Methods for ai driven optimization of intents and resource utilization

Abstract

According to one aspect, a computer-implemented method, executed by a first network node in a communication network environment, for optimizing transmission parameters for at least one user device is provided. The method includes obtaining a first set of at least one intent for at least one user device. The method includes obtaining at least one intent preference value associated with the at least one user device based on the obtained first set of at least one intent. The method includes determining, using a first AI / ML model, a communication configuration for the at least one user device based on the at least one intent preference value. The method includes causing, based on the communication configuration, a change to a configuration of the at least one user device.

Description

METHODS FOR Al DRIVEN OPTIMIZATION OF INTENTS AND RESOURCE UTILIZATIONTECHNICAL FIELD

[0001] Disclosed are embodiments related to optimization of intents and resource utilization using artificial intelligence ("Al”) and machine learning ("ML”) technology.BACKGROUND

[0002] FIG. 1 illustrates 5G RAN (NG-RAN) architecture. The 5G RAN architecture is described in the Third Generation Partnership Project's ("3GPP”) Technical Specification (“TS”) 38.401 v. 15.8.0. The NG-RAN 102 consists of a set of gNBs (104A, 104B) connected to the 5GC (100) through the NG interface. As specified in 3GPP TS 38.300, v. 17.5.0, the NG-RAN 102 could also consist of a set of ng-eNBs, which may consist of an ng-eNB-CU and one or more ng-eNB-DU(s). An ng-eNB-CU and an ng-eNB-DU is connected via W1 interface. A gNB can support FDD mode, TDD mode or dual mode operation. gNBs can be interconnected through the Xn interface.

[0003] A gNB 104A, 104B may consist of a gNB-CU 106 and one or more gNB-DU(s) 108A-B. A gNB-CU and a gNB-DU is connected via F1 interface. One gNB-DU is connected to only one gNB-CU. NG, Xn and F1 are logical interfaces. For NG-RAN, the NG and Xn-C interfaces for a gNB consisting of a gNB-CU and gNB-DUs terminate in the gNB-CU. For EN-DC, the S1-U and X2-C interfaces for a gNB consisting of a gNB-CU and gNB-DUs, terminate in the gNB-CU. The gNB-CU and connected gNB-DUs are only visible to other gNBs and the 5GC as a gNB.

[0004] FIG. 2 illustrates an architecture for separation of gNB-CU-CP and gNB-CU-UP. FIG. 2 includes a gNB-CU- CP 206, gNB-CU-UP 210, and gNB-DU 208A-B.

[0005] A Study Item (SI) "Enhancement for Data Collection for NR and EN-DC” is defined in 3GPP RP-201620. The study item aims to study the functional framework for RAN intelligence enabled by further enhancement of data collection through use cases, examples etc. and identify the potential standardization impacts on current NG-RAN nodes and interfaces.

[0006] The detailed objectives of the SI are listed as follows: Study high level principles for RAN intelligence enabled by Al, the functional framework (e.g., the Al functionality and the input / output of the component for Al enabled optimization) and identify the benefits of Al enabled NG-RAN through possible use cases e.g., energy saving, load balancing, mobility management, coverage optimization, etc.:

[0007] (a) Study standardization impacts for the identified use cases including: the data that may be needed by an Al function as input and data that may be produced by an Al function as output, which is interpretable for multi-vendor support.

[0008] (b) Study standardization impacts on the node or function in current NG-RAN architecture to receive / provide the input / output data.

[0009] (c) Study standardization impacts on the network interface(s) to convey the input / output data among network nodes or Al functions.

[0010] As part of the SI work, an interim version of 3GPP Technical Report (TR) 37.817, v. 17.0.0 is available post RAN3#114-bis-e and indicates the text below.

[0011] The following high-level principles should be applied for Al-enabled RAN intelligence:

[0012] The detailed AI / ML algorithms and models for use cases are implementation specific and out of RAN3 scope.

[0013] The study focuses on AI / ML functionality and corresponding types of inputs / outputs.

[0014] The input / output and the location of the Model Training and Model Inference function should be studied case by case.

[0015] The study focuses on the analysis of data needed at the Model Training function from Data Collection, while the aspects of how the Model Training function uses inputs to train a model are out of RAN3 scope.

[0016] The study focuses on the analysis of data needed at the Model Inference function from Data Collection, while the aspects of how the Model Inference function uses inputs to derive outputs are out of RAN3 scope.

[0017] Where AI / ML functionality resides within the current RAN architecture, depends on deployment and on the specific use cases.

