The present disclosure relates to machine-learning or artificial intelligence within the context of a wireless network and, more specifically, to activating intelligent wireless communication device reporting in a wireless network.
The current Fifth Generation (5G) Radio Access Network (RAN) architecture, referred to as the Next Generation RAN (NG-RAN) architecture, is depicted and described in Third Generation Partnership Project (3GPP) Technical Specification (TS) 38.401 (see, e.g., v16.3.0). In particular,
A gNB may also be connected to a Long Term Evolution (LTE) evolved Node B (eNB) via the X2 interface. Another architectural option is that where an LTE eNB connected to the Evolved Packet Core (EPC) network is connected over the X2 interface with a so called nr-gNB. The latter is a gNB not connected directly to a core network (CN) and connected via X2 to an eNB for the sole purpose of performing dual connectivity.
The architecture in
It is important to fully utilize the potential of machine learning (ML) for wireless networks such as the NG-RAN of a 5GS, for example by extracting more data from all nodes in the network. One problem in applying ML for wireless networks is the variable data transfer cost depending on wired or over-the-air transmission. Enabling Artificial Intelligence (AI) or ML by extending the device reporting by including different types of information, from radio to physical measurements, would lead to increased signaling. The trade-off between increased data signaling versus enabling improved intelligence at the network is a challenging problem.
Another alternative is to explore the use of potential augmentation information provided by an AI-model at the device using so-called “intelligent devices”.
The UE can have machine learning models able to predict a certain quantity. The model could for example be:
In regard to the use of intelligent networks, the use cases can comprise:
Signal quality prediction is of particular interest. Based on received User Equipment (UE) data from measurement reports, the network can learn, for example, what sequence of signal quality measurements (e.g. Reference Signal Received Power (RSRP)) that result in a large signal quality drop (e.g. turning around the corners in
In the example in
Systems and methods for activating intelligent wireless communication device reporting in a wireless network are disclosed. Embodiments of a method performed by a wireless communication device for machine-learned optimization of wireless networks is proposed. In one embodiment, the method includes sending, to a network node, information that indicates one or more capabilities of the wireless communication device for reporting of predicted values that are predicted by the wireless communication device using one or more machine learning capabilities of the wireless communication device. The method further includes receiving, from the network node, a request. The request includes (a) a request to start reporting predicted values, (b) a request to start training a machine learning model for generating predicted values, or (c) both (a) and (b). The method further includes performing one or more actions in response to receiving the request. In this manner, network operation may be improved by supporting intelligent reporting.
In one embodiment, performing the one or more actions includes generating one or more reports comprising one or more predicted values based on the machine learning model sending the one or more reports to the network node. In one embodiment, performing the one or more actions further comprises training the machine learning model for generating the predicted values. In one embodiment, the one or more reports further comprise information that indicates an accuracy or confidence level of the one or more predicted values.
In one embodiment, generating and sending the one or more reports is activated when a triggering criterion is satisfied. In one embodiment, the triggering criterion is a required accuracy level for the one or more predicted values. In one embodiment, the triggering criterion is a required confidence level for the one or more predicted values. In one embodiment, the triggering criterion is a time-based triggering criterion. In one embodiment, the triggering criterion is a prediction performance-based triggering criterion. In one embodiment, the triggering criterion is based on: availability of network capabilities at the network node; subscription to one or more services at the network node; configuration at the wireless communication device for (a) Guaranteed Flow Bit Rate (GFBR) for Upload and Download, (b) Maximum Packet Loss Rate for Upload and Download, (c) reporting of Quality of Experience (QoE) measurements for at least one application, or (d) any two or more of (a)-(c); detection of a change of Quality of Service (QoS) parameters associated with the wireless communication device; the wireless communication device being served by a certain slice; the wireless communication device being located within a geographic area; the wireless communication device having a specific Service Profile Identifier (SPID); or a mobility pattern of the wireless communication device.
In one embodiment, the request comprises a request to start reporting predicted values at a particular time(s) or during a particular time window(s). In some embodiments, the predicted values comprise predicted Radio Resource management (RRM) related values. In one embodiment, the predicted values comprise predicted beam related values. In one embodiment, the predicted values comprise predicted values for future traffic needs of the wireless communication device.
In one embodiment, the predicted values comprise predicted measurement values for (a) one or more frequencies, (b) traffic steering, (c) serving cell selection, (d) QoS prediction, (e) RRM, or (f) any two or more of (a)-(e).
In one embodiment, the information that indicates one or more capabilities of the wireless communication device for reporting of predicted values further comprises a performance metric indicative of an accuracy of the wireless communication device for performance of the one or more capabilities
In one embodiment, the information that indicates the one or more capabilities of the wireless communication device for reporting of predicted values comprises physical characteristic data for the wireless communication device descriptive of: (a) battery power, (b) available memory, (c) computational capacity, (d) sensor capabilities, (e) parameters descriptive of a physical environment of the wireless communication device, (f) acceleration or velocity of the wireless communication device, (g) nearby network infrastructure, or (h) any two or more of (a)-(g).
In one embodiment, prior to sending the information that indicates the one or more capabilities, the method further includes receiving a request from the network node for the one or more capabilities of the wireless communication device for reporting of predicted values. Sending the information that indicates the one or more capabilities of the wireless communication device for reporting of predicted values comprises sending the information that indicates the one or more capabilities of the wireless communication device for reporting of predicted values responsive to receiving the request from the network node for the one or more capabilities of the wireless communication device for reporting of predicted values.
