Base station, apparatus, method and computer readable medium for mixed-numerology configuration selection

1 . An apparatus for mixed-numerology configuration selection, the apparatus comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to:

extract statistical features related to service requirements of UEs associated with a base station;

extract statistical features related to incoming and outgoing traffic of the base station;

extract statistical features related to a wireless channel condition and a wireless channel configuration;

generate a context vector based on the statistical features related to the service requirements, the statistical features related to the incoming and outgoing traffic, and the statistical features related to the wireless channel condition and the wireless channel configuration;

estimate Quality of Service (QoS) for mixed-numerology configurations based on the generated context vector using a first Machine Learning (ML) model array including a plurality of ML models, each of the ML models being configured to estimate QoS for a mixed-numerology configuration; and

select a mixed-numerology configuration to be used for data transmission and data reception based on the QoS estimated by the ML models.

2 . The apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to concatenate the statistical features related to the service requirement, the statistical features related to the overall incoming and outgoing traffic, and the statistical features related to the wireless channel condition and the wireless channel configuration to generate the context vector.

3 . The apparatus according to claim 1 , wherein the QoS is a ratio of a number of the QoS flows corresponding to the UEs for which all service requirements are satisfied and a total number of the QoS flows.

4 . The apparatus according to claim 1 , wherein each of the ML models included in the first ML model array is pre-trained in an offline fashion using collected real-world data or using synthetic data generated by a communication network simulation.

5 . The apparatus according to claim 1 , wherein the context vector is input to the each of the ML models included in the first ML model array, and the ML models output estimated QoS for mutually different mixed-numerology configurations.

6 . The apparatus according to claim 1 , wherein the at least one processor is further configured to execute the instructions to:

compute QoS obtained for the selected mixed-numerology configuration;

to store, in a database, the context vector, the selected mixed-numerology configuration and the QoS computed for the selected mixed-numerology configuration as context information; and

update the ML models included in the first ML model array based on the context information.

7 . The apparatus according to claim 6 , wherein the at least one processor is configured to execute the instructions to:

obtain the context information from the database and update the ML models included in the first ML model array using the obtained context information, and a pair of a second ML model array and a third ML model array, each of the second ML model array and the third ML model array including a plurality of ML models each configured to estimate QoS for a mixed-numerology configuration based on the context vector.

8 . The apparatus according to claim 7 , wherein the at least one processor is configured to execute the instructions to:

generate a test mini-batch of the context information and a train mini-batch of the context information;

update parameters of the ML models included in the third ML model array;

input the test mini-batch to the second ML model array and the third ML model array to cause the second ML model array and the third ML model array to estimate the QoS respectively;

compute error difference of the QoS estimated by the second ML model array and error difference of the QoS estimated by the third ML model array;

determine whether to update the ML models included in the first ML model array or not based on the computed error differences;

update, when it is determined to update the ML models included in the first ML model array, the parameters of the ML models included in the first ML model array with the parameters of the ML models included in the third ML model array; and

copy the parameters of the ML models included in the third ML model array into the ML models included in the second ML model array.

9 . The apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to:

extract values each indicating a QoS from data packets;

obtain service requirements corresponding to each of the extracted values using a table mapping the value to service requirements; and

compute at least one statistical value of the obtained service requirements as the statistical features related to the service requirements.

10 . The apparatus according to claim 1 , wherein the service requirements include at least one of a delay budget and packet error rate tolerance of the QoS flows corresponding to the UEs.

11 . The apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to:

collect traffic feature information; and

compute at least one statistical value based on the collected traffic feature information as the statistical features related to incoming and outgoing traffic.

12 . The apparatus according to claim 11 , wherein the traffic feature information includes at least one of UE ID's (Identifiers), packet size, and a packet arrival time-stamp.

13 . The apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to:

collect channel feature information; and

compute at least one statistical value based on the collected channel feature information as the statistical features related to incoming and outgoing traffic.

14 . The apparatus according to claim 13 , wherein the channel feature information includes at least one of a Channel Quality Indicator (CQI) report, a number of available resource blocks, values of guard bands, downlink and uplink share percentages and Hybrid Automatic Repeat reQuest (HARQ) process identifiers.

15 . The apparatus according to claim 1 , wherein the at least one processor is configured to execute the instructions to select a mixed-numerology configuration with a highest estimated QoS among the QoS estimated by the ML models included in the first ML model array.

16 . A base station, comprising:

a communication apparatus configured to communicate with a core network and UEs, and

the mixed-numerology configuration selection apparatus according to claim 1 .

17 . The base station according to claim 16 , wherein the at least one processor is configured the instructions to concatenate the statistical features related to the service requirement, the statistical features related to the overall incoming and outgoing traffic, and the statistical features related to the wireless channel condition and the wireless channel configuration to generate the context vector.

18 . The base station according to claim 16 , wherein the QoS is a ratio of a number of the QoS flows corresponding to the UEs for which all service requirements are satisfied and a total number of the QoS flows.

19 . A method for mixed-numerology configuration selection, the method comprising:

extracting statistical features related to service requirements of UEs associated with a base station,

extracting statistical features related to incoming and outgoing traffic of the base station,

extracting statistical features related to a wireless channel condition and a wireless channel configuration,

generating a context vector based on the statistical features related to the service requirements, the statistical features related to the incoming and outgoing traffic, and the statistical features related to the wireless channel condition and the wireless channel configuration,

estimating Quality of Service (QoS) for mixed-numerology configurations based on the context vector using a Machine Learning (ML) model array including a plurality of ML models, each of the ML models being configured to estimate QoS for a mixed-numerology configuration, and

selecting a mixed-numerology configuration to be used for data transmission and data reception based on the QoS estimated by the ML models.

20 . A non-transitory computer readable medium for mixed-numerology configuration selection, the non-transitory computer readable medium stores a program for a causing a computer to:

extract statistical features related to service requirements of UEs associated with a base station,

extract statistical features related to incoming and outgoing traffic of the base station,

extract statistical features related to a wireless channel condition and a wireless channel configuration,

generate a context vector based on the statistical features related to the service requirements, the statistical features related to the incoming and outgoing traffic, and the statistical features related to the wireless channel condition and the wireless channel configuration,

input the context vector to a Machine Learning (ML) model array including a plurality of ML models, each of the ML models being configured to estimate Quality of Service (QoS) for a mixed-numerology configuration, to cause the ML models to estimate the QoS, and

select a mixed-numerology configuration to be used for data transmission and data reception based on the QoS estimated by the ML models.