[0018] The Model Training and Model Inference functions should be able to request, if needed, specific information to be used to train or execute the AI / ML algorithm and to avoid reception of unnecessary information. The nature of such information depends on the use case and on the AI / ML algorithm.

[0019] The Model Inference function should signal the outputs of the model only to nodes that have explicitly requested them (e.g., via subscription), or nodes that take actions based on the output from Model Inference.

[0020] An AI / ML model used in a Model Inference function has to be initially trained, validated and tested before deployment.

[0021] The Functional Framework for RAN Intelligence is shown in FIG. 3. Data Collection 302 is a function that provides input data to Model training 304 and Model inference 306 functions. AI / ML algorithm specific data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) is not carried out in the Data Collection function. Examples of input data may include measurements from UEs or different network entities, feedback from Actor, output from an AI / ML model. Training Data includes data needed as input for the AI / ML Model Training function. Inference Data includes data needed as input for the AI / ML Model Inference function.

[0022] Model Training 304 is a function that performs the ML model training, validation, and testing which may generate model performance metrics as part of the model testing procedure. The Model Training function is also responsible for data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) based on Training Data delivered by a Data Collection function, if required. Model Deployment / Update is used to initially deploya trained, validated, and tested AI / ML model to the Model Inference function or to deliver an updated model to the Model Inference function.

[0023] Model Inference 306 is a function that provides AI / ML model inference output (e.g. predictions or decisions). Model Inference function may provide model performance feedback to Model Training function when applicable. The Model inference function is also responsible for data preparation (e.g. data pre-processing and cleaning, formatting, and transformation) based on Inference Data delivered by a Data Collection function, if required. Output includes the inference output of the AI / ML model produced by a Model Inference function. Details of inference output are use case specific. Model Performance Feedback may be used for monitoring the performance of the AI / ML model.

[0024] Actor 308 is a function that receives the output from the Model inference function and triggers or performs corresponding actions. The Actor may trigger actions directed to other entities or to itself. Feedback includes information that may be needed to derive training or inference data or performance feedback.

[0025] Three use cases have been identified for RAN Intelligence, namely: Network Energy Saving, Load Balancing and Mobility Optimization and the potential standard impacts are described in the same TR 37.817 as follows.

[0026] For Network energy saving use case: AI / ML Model Training is located in the Operations, Administration, and Maintenance function (0AM) and AI / ML Model Inference is located in the gNB; and AI / ML Model Training and AI / ML Model Inference are both located in the gNB. Note: gNB is also allowed to continue model training based on AI / ML model trained in the 0AM. In case of CU-DU split architecture, the following solutions are possible: AI / ML Model Training is located in the 0AM and AI / ML Model Inference is located in the gNB-CU; or AI / ML Model Training and Model Inference are both located in the gNB-CU.

[0027] For the Mobility Optimization use case, the following two options are considered for the locations of AI / ML Model Training and AI / ML Model Inference: The AI / ML Model Training function is deployed in 0AM, while the Model Inference function resides within the RAN node; or both the AI / ML Model Training function and the AI / ML Model Inference function reside within the RAN node. Furthermore, for CU-DU split scenario, following option is possible: AI / ML Model Training is located in CU-CP or CAM, and AI / ML Model Inference function is located in CU-CP. Note: the gNB is also allowed to continue model training based on AI / ML model trained in the CAM.

[0028] For the Load Balancing use case, the following solutions can be considered for supporting AI / ML-based load balancing: AI / ML Model Training is located in the CAM and AI / ML Model Inference is located in the gNB, or AI / ML Model Training and AI / ML Model Inference are both located in the gNB. In the case of CU-DU split architecture, the following solutions are possible: AI / ML Model Training is located in the CAM and AI / ML Model Inference is located in the gNB-CU or AI / ML Model Training and Model Inference are both located in the gNB-CU. Note: gNB is also allowed to continue model training based on AI / ML model trained in the CAM.

[0029] At RAN3#114-e meeting, contribution R3-215244 has proposed to introduce a model management function in the Functional Framework for RAN Intelligence, as reported below.

[0030] FIG. 4 illustrates a functional framework for RAN Intelligence. As shown in FIG. 4, Model deployment / update may be decided by model management 410 instead of model training 404. The model management 410 may also host a model repository. The model deployment / update should be performed by model management. Model performance monitoring is a key function to assist and control model inference 406. The model performance feedback from model inference 406 should be first sent to model management 410. If the performance is not ideal, the model management 410 may decide to fallback to traditional algorithm or change / update the model.