In one embodiment, performing the one or more actions in response to receiving the request comprises activating one or more procedures that replace measurements with predicted values.
In one embodiment, performing the one or more actions comprises, after transitioning from a connected state to an inactive state and subsequently transitioning back to the connected state in association to a second network node, providing data resulting from performing the one or more actions to the second network node.
Corresponding embodiments of a wireless communication device are disclosed. In one embodiment, a wireless communication device is adapted to send, to a network node, information that indicates one or more capabilities of the wireless communication device for reporting of predicted values that are predicted by the wireless communication device using one or more machine learning capabilities of the wireless communication device. The wireless communication device is adapted to receive, from the network node, a request, the request comprising: (a) a request to start reporting predicted values, (b) a request to start training a machine learning model for generating predicted values, or (c) both (a) and (b). The wireless communication device is adapted to perform one or more actions in response to receiving the request.
In one embodiment, a wireless communication device includes one or more transmitters, one or more receivers, and processing circuitry associated with the one or more transmitters and the one or more receivers. The processing circuitry is configured to cause the wireless communication device to send, to a network node, information that indicates one or more capabilities of the wireless communication device for reporting of predicted values that are predicted by the wireless communication device using one or more machine learning capabilities of the wireless communication device. The processing circuitry is further configured to cause the wireless communication device to receive, from the network node, a request, the request comprising: (a) a request to start reporting predicted values, (b) a request to start training a machine learning model for generating predicted values, or (c) both (a) and (b). The processing circuitry is further configured to cause the wireless communication device to perform one or more actions in response to receiving the request.
In some embodiments, a method performed by a network node for machine-learned optimization of wireless networks is proposed. The method includes receiving, from a plurality of wireless communication devices, information that indicates one or more capabilities of the plurality of wireless communication devices for reporting of predicted values. The method includes either or both of: (a) determining one or more wireless communication devices from which to request reporting of predicted values from the plurality of wireless communication devices based on the received information and (b) determining one or more reports to request from one or more wireless communication devices from among the plurality of wireless communication devices based on the received information. The method includes sending, to the one or more wireless communication devices, one or more messages related to activation of intelligent reporting. The one or more messages include (a) a request to start reporting predicted values, (b) a request to start training a machine learning model for generating predicted values, or (c) both (a) and (b).
Corresponding embodiments of a network node are also disclosed.
In another embodiment, a method performed by a network node includes sending, to a supervising network node, data indicative of a request to configure one or more wireless communication devices for reporting of predicted values. The method further includes responsive to sending the data indicative of the request, receiving, from the supervising network node, predicted values from the one or more wireless communication devices. The method further includes performing one or more actions based at least in part on the predicted values from the one or more wireless communication devices.
In one embodiment, sending the data indicative of the request to configure the one or more wireless communication devices comprises sending, to the supervising network node, data indicative of a request to configure one or more wireless communication devices for reporting of predicted values from the one or more wireless communication devices for beam level coverage for one or more of served beams or neighbor cell beams.
In one embodiment, performing the one or more actions comprises (a) adjusting a shape of one or more beams, (b) extending coverage of one or more beams, (c) reducing coverage of one or more beams, (d) reducing a number of beams for a geographic area, or (e) two or more of any of (a)-(d).
In one embodiment, the network node comprises a New Radio (NR) base station (gNB) Distributed Unit (DU), and the supervising node comprises a gNB Central Unit (CU).
The accompanying drawing figures incorporated in and forming a part of this specification illustrate several aspects of the disclosure, and together with the description serve to explain the principles of the disclosure.
The embodiments set forth below represent information to enable those skilled in the art to practice the embodiments and illustrate the best mode of practicing the embodiments. Upon reading the following description in light of the accompanying drawing figures, those skilled in the art will understand the concepts of the disclosure and will recognize applications of these concepts not particularly addressed herein. It should be understood that these concepts and applications fall within the scope of the disclosure.
Some of the embodiments contemplated herein will now be described more fully with reference to the accompanying drawings. Other embodiments, however, are contained within the scope of the subject matter disclosed herein, the disclosed subject matter should not be construed as limited to only the embodiments set forth herein; rather, these embodiments are provided by way of example to convey the scope of the subject matter to those skilled in the art.
Generally, all terms used herein are to be interpreted according to their ordinary meaning in the relevant technical field, unless a different meaning is clearly given and/or is implied from the context in which it is used. All references to a/an/the element, apparatus, component, means, step, etc. are to be interpreted openly as referring to at least one instance of the element, apparatus, component, means, step, etc., unless explicitly stated otherwise. The steps of any methods disclosed herein do not have to be performed in the exact order disclosed, unless a step is explicitly described as following or preceding another step and/or where it is implicit that a step must follow or precede another step. Any feature of any of the embodiments disclosed herein may be applied to any other embodiment, wherever appropriate. Likewise, any advantage of any of the embodiments may apply to any other embodiments, and vice versa. Other objectives, features, and advantages of the enclosed embodiments will be apparent from the following description.
Radio Node: As used herein, a “radio node” is either a radio access node or a wireless communication device.
