell.

[0170] In step 1202, the method may further include receiving a first identifier from a first candidate cell.

[0171] In step 1203, the method may further include sending to the user equipment a plurality of identifiers including at least a first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0172] In some example embodiments, at least one downlink reference signal may include at least one synchronization signal / physical broadcast channel block (SSB).

[0173] In some example embodiments, the method may further include sending downlink reference signal transmissions to the user equipment.

[0174] Figure 13 An example of a system according to certain example embodiments is illustrated. In one example embodiment, the system may include multiple devices, such as, for example, NE 1310 and / or UE 1320.

[0175] NE 1310 can be one or more of the following, or a combination thereof: a base station (e.g., a 3G UMTS NodeB, a 4G LTE evolved NodeB, or a 5G NR next-generation NodeB), a serving gateway, a server, and / or any other access node.

[0176] NE 1310 may also include at least one gNB centralized unit (CU), which may be associated with at least one gNB distributed unit (DU). At least one gNB-CU and at least one gNB-DU may be connected via at least one F1 interface and at least one X... n -C interface, and / or communicate via at least one NG interface of the fifth generation core (5GC).

[0177] UE 1320 may include one or more mobile devices, such as mobile phones, smartphones, personal digital assistants (PDAs), tablet computers, or portable media players, digital cameras, pocket cameras, video game consoles, navigation units (such as GPS devices), desktop or laptop computers, single-location devices (such as sensors or smart meters), or any combination thereof. Furthermore, NE 1310 and / or UE 1320 may be one or more Citizen Broadband Radio Service (CBSD) devices.

[0178] NE 1310 and / or UE 1320 may include at least one processor, designated as 1311 and 1321, respectively. Processor 1311 and processor 1321 may be embodied by any computing or data processing device, such as a central processing unit (CPU), application-specific integrated circuit (ASIC), or similar device. The processor may be implemented as a single controller or multiple controllers or processors.

[0179] At least one memory may be provided in one or more devices, as indicated by 1312 and 1322. The memory may be fixed or removable. The memory may include computer program instructions or computer code contained therein. Memory 1312 and memory 1322 may independently be any suitable storage device, such as a non-transitory computer-readable medium. The term "non-transitory" as used herein may correspond to a limitation on the medium itself (i.e., tangible, not tactile) rather than a limitation on the persistence of data storage (e.g., random access memory (RAM) versus read-only memory (ROM)). Hard disk drives (HDDs), random access memory (RAM), flash memory, or other suitable memories may be used. The memory may be integrated as a processor on a single integrated circuit or may be separate from one or more processors. Furthermore, the computer program instructions stored in the memory and processed by the processor may be any suitable form of computer program code, such as a compiled or interpreted computer program written in any suitable programming language.

[0180] Processors 1311 and 1321, memory 1312 and 1322, and any subset thereof can be configured to provide with Figures 2 to 12 The components corresponding to each box. Although not shown, the device may also include positioning hardware, such as GPS or microelectromechanical systems (MEMS) hardware, which can be used to determine the device's location. Other sensors are also permitted and can be configured to determine position, altitude, speed, orientation, etc., such as barometers, compasses, etc.

[0181] like Figure 13 As shown, transceiver 1313 and transceiver 1323 may be provided, and one or more devices may further include at least one antenna, illustrated as 1314 and 1324 respectively. The device may have a plurality of antennas, such as an antenna array configured for multiple-input multiple-output (MIMO) communication or multiple antennas for various RATs. Other configurations of these devices may be provided, for example. Transceiver 1313 and transceiver 1323 may be a transmitter, a receiver, both a transmitter and a receiver, or units or devices that can be configured for both Tx and Rx.

[0182] The memory and computer program instructions can be configured using a processor for a particular device to cause hardware devices such as a UE to perform any of the processes described above (i.e., Figures 2 to 12 Therefore, in some example embodiments, the non-transitory computer-readable medium may be encoded with computer instructions that, when executed in hardware, perform a process, such as one of the processes described herein. Alternatively, some example embodiments may be executed entirely in hardware.