[0031] In this scenario, the model training should be also controlled by model management. The model management function may be taken by either GAM or CU or other network entities depending on the use cases. Clearly defining a model management function is useful for future signaling design and analysis.

[0032] Model management function supports following roles: Requesting model training and receiving the model training result; Model deployment / updates for inference; Model performance monitoring, including receiving performance feedback from model inference and taking necessary action, e.g., keep the model, fallback to traditional algorithm, change or update the model; and model storageSUMMARY

[0033] Designing generalizable Al applications for communication networks requires considering the diverse intents driving network functions. In RAN systems, performance is not defined by a single key performance indicator (KPI), but by multiple quantifiable and often competing KPIs. The 3GPP TS 28.312, e.g. v. 18.0.0, for instance, describes the concept of intent as a tool to specify objectives, requirements, and constraints for a service or network management workflow. Hence, RAN control problems, like radio resource management (RRM), including user and resource scheduling, do not have a single optimum-for-RAN KPI but a Pareto frontier of RAN KPIs, each trading off against another. An example is the tradeoff between optimization of spectral efficiency or throughput against transmission reliability or latency.

[0034] Most Al research in communication networks proposes algorithms that optimize a specific intent function at runtime, in particular for the allocation of radio resources and link adaptation for user devices. Addressing different intents at runtime would require training multiple models for each RAN functionality (e.g., user scheduling or link adaptation), which poses significant challenges. Firstly, developing unique Al models for each RAN function (e.g., link adaptation) to optimize a specific intent, such as a specific reward function in reinforcement learning methods, necessitates training and managing multiple Al models to address and optimize the performance for different intents (e.g., reward functions) at runtime. This is costly and impractical, even with a fully reliable and automated system. Secondly, specialized Al model to optimize a certain intent reduces the system robustness for handling and adapting to changes that may render achieving a certain intent infeasible. Consequently, the system may be driven to operate in suboptimal configuration, potentially leading to overprovisioning of resources and performance loss.

[0035] Aspects of the present disclosure provide a method for a first network node for optimizing transmission parameters in downlink or uplink for at least a user device based on a plurality of at least one intent associated to the user device. Aspects of the present disclosure further provide a method for training and executing a single AI / ML model that at runtime can select transmission parameters for the user device to optimize multiple intents, thereby improving system performance and scalability. At runtime, the intent used to select transmission parameters can be user specific or user-group specific. An intent associated to a user device, or a group of user devices can be represented, for instance, by an objective function, utility function, performance metrics, key performance indicators, associated to an individual, as well as quality of service requirements, quality of experience requirements. When multiple intents are associated to a user device or a group of user device, the method offers a solution to find the optimal or a near-optimal tradeoff to fulfil the intents.

[0036] One advantage of aspects of the present disclosure is to enable the first network node to dynamically optimize and adapt the configuration of multiple transmission parameters for a user device, such as allocation of radio resources and link adaptation parameters, to one or more communication intents or objectives associated to the user device. Therefore, system resources can be optimized to meet the needs of individual user devices, resulting in improved system spectral efficiency and system performance.

[0037] A further advantage of aspects of the present disclosure is to enable the first network node use a single solution, such as a single AI / ML model, to optimize multiple, often competing, communication intents for each individual user device. This enables to reduce the complexity of the first network node compared to the case where multiple individual solutions are required to optimize different communication intents of a user device. Therefore, the first network node may more efficiently determine the configuration of multiple transmission parameters for a user device enabling, for instance, energy savings.

[0038] According to one aspect, a computer-implemented method, executed by a first network node in a communication network environment, for optimizing transmission parameters for at least one user device is provided. The method includes obtaining a first set of at least one intent for at least one user device. The method includes obtaining at least one intent preference value associated with the at least one user device based on the obtained first set of at least one intent. The method includes determining, using a first AI / ML model, a communication configuration for the at least one user device based on the at least one intent preference value. The method includes causing, based on the communication configuration, a change to a configuration of the at least one user device.

[0039] In some embodiments, the first set of at least one intent comprises one or more: objective functions, utility functions, performance metrics, key performance indicators, KPIs, quality of service, QoS, requirements, or quality of experience, QoE, requirements.