Radio Access Node: As used herein, a “radio access node” or “radio network node” or “radio access network node” is any node in a Radio Access Network (RAN) of a cellular communications network that operates to wirelessly transmit and/or receive signals. Some examples of a radio access node include, but are not limited to, a base station (e.g., a New Radio (NR) base station (gNB) in a Third Generation Partnership Project (3GPP) Fifth Generation (5G) NR network or an enhanced or evolved Node B (eNB) in a 3GPP Long Term Evolution (LTE) network), a high-power or macro base station, a low-power base station (e.g., a micro base station, a pico base station, a home eNB, or the like), a relay node, a network node that implements part of the functionality of a base station (e.g., a network node that implements a gNB Central Unit (gNB-CU) or a network node that implements a gNB Distributed Unit (gNB-DU)) or a network node that implements part of the functionality of some other type of radio access node.
Core Network Node: As used herein, a “core network node” is any type of node in a core network or any node that implements a core network function. Some examples of a core network node include, e.g., a Mobility Management Entity (MME), a Packet Data Network Gateway (P-GW), a Service Capability Exposure Function (SCEF), a Home Subscriber Server (HSS), or the like. Some other examples of a core network node include a node implementing a Access and Mobility Management Function (AMF), a User Plane Function (UPF), a Session Management Function (SMF), an Authentication Server Function (AUSF), a Network Slice Selection Function (NSSF), a Network Exposure Function (NEF), a Network Function (NF) Repository Function (NRF), a Policy Control Function (PCF), a Unified Data Management (UDM), or the like.
Communication Device: As used herein, a “communication device” is any type of device that has access to an access network. Some examples of a communication device include, but are not limited to: mobile phone, smart phone, sensor device, meter, vehicle, household appliance, medical appliance, media player, camera, or any type of consumer electronic, for instance, but not limited to, a television, radio, lighting arrangement, tablet computer, laptop, or Personal Computer (PC). The communication device may be a portable, hand-held, computer-comprised, or vehicle-mounted mobile device, enabled to communicate voice and/or data via a wireless or wireline connection.
Wireless Communication Device: One type of communication device is a wireless communication device, which may be any type of wireless device that has access to (i.e., is served by) a wireless network (e.g., a cellular network). Some examples of a wireless communication device include, but are not limited to: a User Equipment device (UE) in a 3GPP network, a Machine Type Communication (MTC) device, and an Internet of Things (IoT) device. Such wireless communication devices may be, or may be integrated into, a mobile phone, smart phone, sensor device, meter, vehicle, household appliance, medical appliance, media player, camera, or any type of consumer electronic, for instance, but not limited to, a television, radio, lighting arrangement, tablet computer, laptop, or PC. The wireless communication device may be a portable, hand-held, computer-comprised, or vehicle-mounted mobile device, enabled to communicate voice and/or data via a wireless connection.
Network Node: As used herein, a “network node” is any node that is either part of the RAN or the core network of a cellular communications network/system.
Intelligent RRM Reporting: As used herein, “intelligent RRM reporting” is RRM reporting by a wireless communication device that is based on a machine learning model or artificial intelligence at the wireless communication device. Similarly, as used herein, an “intelligent RRM report” is a RRM report generated and sent by a wireless communication device to convey RRM related information that is based on a machine learning model or artificial intelligence at the wireless communication device.
Note that the description given herein focuses on a 3GPP cellular communications system and, as such, 3GPP terminology or terminology similar to 3GPP terminology is oftentimes used. However, the concepts disclosed herein are not limited to a 3GPP system.
Note that, in the description herein, reference may be made to the term “cell”; however, particularly with respect to 5G NR concepts, beams may be used instead of cells and, as such, it is important to note that the concepts described herein are equally applicable to both cells and beams.
There currently exist certain challenge(s). It is important to fully utilize the potentials in machine learning for wireless networks, for example by extracting more data from all nodes in the network. One method is to extend UE reporting by including different types of information, from radio to physical measurements. This would however lead to increased signaling. With the densification of radio networks, utilization of higher chunks of spectrum, and higher complexity of the network, it is important to leverage intelligence in all aspects of the network with minimal network impact. One method is to move some intelligence to the UE, by downloading such intelligence to the device. Methods to receive knowledge on whether the device is capable or configured to use machine learning (ML) models/algorithms has been proposed. Such method includes receiving from the device a message indicating whether it is either capable or configured to use (or it is using) machine learning models and algorithms for its operation. However, the existing solutions do not cover aspects on selecting and activating/deactivating the use of the ML-models for intelligent reporting. It is important to keep the signaling overhead to a minimum by making sure to only provide e.g. predicted signal quality information when it is needed at the network.
Certain aspects of the present disclosure and their embodiments may provide solutions to the aforementioned or other challenges. Embodiments of the present disclosure provide a framework for selecting and activating one or more intelligent reports from UEs, e.g., based on base station information and UE capabilities. In one embodiment, the intelligent report comprises a value estimated by a machine-learning algorithm such as, for example, a predicted future signal quality value or a value associated to one or more reference signals. In one embodiment, the intelligent report is used to configure one or more Radio Resource Management (RRM) related parameters.
Certain embodiments may provide one or more of the following technical advantage(s). Embodiments disclosed herein may improve RRM operation by leveraging improved capabilities from UEs supporting intelligent RRM reporting, by activating such reporting. The UE has more information (data) regarding its experienced radio environment and also of its surroundings using any type of available information (e.g., camera, Light Detection and Ranging (LIDAR), Global Naviation Satellite System (GNSS), Inertial Measurement Unit (IMU)). The amount of data available at the UE leads to improved machine learning models.
Embodiments disclosed herein may:
Embodiments disclosed herein also list potential new intelligent reports from UEs, which enable better RRM. For example, signaling a future signal quality value as a probability density function.
