Predictive decision-making in a wireless communication network
A network node ( 16 ) obtains, for each of one or more wireless devices, radio performance information ( 18 ) indicating measured radio performances that multiple cells respectively provided to the wireless device ( 20 A) while the wireless device ( 20 A) was simultaneously connected to the multiple cells in multi-connectivity operation under certain conditions. Based on the obtained radio performance information ( 18 ), the network node ( 16 ) builds a prediction model ( 24 ) which models radio performance that a cell will provide under one or more conditions. The network node ( 16 ) then makes or configures another network node to make decisions on mobility, scheduling and/or radio configuration of wireless devices based on the prediction model ( 24 ).
1 . A method performed by a network node, the method comprising:
obtaining, for each of one or more wireless devices, radio performance information indicating measured radio performances that multiple cells respectively provided to the wireless device while the wireless device was simultaneously connected to the multiple cells in multi-connectivity operation under certain conditions;
based on the obtained radio performance information, building a prediction model which models radio performance that a cell will provide under one or more conditions, wherein the prediction model models radio performance that a cell will provide in terms of one or more performance metrics predicted for the cell under the one or more conditions, wherein the one or more performance metrics include one or more absolute or relative measures of data throughput, radio link control throughput, signal strength, signal quality, spectral efficiency, packet error rate, packet delay, packet loss rate, channel quality information, and/or block error rate for the cell; and
making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of wireless devices based on comparing radio performances that different cells will provide under the same one or more conditions according to the prediction model, wherein the radio performances that the different cells will provide under the same one or more conditions are modeled based on radio performances that the different cells respectively provided in multi-connectivity operation.
2 . The method of claim 1 , wherein the one or more wireless devices are one or more first wireless devices, wherein, for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation during certain times and/or at certain locations, wherein the prediction model models radio performance that a cell will provide during one or more times and/or at one or more locations, and wherein making, or configuring another network node to make, the decisions comprises making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide during a given time and/or at a given location as predicted according to the prediction model.
3 . The method of claim 1 , wherein the one or more wireless devices are one or more first wireless devices, wherein, for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation under a certain cell load level, wherein the prediction model models radio performance that a cell will provide under one or more cell load levels, and wherein making, or configuring another network node to make, the decisions comprises making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide under a given cell load level as predicted according to the prediction model.
4 . The method of claim 1 , wherein the one or more wireless devices are one or more first wireless devices, wherein, for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation using a certain radio configuration, wherein the prediction model models radio performance that a cell will provide under one or more radio configurations, and wherein making, or configuring another network node to make, the decisions comprises making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide using a given radio configuration as predicted according to the prediction model.
5 . The method of claim 4 , wherein the certain radio configuration is a multi-connectivity configuration, wherein the one or more radio configurations are one or more multi-connectivity configurations, and wherein the given radio configuration is a multi-connectivity configuration.
6 . The method of claim 1 , wherein the one or more wireless devices are one or more first wireless devices, wherein, for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation with certain device capabilities, wherein the prediction model models radio performance that a cell will provide to any wireless device with one or more device capabilities, and wherein making, or configuring another network node to make, the decisions comprises making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide to the second wireless device with one or more given device capabilities as predicted according to the prediction model.
7 . The method of claim 1 , comprising making, or configuring another network node to make, decisions on mobility of wireless devices based on the prediction model, wherein said mobility comprises inter-frequency mobility or inter radio access technology mobility.
8 . The method of claim 1 , comprising making, or configuring another network node to make, decisions on mobility of wireless devices based on the prediction model, wherein said decisions on mobility include at least one of any one or more of:
a decision on which cell is to be a target of a mobility procedure for a second wireless device;
a decision on which wireless device is to perform a mobility procedure; and
a decision on where or when a second wireless device is to perform a mobility procedure.