[0183] In some example embodiments, the apparatus may include being configured to perform Figures 2 to 12 The circuit system of any process or function illustrated. As used in this application, the term "circuit system" may refer to one or more or all of the following: (a) a hardware circuit implementation only (such as an implementation in an analog and / or digital circuit system only); (b) a combination of hardware circuitry and software, such as (if applicable): (i) a combination of (multiple) analog and / or digital hardware circuitry with software / firmware, and (ii) any portion of (multiple) hardware processors with software (including (multiple) digital signal processors, software, and (multiple) memories, which work together to enable a device such as a mobile phone or a server to perform various functions); and (c) (multiple) hardware circuitry and / or (multiple) processors, such as (multiple) microprocessors or portions thereof, which require software (e.g., firmware) for operation, but may be absent when not required for operation. This definition of circuit system applies to all uses of the term in this application, including in any claim. As yet another example, as used in this application, the term circuit system may also cover an implementation of hardware circuitry or a processor (or multiple processors) only, or a portion thereof, and its (or their) accompanying software and / or firmware. For example, and if applicable to certain claim elements, the term "circuit system" will also cover baseband integrated circuits or processor integrated circuits for mobile devices, or similar integrated circuits in servers, cellular network devices, or other computing or network devices.

[0184] Figure 14 Examples of 5G network and system architectures according to certain example embodiments are illustrated. Multiple network functions are shown, which can be implemented as software operating as part of a network device or dedicated hardware, the network device itself or dedicated hardware, or virtual functions operating as a network device or dedicated hardware. Figure 14 The illustrated NE and UE can be similar to NE 1310 and UE 1320, respectively. User plane functions (UPF) can provide services such as intra-RAT and inter-RAT mobility, data packet routing and forwarding, packet inspection, user plane quality of service (QoS) processing, DL packet buffering, and / or triggering DL data notifications. Application functions (AF) can primarily connect to the core network interface to facilitate the application use of service routing and interact with the policy framework.

[0185] According to some example embodiments, processors 1311 and 1321, as well as memories 1312 and 1322, may be included in or form part of a processing circuitry or control circuitry. Additionally, in some example embodiments, transceivers 1313 and 1323 may be included in or form part of a transceiver circuitry.

[0186] In some example embodiments, the apparatus (e.g., NE 1310 and / or UE 1320) may include components for performing methods, procedures, or any variations discussed herein. Examples of such components may include one or more processors, memories, controllers, transmitters, receivers, and / or computer program code for inducing the execution of operations.

[0187] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell and the second identifier identifying at least one downlink reference signal for a second candidate cell; perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; rank a subset of the plurality of identifiers according to parameters of the downlink reference signal measurements using a configured rule set; and train at least one machine learning model using the ranked subset of the plurality of identifiers.

[0188] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of a first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell; components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; components for ranking a subset of the plurality of identifiers according to parameters measured on the downlink reference signal using a configured rule set; and components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0189] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a mobility data collection configuration including a plurality of identifiers; select a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal; and use the machine learning model to perform at least one inference on at least one measurement of the plurality of cells and the plurality of identifiers.

[0190] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a mobility data collection configuration including a plurality of identifiers; components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured based on the downlink reference signal; and components for performing at least one inference using the machine learning model on at least one measurement of the plurality of identifiers and the plurality of cells.

[0191] In various example embodiments, the apparatus 1310 may be controlled by the memory 1312 and the processor 1311 to send a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; receive the first identifier from the first candidate cell; and send to the user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell and the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0192] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; components for receiving the first identifier from the first candidate cell; and components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell.

[0193] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets; perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; rank a subset of the plurality of identifiers according to parameters of the downlink reference signal measurements using a configured rule set; and train at least one machine learning model using the ranked subset of the plurality of identifiers.

[0194] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets; components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; components for ranking a subset of the plurality of identifiers according to parameters measured on the downlink reference signal using a configured rule set; and components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0195] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a mobility data collection configuration including a plurality of identifiers; select a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets; and use the machine learning model to perform at least one inference on at least one measurement of the plurality of cells and the plurality of identifiers.

[0196] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a mobility data collection configuration including a plurality of identifiers; components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets; and components for performing at least one inference on the plurality of identifiers and at least one measurement of the plurality of cells using the machine learning model.

[0197] In various example embodiments, apparatus 1310 may be controlled by memory 1312 and processor 1311 to send a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; receive the first identifier from the first candidate cell; and send to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0198] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; components for receiving the first identifier from the first candidate cell; and components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

[0199] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell and the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets; perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; rank a subset of the plurality of identifiers according to parameters of the downlink reference signal measurements using a configured rule set; and train at least one machine learning model using the ranked subset of the plurality of identifiers.

[0200] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets; components for performing at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; components for ranking a subset of the plurality of identifiers according to parameters measured on the downlink reference signal using a configured rule set; and components for training at least one machine learning model using the ranked subset of the plurality of identifiers.