Note that while many of the embodiments described herein focus on the example of intelligent RRM reporting, the present disclosure is not limited thereto.
The base stations 502 and the low power nodes 506 provide service to wireless communication devices 512-1 through 512-5 in the corresponding cells 504 and 508. The wireless communication devices 512-1 through 512-5 are generally referred to herein collectively as wireless communication devices 512 and individually as wireless communication device 512. In the following description, the wireless communication devices 512 are oftentimes UEs and as such sometimes referred to herein as UEs 512, but the present disclosure is not limited thereto.
The network node receives information that indicates capabilities of one or more wireless communication devices 512 for intelligent RRM reporting (step 604). The information in step 604 may be received in response to the request of step 602. The received information may also include information that indicates a performance metric for each intelligent RRM report supported by the wireless communication device 512.
The network node determines one or more intelligent RRM reports to be activated and one or more wireless communication devices 512 for which the intelligent RRM reports are to be activated, based on the received capability information from step 604 (step 606). The network node then activates the determined intelligent RRM report(s) for the determined wireless communication device(s) 512 (step 608). In other words, the network node sends a message(s) to the determined wireless communication device(s) 512 that instruct the determined wireless communication device(s) 512 to activate the determined intelligent RRM report(s). Optionally, message(s) sent to the wireless communication device(s) 512 to activate the intelligent RRM report(s) may include a triggering criterion that describes when to activate the intelligent RRM report(s). Optionally, the message(s) may include information that indicates one or more time-windows for which the wireless communication device(s) is to produce the intelligent RRM report(s). Optionally, the message(s) may include information that indicates that the wireless communication device(s) 512 are to start training a model (e.g. a ML model such as, e.g., a neural network) for intelligent RRM reporting. For example, the network node may only activate the intelligent RRM report(s) during a time period(s) when such reporting is useful where the time period(s) can be based on, e.g., networking scheduling and/or UE service type. Further, the determined wireless communication device(s) 512 for which the intelligent RRM report(s) are activated may be selected based on the capabilities of those wireless communication device(s) 512. Further, in one embodiment, intelligent RRM reporting is triggered when it fulfills a certain accuracy requirement. The accuracy requirement may be configured by the network node or predefined or otherwise known to the wireless communication device(s) 512. In one embodiment, prioritization of the intelligent reporting process at the wireless communication device(s) 512 may be ensured by providing the wireless communication device(s) 512 with priority levels for the reporting process so that lower priority processes can, if needed, be deferred or interrupted in favor of higher priority processes.
The network node may then receive the activated intelligent RRM report(s) from the wireless communication device(s) for which it(they) are activated and perform one or more RRM actions based on these reports (step 610).
The wireless communication device 512 receives, from a network node, a request to start intelligent RRM reporting and/or a request to start training a model (e.g., a ML model such as, e.g., a neural network) for intelligent RRM reporting (step 704). Optionally, the message(s) received by the wireless communication device 512 to activate the intelligent RRM reporting may include a triggering criterion that describes when to activate the intelligent RRM report. Optionally, the message(s) may include information that indicates one or more time-windows for which the wireless communication device 512 is to produce intelligent RRM reports. Optionally, the message(s) may include information that indicates that the wireless communication device 512 is to start training a model (e.g. a ML model such as, e.g., a neural network) for intelligent RRM reporting. For example, the network node may only activate the intelligent RRM report(s) during a time period(s) when such reporting is useful where the time period(s) can be based on, e.g., networking scheduling and/or UE service type. Further, in one embodiment, intelligent RRM reporting is triggered when it fulfills a certain accuracy requirement. The accuracy requirement may be configured by the network node or predefined or otherwise known to the wireless communication device 512. In one embodiment, prioritization of the intelligent reporting process at the wireless communication device 512 may be ensured by priority information provided to the wireless communication device 512 (e.g., in step 704) that defines priority levels for the reporting process so that lower priority processes can, if needed, be deferred or interrupted in favor of higher priority processes.
The wireless communication device 512 performs one or more actions in response to the request received in step 704. More specifically, the wireless communication device 512 generates and sends an intelligent RRM report(s) to the network node in accordance with the received request of step 704 (step 706). Notably, as will be understood by those of skill in the art, in one embodiment, the request received in step 704 includes a request to train the machine-learning model used to produce the predicted values for the intelligent RRM report(s) prior to generating and sending the intelligent RRM report(s) in step 706. As such, in one embodiment, the one or more actions performed by the wireless communication device 512 in response to the request received in step 704 includes training the machine-learning model. The wireless communication device 512 may receive an update to information related to intelligent RRM reporting (e.g., an update to the triggering criterion, an update to deactivate intelligent RRM reporting, or the like) (step 708). The wireless communication device 512 then proceeds in accordance with the update.
Now, a discussion of numerous aspects that are related to the processes of
I. When to Activate Intelligent RRM Capabilities Reporting?
Requesting intelligent RRM reporting capabilities of a wireless communication device 512 (e.g., in step 602 of
A. Performance Based
In one embodiment, the network node decides (e.g., in step 600 of
B. Energy Info.
In one embodiment, the network node decides (e.g., in step 600 of
C. Network Information
In one embodiment, the network node decides (e.g., in step 600 of
D. Base Station Capabilities
In one embodiment, the network node decides (e.g., in step 600 of
E. UE Info
In one embodiment, the network node decides (e.g., in step 600 of
F. Historical Information
In one embodiment, the network node decides (e.g., in step 600 of
II. Selecting and Activating Intelligent RRM Reporting
A. Select a Number of Wireless Communication Devices for Intelligent Reporting
The network node (e.g., base station 502 such as, e.g., a gNB) can select a certain number of UEs to use intelligent RRM reporting (e.g., in step 606 of
It should be noted that wireless communication devices 512 can be configured to report prediction information while being in RRC_Connected or while being in other states, e.g. RRC_Idle or RRC_Inactive.