9 . The method of claim 1 , comprising making, or configuring another network node to make, decisions on scheduling of wireless devices based on the prediction model, wherein making, or configuring another network node to make, decisions on scheduling of wireless devices based on the prediction model comprises making, or configuring another network node to make, a decision that a second wireless device is to be prioritized for scheduling before the second wireless device arrives at a location where, according to the prediction model, one or more cells will provide radio performance below a threshold.
10 . The method of claim 1 , comprising making, or configuring another network node to make, decisions on radio configuration of wireless devices based on the prediction model, wherein said decision on radio configuration comprise at least one of any one or more of:
a decision on whether or how to configure the multi-connectivity operation of a second wireless device;
a decision on whether or how to configure carrier aggregation for a second wireless device; and
a decision on whether or how to configure power boosting for a second wireless device.
11 . The method of claim 1 , wherein said obtaining comprises inferring or deducing at least some of the radio performance information, wherein inferring or deducing at least some of the radio performance information comprises inferring or deducing radio performance information that indicates radio performance of a cell, based on:
a change in an extent to which the cell transmits or receives user plane traffic to a second wireless device; and/or
a change in an allocation of resources for transmitting or receiving user plane traffic to a second wireless device in the cell.
12 . The method of claim 1 , wherein the network node is:
a radio network node configured as a master radio network node for multi-connectivity operation of at least one of the one or more wireless devices;
a secondary radio network node for multi-connectivity operation of at least one of the one or more wireless devices; or
an operation and maintenance node.
13 . The method of claim 1 , wherein said configuring another network node to make decisions on mobility, scheduling, and/or radio configuration of wireless devices based on the prediction model comprises transmitting the prediction model to said another network node.
14 . The method of claim 1 , wherein the prediction model is a machine learning model built using machine learning.
15 . The method of claim 1 , wherein making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of wireless devices based on the prediction model comprises making, or configuring another network node to make, decisions on mobility, scheduling, and/or radio configuration of wireless devices based on comparing radio performances that different cells will provide under the same one or more conditions according to the prediction model, wherein the radio performances that the different cells will provide under the same one or more conditions are modeled based on radio performances that the different cells respectively provided in multi-connectivity operation.
16 . The method of claim 1 , wherein:
said obtaining comprises inferring or deducing at least some of the radio performance information, wherein inferring or deducing at least some of the radio performance information comprises inferring or deducing radio performance information that indicates radio performance of a cell, based on whether the cell has transmitted or received user plane traffic to a second wireless device within a certain time interval; and/or
the method comprises making, or configuring another network node to make, decisions on radio configuration of wireless devices based on the prediction model, wherein said decision on radio configuration comprises a decision on whether or how to configure transmission time interval (TTI) bundling for a second wireless device.
17 . The method of claim 1 , wherein the radio performances that the different cells will provide under the same one or more conditions are modeled based on parallel measurements of radio performances that the different cells respectively provided to a wireless device while simultaneously serving that wireless device in multi-connectivity operation, the multi-connectivity operation enabling the parallel measurements to be obtained under the same one or more conditions such that the radio performances are comparable.
18 . A network node configured to:
obtain, for each of one or more wireless devices, radio performance information indicating measured radio performances that multiple cells respectively provided to the wireless device while the wireless device was simultaneously connected to the multiple cells in multi-connectivity operation under certain conditions;
based on the obtained radio performance information, build a prediction model which models radio performance that a cell will provide under one or more conditions, wherein the prediction model models radio performance that a cell will provide in terms of one or more performance metrics predicted for the cell under the one or more conditions, wherein the one or more performance metrics include one or more absolute or relative measures of data throughput, radio link control throughput, signal strength, signal quality, spectral efficiency, packet error rate, packet delay, packet loss rate, channel quality information, and/or block error rate for the cell; and
make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of wireless devices based on comparing radio performances that different cells will provide under the same one or more conditions according to the prediction model, wherein the radio performances that the different cells will provide under the same one or more conditions are modeled based on radio performances that the different cells respectively provided in multi-connectivity operation.