[0201] In various example embodiments, device 1320 may be controlled by memory 1322 and processor 1321 to receive a mobility data collection configuration including a plurality of identifiers; select a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets; and use the machine learning model to perform at least one inference on at least one measurement of the plurality of cells and the plurality of identifiers.

[0202] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for receiving a mobility data collection configuration including a plurality of identifiers; components for selecting a machine learning model trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal corresponding to the candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets; and components for performing at least one inference using the machine learning model on the plurality of identifiers and at least one measurement of the plurality of cells.

[0203] In various example embodiments, apparatus 1310 may be controlled by memory 1312 and processor 1311 to send a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; receive the first identifier from the first candidate cell; and send to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0204] Some example embodiments may relate to an apparatus including components for performing any of the methods described herein, such as: components for sending a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; components for receiving the first identifier from the first candidate cell; and components for sending to a user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and a second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of a second candidate cell, wherein the number of the plurality of identifiers is not equal to the number of downlink reference signal resource sets.

[0205] The features, structures, or characteristics of the exemplary embodiments described throughout this specification may be combined in any suitable manner in one or more exemplary embodiments. For example, the use of phrases such as “various embodiments,” “some embodiments,” “a few embodiments,” or other similar language throughout this specification refers to the fact that a particular feature, structure, or characteristic described in connection with an exemplary embodiment may be included in at least one exemplary embodiment. Therefore, the appearance of phrases such as “in various embodiments,” “in some embodiments,” “in some embodiments,” or other similar language throughout this specification does not necessarily refer to the same set of exemplary embodiments, and the described features, structures, or characteristics may be combined in any suitable manner in one or more exemplary embodiments.

[0206] As used herein, “at least one of the following: ” and “at least one of ” and similar wording (where the list of two or more elements is connected by “and” or “or”) means at least any one element, or at least any two or more elements, or at least all elements.

[0207] Additionally, the different functions or processes discussed above may be executed in different orders and / or concurrently with each other, if necessary. Furthermore, one or more of the described functions or processes may be optional or may be combined, if necessary. Therefore, the above description should be considered as an illustration of the principles and teachings of certain exemplary embodiments, and not as a limitation thereof.

[0208] Those skilled in the art will readily understand that the exemplary embodiments discussed above can be practiced with different sequences of processes and / or with hardware elements in configurations different from the disclosed configuration. Therefore, although some embodiments have been described based on these exemplary embodiments, it will be apparent to those skilled in the art that certain modifications, variations, and alternative constructions will be readily apparent while remaining within the spirit and scope of the exemplary embodiments.

[0209] Partial vocabulary list

[0210] 3GPP Third Generation Partnership Project

[0211] 5G (Fifth Generation)

[0212] 5GC fifth-generation core

[0213] 6G sixth generation

[0214] AF application functions

[0215] AI (Artificial Intelligence)

[0216] ASIC (Application-Specific Integrated Circuit)

[0217] BM Beam Management

[0218] CBSD Citizen Broadband Radio Service Equipment

[0219] CPU (Central Processing Unit)

[0220] CSI Channel State Information

[0221] CU Centralized Unit

[0222] DL downlink

[0223] DU Distributed Unit

[0224] eMBB Enhanced Mobile Broadband

[0225] eNB Evolutionary Node B

[0226] FR frequency range

[0227] gNB Next Generation Node B

[0228] GPS Global Positioning System

[0229] HDD (Hard Disk Drive)

[0230] ID identifier

[0231] IE Information Elements

[0232] IoT (Internet of Things)

[0233] L1 Floor 1

[0234] L3 Floor 3

[0235] LCM Lifecycle Management

[0236] LTE Long Term Evolution

[0237] LTE-A Long Term Evolution Advanced

[0238] MEMS (Micro-Electro-Mechanical Systems)

[0239] MIMO (Multiple Input Multiple Output)

[0240] ML Machine Learning

[0241] mMTC (Mass Machine Type Communication)

[0242] NE network entity

[0243] NG Next Generation

[0244] NG-eNB Next Generation Evolution Node B

[0245] NG-RAN (Next Generation Radio Access Network)

[0246] NR New Radio

[0247] PBCH (Physical Block Channel)

[0248] PDA (Personal Digital Assistant)

[0249] QoS (Quality of Service)

[0250] RAM (Random Access Memory)

[0251] RAN (Radio Access Network)

[0252] RAT Radio Access Technology

[0253] RF (Radio Frequency)