RRC_Connected: The serving network node is aware and in control of the process of configuration and reporting of prediction information. The serving network node can therefore select, configure, de-configure and re-configure a wireless communication device 512 as per criteria described herein.
RRC_Idle and RRC_Inactive: The serving network node may select and configure a wireless communication device 512 for intelligent RRM reporting; however the wireless communication device 512 might move to the coverage area of other network nodes while being configured with the intelligent RRM reporting configuration. In this case, there are two options to handle the intelligent RRM reporting process:
B. Triggering Criterion
The network node can signal a triggering criterion to the wireless communication device 512 describing when the wireless communication device 512 should activate intelligent RRM reporting. This may be in case of replacing measurements with predictions or in case or reporting both measured values and predicted events. This triggering criterion may be signaled to the wireless communication device 512 in step 608 of
In another example, in case of future signal quality prediction, the wireless communication device 512 may be configured with a triggering criterion such that the wireless communication device 512 only sends the predictions if its prediction performance is within a certain threshold range for the last x seconds, e.g., for a specific reference signal. The threshold and time window could be configured by the network. In case of a future value predicted, the network node can request a predicted value for a set of future time instances, for example based on its scheduling state.
In another embodiment, the network node triggers a periodical reporting of the predicted value at the wireless communication device 512. In other words, the triggering criterion may be a time-based criterion that triggers periodical reporting of the predicted value.
At the time of triggering an intelligent reporting process, the network node can also assign to the wireless communication device 512 a priority for the prediction process (e.g., in step 608 of
In another embodiment, the network node can trigger the reporting of predicted values based various aspects such as:
III. Intelligent RRM Report
The wireless communication device 512 can signal its capabilities in predicting a certain quantity (e.g., in step 604 of
A. Measurement Replacement
An intelligent wireless communication device 512 (i.e., a wireless communication device 512 capable of intelligent RRM reporting) can reduce its need for beam measurements and, for example, use a subset of beam measurements and predict the rest. For example, the wireless communication device 512 could measure 2-beams, but provide e.g. 3 beam-reports (e.g., in step 706 of
B. Future Traffic
The network node can request (e.g., in step 608 of
C. Other Type of Information
The capabilities (e.g., reported by the wireless communication device in step 702 of
IV. CU-CP and DU-CP Aspects
In one embodiment, the network node in charge of managing Intelligent Reporting from the wireless communication device 512 is a gNB-CU. However, it is possible for an associated gNB-DU to task the gNB-CU with configuration of specific Intelligent Reporting processes that might be beneficial for the gNB-DU 1602. In such case, as illustrated in
An example of this embodiment could include the gNB-DU requesting predictions on beam level coverage for the served beams and for neighbor cell beams. The gNB-DU may use such predictions to optimize management of its served beams and for example to shape or extend or reduce beam coverage or to increase or reduce the number of beams serving a given area.
V. Additional Description
of a network node, base station 502/506, gNB, gNB-CU, gNB-DU, or the like, as described herein) are implemented at the one or more processing nodes 1200 or distributed across the one or more processing nodes 1200 and the control system 1102 and/or the radio unit(s) 1110 in any desired manner. In some particular embodiments, some or all of the functions 1210 of the network node 1100 described herein are implemented as virtual components executed by one or more virtual machines implemented in a virtual environment(s) hosted by the processing node(s) 1200. As will be appreciated by one of ordinary skill in the art, additional signaling or communication between the processing node(s) 1200 and the control system 1102 is used in order to carry out at least some of the desired functions 1210. Notably, in some embodiments, the control system 1102 may not be included, in which case the radio unit(s) 1110 communicate directly with the processing node(s) 1200 via an appropriate network interface(s).
In some embodiments, a computer program including instructions which, when executed by at least one processor, causes the at least one processor to carry out the functionality of the network node 1100 or a node (e.g., a processing node 1200) implementing one or more of the functions 1210 of the network node 1100 in a virtual environment according to any of the embodiments described herein is provided. In some embodiments, a carrier comprising the aforementioned computer program product is provided. The carrier is one of an electronic signal, an optical signal, a radio signal, or a computer readable storage medium (e.g., a non-transitory computer readable medium such as memory).
In some embodiments, a computer program including instructions which, when executed by at least one processor, causes the at least one processor to carry out the functionality of the wireless communication device 1400 according to any of the embodiments described herein is provided. In some embodiments, a carrier comprising the aforementioned computer program product is provided. The carrier is one of an electronic signal, an optical signal, a radio signal, or a computer readable storage medium (e.g., a non-transitory computer readable medium such as memory).
Some example embodiments of the present disclosure are as follows:
Embodiment 1: A method performed by a wireless communication device (512), the method comprising one or more of the following actions:
Embodiment 2: The method of embodiment 1 wherein performing the one or more actions comprises:
Embodiment 3: The method of embodiment 2 wherein the one or more reports further comprise information that indicates an accuracy or confidence level of the one or more predicted values.
Embodiment 4: The method of any of embodiments 1 to 3 wherein reporting of predicted values is activated when a triggering criterion is satisfied.