19 . The network node of claim 18 , wherein the one or more wireless devices are one or more first wireless devices, wherein:
for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation during certain times and/or at certain locations, wherein the prediction model models radio performance that a cell will provide during one or more times and/or at one or more locations, and wherein the processing circuitry is configured to make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide during a given time and/or at a given location as predicted according to the prediction model; and/or
for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation under a certain cell load level, wherein the prediction model models radio performance that a cell will provide under one or more cell load levels, and wherein the processing circuitry is configured to make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide under a given cell load level as predicted according to the prediction model; and/or
for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation using a certain radio configuration, wherein the prediction model models radio performance that a cell will provide under one or more radio configurations, and wherein the processing circuitry is configured to make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide using a given radio configuration as predicted according to the prediction model; and/or
for each of the one or more first wireless devices, the radio performance information obtained for the first wireless device indicates measured radio performances that the multiple cells respectively provided to the first wireless device while the first wireless device was simultaneously connected to the multiple cells in the multi-connectivity operation with certain device capabilities, wherein the prediction model models radio performance that a cell will provide to any wireless device with one or more device capabilities, and wherein the processing circuitry is configured to make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of a second wireless device based on comparing radio performances that different cells will provide to the second wireless device with one or more given device capabilities as predicted according to the prediction model.
20 . The network node of claim 18 , comprising making, or configuring another network node to make, decisions on mobility of wireless devices based on the prediction model, wherein said decisions on mobility include at least one of any one or more of:
a decision on which cell is to be a target of a mobility procedure for a second wireless device;
a decision on which wireless device is to perform a mobility procedure; and
a decision on where or when a second wireless device is to perform a mobility procedure.
21 . The network node of claim 18 , wherein the processing circuitry is configured to make, or configure another network node to make, decisions on scheduling of wireless devices based on the prediction model, by making, or configuring another network node to make, a decision that a second wireless device is to be prioritized for scheduling before the second wireless device arrives at a location where, according to the prediction model, one or more cells will provide radio performance below a threshold.
22 . The network node of claim 18 , wherein the processing circuitry is configured to make, or configure another network node to make, decisions on radio configuration of wireless devices based on the prediction model, wherein a decision on radio configuration comprise at least one of any one or more of:
a decision on whether or how to configure multi-connectivity operation of a second wireless device;
a decision on whether or how to configure carrier aggregation for a second wireless device; and
a decision on whether or how to configure power boosting for a second wireless device.
23 . The network node of claim 18 , wherein the processing circuitry is configured to obtain the radio performance information by comprises inferring or deducing at least some of the radio performance information, wherein inferring or deducing at least some of the radio performance information comprises inferring or deducing radio performance information that indicates radio performance of a cell, based on:
a change in an extent to which the cell transmits or receives user plane traffic to a second wireless device; and/or
a change in an allocation of resources for transmitting or receiving user plane traffic to a second wireless device in the cell.
24 . The network node of claim 18 , wherein the processing circuitry is configured to make, or configure another network node to make, decisions on mobility, scheduling, and/or radio configuration of wireless devices based on the prediction model by making, or configuring another network node to make, decisions on mobility of a given wireless device based on comparing radio performances that different cells will provide to the given wireless device under the same one or more conditions according to the prediction model, wherein the radio performances that the different cells will provide under the same one or more conditions are modeled by the prediction model based on radio performances that the different cells respectively provided in multi-connectivity operation.
25 . The network node of claim 18 , wherein:
the processing circuitry is configured to obtain the radio performance information by comprises inferring or deducing at least some of the radio performance information, wherein inferring or deducing at least some of the radio performance information comprises inferring or deducing radio performance information that indicates radio performance of a cell, based on whether the cell has transmitted or received user plane traffic to a second wireless device within a certain time interval; and/or
the processing circuitry is configured to make, or configure another network node to make, decisions on radio configuration of wireless devices based on the prediction model, wherein a decision on radio configuration comprises a decision on whether or how to configure transmission time interval (TTI) bundling for a second wireless device.