[0254] ROM (Read-Only Memory)

[0255] RRC Radio Resource Control

[0256] RRM Radio Resource Management

[0257] RS reference signal

[0258] RSRP reference signal received power

[0259] Rx Receive

[0260] SS synchronization signal

[0261] SSB Synchronization Signal Block

[0262] Tx transfer

[0263] UE User Equipment

[0264] UMTS (Universal Mobile Telecommunications System)

[0265] UPF User Plane Functions

[0266] URLLC Ultra-Reliable Low-Latency Communication

[0267] UTRAN (Universal Mobile Telecommunications System Terrestrial Radio Access Network)

[0268] WLAN (Wireless Local Area Network)

Claims

1. A device for communication, comprising: At least one processor; as well as At least one memory stores instructions that, when executed by the at least one processor, cause the device to at least: Receive a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. Perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; Using a set of configuration rules, a subset of the plurality of identifiers is ranked based on parameters measured by the downlink reference signal; as well as At least one machine learning model is trained using the ranked subset of the plurality of identifiers.

2. The apparatus of claim 1, wherein the measurement performed of the at least one includes a channel state information reference signal CSI-RS, a synchronization signal reference signal received power SS-RSRP, and a synchronization signal received power SSB-RP.

3. The apparatus according to claim 1 or 2, wherein at least one of the following: The multiple identifiers are associated with the serving cell configuration; or The at least one downlink reference signal includes at least one synchronization signal / physical broadcast channel block (SSB).

4. A device for communication, comprising: At least one processor; as well as At least one memory stores instructions that, when executed by the at least one processor, cause the device to at least: Receive a mobility data collection configuration, which includes multiple identifiers; A machine learning model is selected to be trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for the corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. Using the machine learning model, perform inferences on at least one of the following: the plurality of identifiers and at least one measurement of the plurality of cells.

5. The apparatus of claim 4, wherein the at least one memory and the instructions, when executed by the at least one processor, further cause the apparatus to perform at least one or more of the following: For a list of Synchronization Signal / Physical Broadcast Channel Block (SSB) resource sets associated with multiple candidate cells, determine the order of the identifiers; or Receive downlink reference signal transmission.

6. The apparatus according to claim 4 or 5, wherein at least one of the following: The multiple identifiers are associated with the serving cell configuration; or The at least one downlink reference signal includes at least one synchronization signal / physical broadcast channel block (SSB).

7. A communication apparatus, comprising: At least one processor; as well as At least one memory stores instructions that, when executed by the at least one processor, cause the device to at least: Send a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; Receive the first identifier from the first candidate cell; as well as Send to the user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and the second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of the second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

8. A method for communication, comprising: Receive a plurality of identifiers configured for mobility data collection, each of the plurality of identifiers indicating at least one downlink reference signal for a corresponding candidate cell, the plurality of identifiers including a first identifier and a second identifier, the first identifier identifying at least one downlink reference signal for a first candidate cell, the second identifier identifying at least one downlink reference signal for a second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. Perform at least one measurement on the downlink reference signal indicated by each of the plurality of identifiers; Using a set of configuration rules, a subset of the plurality of identifiers is ranked based on parameters measured by the downlink reference signal; as well as At least one machine learning model is trained using the ranked subset of the plurality of identifiers.

9. A method for communication, comprising: Receive mobility data collection configuration including multiple identifiers; A machine learning model is selected to be trained on a ranked subset of the plurality of identifiers, wherein the ranked subset of the plurality of identifiers includes a first identifier identifying a first candidate cell and a second identifier identifying a second candidate cell, each of the plurality of identifiers indicating at least one downlink reference signal for the corresponding candidate cell, and the subset of the plurality of identifiers is ranked according to parameters measured by the downlink reference signal, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets. Using the machine learning model, at least one inference is performed on at least one measurement of the plurality of identifiers and the plurality of cells.

10. A method for communication, comprising: Send a request to a first candidate cell for a first identifier configured for mobility data collection, wherein the first identifier identifies the first candidate cell; Receive the first identifier from the first candidate cell; as well as Send to the user equipment a plurality of identifiers including at least the first identifier, each of the plurality of identifiers indicating at least one downlink reference signal of a corresponding candidate cell, the plurality of identifiers including the first identifier and the second identifier, the first identifier identifying at least one downlink reference signal of the first candidate cell, the second identifier identifying at least one downlink reference signal of the second candidate cell, wherein the number of the plurality of identifiers is equal to the number of downlink reference signal resource sets.

Images