Embodiment 5: The method of embodiment 4 wherein the triggering criterion is a required accuracy level for the one or more predicted values.
Embodiment 6: The method of embodiment 4 wherein the triggering criterion is a time-based triggering criterion.
Embodiment 7: The method of embodiment 4 wherein the triggering criterion is a prediction performance based triggering criterion.
Embodiment 8: The method of any of embodiments 1 to 7 wherein the request comprises a request to start reporting predicted values at a particular time(s) or during a particular time window(s).
Embodiment 9: The method of any of embodiments 1 to 8 wherein the predicted values comprise predicted RRM related values.
Embodiment 10: The method of any of embodiments 1 to 8 wherein the predicted values comprise predicted beam related values.
Embodiment 11: The method of any of embodiments 1 to 8 wherein the predicted values comprise predicted measurement values for one or more frequencies.
Embodiment 12: The method of any of embodiments 1 to 8 wherein the predicted values comprise predicted values for future traffic needs of the wireless communication device (512).
Embodiment 13: The method of any of embodiments 1 to 12 further comprising:
Embodiment 14: A wireless communication device (512) adapted to perform the method of any of embodiments 1 to 13.
Embodiment 15: A wireless communication device (512; 1400) comprising:
Embodiment 16: A method performed by a network node (1100; 502), the method comprising one or more of the following actions:
Embodiment 17: The method of embodiment 16 further comprising receiving (710) one or more reports from the one or more wireless communication devices (512), the one or more reports comprising predicted values.
Embodiment 18: The method of embodiment 17 further comprising performing one or more actions (e.g., one or more RRM related actions) based on the one or more reports.
Embodiment 19: The method of embodiment 17 or 18 wherein the one or more reports further comprise information that indicates an accuracy or confidence level of the predicted values.
Embodiment 20: The method of any of embodiments 17 to 19 further comprising sending (708), to the one or more wireless communication devices (512), a triggering criterion that defines when reporting of predicted values is to be activated at the one or more wireless communication devices (512).
Embodiment 21: The method of embodiment 20 wherein the triggering criterion is a required accuracy level for the one or more predicted values.
Embodiment 22: The method of embodiment 20 wherein the triggering criterion is a time-based triggering criterion.
Embodiment 23: The method of embodiment 20 wherein the triggering criterion is a prediction performance based triggering criterion.
Embodiment 24: The method of any of embodiments 16 to 23 wherein the request comprises a request to start reporting predicted values at a particular time(s) or during a particular time window(s).
Embodiment 25: The method of any of embodiments 16 to 24 wherein the predicted values comprise predicted RRM related values.
Embodiment 26: The method of any of embodiments 16 to 24 wherein the predicted values comprise predicted beam related values.
Embodiment 27: The method of any of embodiments 16 to 24 wherein the predicted values comprise predicted measurement values for one or more frequencies.
Embodiment 28: The method of any of embodiments 16 to 24 wherein the predicted values comprise predicted values for future traffic needs of the wireless communication device (512).
Embodiment 29: The method of any of embodiments 16 to 28 further comprising sending (602), to the plurality of wireless communication devices (512), a request for the one or more capabilities of the wireless communication device (512) for reporting of predicted values.
Embodiment 30: A network node (1100; 502) adapted to perform the method of any of embodiments 16 to 29.
Embodiment 31: A network node (1100; 502) comprising processing circuitry (1104; 1204) configured to cause the network node (1100; 502) to perform the method of any of embodiments 16 to 29.
Any appropriate steps, methods, features, functions, or benefits disclosed herein may be performed through one or more functional units or modules of one or more virtual apparatuses. Each virtual apparatus may comprise a number of these functional units. These functional units may be implemented via processing circuitry, which may include one or more microprocessor or microcontrollers, as well as other digital hardware, which may include Digital Signal Processor (DSPs), special-purpose digital logic, and the like. The processing circuitry may be configured to execute program code stored in memory, which may include one or several types of memory such as Read Only Memory (ROM), Random Access Memory (RAM), cache memory, flash memory devices, optical storage devices, etc. Program code stored in memory includes program instructions for executing one or more telecommunications and/or data communications protocols as well as instructions for carrying out one or more of the techniques described herein. In some implementations, the processing circuitry may be used to cause the respective functional unit to perform corresponding functions according one or more embodiments of the present disclosure. While processes in the figures may show a particular order of operations performed by certain
embodiments of the present disclosure, it should be understood that such order is exemplary (e.g., alternative embodiments may perform the operations in a different order, combine certain operations, overlap certain operations, etc.).
Those skilled in the art will recognize improvements and modifications to the embodiments of the present disclosure. All such improvements and modifications are considered within the scope of the concepts disclosed herein.
VI. Additional Aspects
The following pages of the detailed description reproduce text of discussion papers prepared for 3GPP meeting #110-e. This text was included as an appendix to the priority founding application, provisional patent application Ser. No. 63/094,698, filed Oct. 21, 2020. Up to this section, the present disclosure has focused primarily on utilization machine learning to predict values for RRM use cases. However, there are a number of other use cases in which the methods of the present disclosure for generation of predicted values for machine-learned network optimization (e.g., as described with regards to
VI.A. AI/ML based Use Cases
1 Introduction
As described in RP-201620, the study on AI/ML in RAN3 will focus on the following:
In order to explore the areas where AI/ML is most applicable and can improve the network performance for the NG RAN, this paper illustrates use cases that can be taken as reference during the development of AI/ML based solutions.
2 AI/ML Use Cases
It is important to fully utilize the potentials in AI/ML for wireless networks, for example by extracting important data from the system in order to build advanced AI/ML models.
One problem in enabling AI/ML for wireless networks is the variable cost depending on wired or over-the-air data transfer. Enabling AI/ML by extending the UE reporting over-the-air by including different types of information, from radio to physical measurements would lead to increased signalling. The trade-off between increased data signalling versus enabling improved intelligence at the network is a challenging problem. It is important to fully address such trade-offs when evaluating different AI/ML use cases in the SI. One alternative to extending the UE report of radio or physical measurements is to explore the use of potential augmented information provided by the UE, for example generated by an AI-model. This information may be given as input to AI models hosted in the network, hence creating a system where AI models interact between each other to produce the desired final output.
Proposal 1: Explore Potential Augmented Information from the UE and from the RAN in Each Use Case.
Next, use cases covering important areas where AI/ML is likely to improve network performance is described. The use cases are classified in the following families:
2.1 AI/ML for Traffic Steering
AI/ML can be applied to steer traffic more efficiently, both in terms of capacity and energy efficiency.
2.1.1 Reward Information for AI/ML-Based Handovers
Finding the best cell or set of cells to serve a UE is a challenging task due to the densification of networks and introduction of new frequency bands. One of the challenges in finding the best cell for a UE is to evaluate if the new cell was better than a previous serving cell for the UE, hence, it would be beneficial to have richer feedback information available from the new serving cell, so to compare previous and current serving cell performance.
Considering the current handover mechanisms in NR, after a handover to the target cell, the source/serving node would act obliviously about the handed over UE i.e. it would not be interested on that UE any longer. Therefore, if the UE experiences low throughput or poor radio coverage once handed over to the target cell, the source node of the handover would not be able to recognize and take any counteraction preventing such handovers causing poor performance for the UE. It is thus important to design a solution enabling a feedback mechanism after handover, where the UE and the target node provide measurements relative to the performance of the target cell serving the UE. This can enable the source node to update its handover decisions frequently based on the received feedback from target node (which would comprise also feedback from the UE while at target). The feedback from the target could be used as reward information for an AI/ML function that performs handover decisions, one such function could comprise reinforcement learning. Handover decisions consist of a prediction that could take into account possible future performance for a UE once handed over to a certain target cell/node. The feedback provided from target RAN node to source could comprise of:
Proposal 2: Investigate Potential Reward Information for Enabling AI/ML Based Traffic Steering
2.1.2 Traffic Steering Augmented Information
In addition to the reward information provided by the target RAN node, the potential target RAN node could also signal augmented information as illustrated in the message sequence chart below, generated by an ML-model for improved traffic steering, for example its future load information. The predicted future load information can comprise
The UE may also provide augmented information such as its predicted mobility pattern and feed this to the target RAN, which in turn will forward it to the source RAN. Similarly, the serving gNB can provide the target gNB with augmented information related to the UE at handover, for example the predicted UE mobility or traffic.
Proposal 3: Augmented Information Related to Improved Traffic Steering should be Investigated
2.1.3 AI/ML for Energy Efficiency
Energy efficiency is an important aspect in wireless communications networks. One method for providing energy saving is to put capacity cells into a sleep mode. The activation or deactivation of a capacity cell may be triggered from a gNB that provides basic coverage as illustrated in the picture below and is typically a trade-off between energy efficiency and capacity.
In cases when there is quite low traffic around the capacity cell, it may be more energy efficient to turn off the capacity cell until the load increases. The capacity cell may later be activated when the traffic is higher and when there are UEs in the vicinity of the capacity cell which may be moved into the capacity cell by a handover procedure or some other connectivity reconfiguration procedure. However, it may be quite tricky to find out whether or not the communications UEs served by the basic coverage cell may be served by the capacity cell without activating the capacity cell. This means that in some situations when the load increases, the capacity cell is activated in order to determine whether or not one or more UEs served by the basic coverage cell may be served by the capacity cell. In case no such UEs would connect (or it would connect with acceptable radio conditions) to the activated capacity cell, the activation is done in vain, hence leading to a waste of energy.
Furthermore, a capacity cell is often deployed in the handover region of two basic coverage cells, and therefore it is difficult to optimize capacity versus energy consumption. It is important to also look into energy saving application using ML/AI in activating capacity cells efficiently, for example to activate capacity cells based on predictions on traffic that could be offloaded to the capacity cell for all relevant nodes in the network. The signaling of such predictions to the RAN node controlling the activation or the signaling of information that may help to derive a prediction of offloaded traffic to capacity cell, should be investigated. It is also important to investigate whether the UE can provide augmented information to enable a smarter capacity cell activation.
Proposal 4: Energy Efficiency should be Studied, for Example AI/ML for Capacity Cell Activation
2.2 AI/ML for QoS Prediction
Quality of service (QoS) describes the overall performance of a service, for example the latency, reliability or throughput. Service Level Agreements (SLAs) are contractual agreements between an operator and an incumbent for the provisioning of services with a given set of performance requirements. On the basis of the current and predicted QoS target of each served UE, it is possible to determine if SLAs are going to be met. The system in charge for checking fulfillment of SLAs is the OAM. In order to enable better SLA fulfillment prediction at the OAM, one should look into AI/ML in order to provide augmented information helping to forecast SLA fulfilment.
Using AI/ML, the CU-CP can for example predict whether for a group of UEs and services (e.g. for UEs in a certain network slice using a service with 5QI==x) the target QoS requirements will be fulfilled or not. Such prediction can be relative to a specific time window into the future.
Such augmented information can also comprise non-UE specific information, such as a prediction of the expected load per QoS class for a particular time of the day, as well as a prediction of whether QoS requirements for such QoS classes can be fulfilled. The QoS fulfillment prediction could be signaled from the RAN to the OAM upon request from the OAM. The request could also comprise a request for the predicted QoS for a certain type of UE, for example a highly mobile UE or a low-end UE (e.g. IoT).
The OAM receiving such QoS fulfillment prediction can in turn derive whether SLAs can be fulfilled in the future. If for example the OAM determines that SLAs cannot be fulfilled in the future, the OAM can take preventive actions such as to reconfigure resource partition policies per slice at the RAN in order to ensure that the SLAs not fulfilled can be fulfilled by means of a higher amount of resources to be utilized. The general framework is illustrated in the flowchart below.
The augmented information sent to the OAM can be used to change the slice configuration, for example allocate more resources if SLA is predicted to not be fulfilled in a future time window.
Proposal 5: AI/ML for Predicting QoS and SLA Fulfilment should be Studied
2.3 AI/ML for Improved Radio Resource Management (RRM)
The use of AI/ML can provide an improved performance by leveraging new capabilities in learning complex interactions in the environment, one such environment with complex interactions is RRM. Potential RRM algorithms comprise, link-adaptation, rank-selection, power control, mobility decisions. The SI should investigate potential augmented information from UEs or gNBs in order to enable an even better RRM. The augmented information generated by an AI-model could for example comprise forecast values such as the predicted load in a future time frame for one RAN node, or a UE predicted future signal quality value.
As an example, the use case of link adaptation can be considered. Link adaptation is a function that needs to react to rather fast changes of radio conditions. A way to improve the performance of link adaptation would be to gain more granular information about the radio environment and to predict the optimal link adaptation configuration on the basis of a prediction of the radio conditions.
In order to enhance link adaptation performance the UE may provide higher granularity data to the serving RAN, such as more granular L1 measurements, measurements of UE speed, UL queuing delays.
At the same time the serving RAN may receive from neighbor nodes information about cross cell interference, e.g. in the form of number of UEs or resource utilization at cell edge, or indeed information either constituting or helping to extrapolate a prediction of cross cell interference.
With such information the serving RAN is able to derive a prediction of the channel condition for the UE and therefore to adopt a better link adaptation configuration.
Proposal 6: Investigate new AI/ML-based augmented information for improved RRM
VI.A. AI/ML Based Use Cases
1 Introduction
A new SI has been approved in RP-201620: “Enhancement for data collection for NR and ENDO”. As specified in the SID, the study is tasked to address the following objective:
In R3-20xxxx a number of AI/ML use cases were described. The Use Cases could be classified as follows:
This paper addresses the potential Standardization Impact of the Use Cases analyzed.
2 Standardization Impacts on a Per Use Case Class
2.1 Standardisation Impacts of AI/ML for Traffic Steering—for Capacity and Energy Efficiency
This class of Use Cases relies on the ability of the RAN to predict the best cell that will serve the UE. The Use Cases can include mobility scenarios triggered by various reasons (e.g. Energy Efficiency, radio conditions, load conditions) or multi connectivity scenarios (e.g. prediction of best PSCell). In general the use cases provide augmented information about the cell that, given the predicted conditions, will best serve the UE within a future time window.
In this class of Use Cases the main standardization impacts are foreseen to be on the following:
Conclusion 1: The Use Case family of “AI/ML for traffic steering” may generate the following impacts:
2.2 Standardisation Impacts of AI/ML for QoS Prediction
This class of Use Cases relies on the interaction between the RAN and the OAM system. In this class of Use Cases the RAN provides augmented information to the OAM concerning predictions of QoS levels.
Such QoS level predictions may consist of predictions of one or more QoS parameters forming the QoS profile of each bearer at a UE. While it might be considered that predictions could be derived on a per UE per bearer basis, it appears that the amount of information and predictions generated in this case may be overwhelming, as well as the computational effort to derive such number of predications. Instead, an equally effective approach with a lower burden on processing and storage could be that of deriving QoS predictions on a per QoS class basis. For example, QoS prediction could be derived on a per slice and per 5QI granularity.
In this class of Use Cases the main standardization impacts are foreseen to be on the following:
Conclusion 2: The Use Case family of “Standardization Impacts of AI/ML for QoS monitoring” may generate the following impacts:
2.3 Standardisation Impacts of AI/ML for Improved Radio Resource Management
In this class of scenarios it is possible to group all scenarios based on AI/ML model hosting at the RAN, so to allow for optimization of RRM processes via a fast control loop. The output of the AI/ML models in this family are prediction parameters that can be used when applying radio resource management. An example of such input could be a prediction of link adaptation configurations.
The RAN has today a very rich set of information that allow for good configuration of radio resource policies. However, there are information currently missing at the RAN, especially concerning the “view” UEs have of the surrounding conditions.
In this class of Use Cases the main standardization impacts are foreseen to be on the following:
Conclusion 3: The Use Case family of “AI/ML for improved radio resource management” may generate the following impacts:
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2021/079263 | 10/21/2021 | WO |
Number | Date | Country | |
---|---|---|---|
63094698 | Oct 2020 | US |