[0040] In some embodiments, the objective functions, utility functions, performance metrics, and KPIs comprise one or more of: user throughput, user spectral efficiency, transmitted information per packet, resource utilization or allocation, latency, block error rate, bit error rate, reliability, packet error rate, energy consumption, energy savings,discontinuous transmission, discontinuous reception, or signal quality, and wherein the QoS or QoE requirements comprise one or more of: a throughput requirement, a spectral efficiency requirement, a transmission latency requirement, a block error rate requirement, a packet error rate requirement, a bit error rate requirement, a packet error rate requirement, a reliability requirement, a signal to noise ration requirement, a resource utilization requirement, an energy consumption requirement, an energy savings requirement, a resource utilization requirement, a discontinuous transmission requirement, a discontinuous reception requirement, a transport block size requirement, or a resource element requirement.

[0041] In some embodiments, the communication configuration comprises one or more of: an allocation or configuration of at least one communication resource, an allocation or configuration of link adaptation parameters, an allocation or configuration of information bits to be received or transmitted by the at least one user device, an allocation or configuration of downlink or uplink reference signals, a configuration for discontinuous transmission or reception, or a configuration for energy saving.

[0042] In some embodiments, the first set of at least one intent is contained in a first message transmitted by the at least one user device or a second network node.

[0043] In some embodiments, the method further includes transmitting a second message towards the at least one user device or the second network node, the second message comprising a request for the first set of at least one intent.

[0044] In some embodiments, the method further includes transmitting a third message towards the at least one user device, the third message comprising the communication configuration.

[0045] In some embodiments, the obtaining the at least one intent preference value comprises mapping the first set of at least one intent to the at least one intent preference value. In some embodiments, the mapping comprises mapping the first set of at least one intent to one or more user performance metrics parameterized with respect to one or more intent preference values. In some embodiments, the mapping is performed by obtaining the user performance metrics parameterized with respect to one or more intent preference values by testing the first AI / ML model over a range of values assumed for the one or more intent preference values. In some embodiments, the mapping is performed by a second AI / ML model.

[0046] In some embodiments, the first set of at least one intent comprises a first intent and a second intent, and wherein the at least one intent preference value comprises a first intent preference value associated with the first intent and a second intent preference value associated with the second intent. In some embodiments, the first intent and the second intent correspond to a first user device. In some embodiments, the first intent corresponds to a first user device and the second intent corresponds to a second user device different than the first user device. In some embodiments, the first set of at least one intent comprises at least one intent relating to transmission reliability, throughput, spectral efficiency, block error rate, or latency. In some embodiments, the first intent corresponds to a number of information bits carried by a packet transmission to or from the at least one user device, the second intent corresponds to a totalnumber of resource elements used for communicating with the at least one user device until an nthtransmission attempt of the packet, and the communication configuration comprises one or more link adaptation parameters for downlink transmission or uplink transmission.

[0047] In some embodiments, each component / of the at least one intent is comprised in an N-dimensional reward function vector r and wherein each component wj of the at least one intent preference value is associated with a respective component / of the reward function vector r. In some embodiments, the determining using the first AI / ML model comprises providing to the first AI / ML model at least one of: a set of one or more information elements indicating or characterizing a state of the first network node, a state of the at least one user device, a state of the communication network environment relevant to communication between the first network node and the at least one user device, or a combination thereof; an indication of the at least one intent; an indication of the at least intent preference value to be optimized for the at least one user device; or a combination thereof.

[0048] In some embodiments, the method further includes obtaining at least one training data sample associated to the first AI / ML model or the second AI / ML model. In some embodiments, the method further includes training the first AI / ML model based on the at least one training data sample; or transmitting a fourth message to a third network node comprising the at least one training data sample. In some embodiments, the method further includes receiving a fifth message, the fifth message comprising the first AI / ML model.

[0049] In another aspect there is provided a network node in a communication network environment for optimizing transmission parameters for at least one user device, the node being configured to perform the methods described above.

[0050] According to another aspect, a node in a communication network environment for optimizing transmission parameters for at least one user device is provided. The node comprises one or more memories comprising instruction data representing a set of instructions; and one or more processors configured to communicate with the one or more memories and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the one or more processors to perform the methods described above.

[0051] According to another aspect, a computer-implemented method, executed by a first user device in a communication network environment for optimizing transmission parameters for the user device, is provided. The method includes generating a first message, the first message comprising a first set of at least one intent for the user device. The method includes transmitting the first message towards a first network node. The method includes obtaining a communication configuration. The method includes updating one or more transmission parameters based on the communication configuration.

[0052] In some embodiments, the method further includes receiving a second message, the second message comprising a request for the first set of at least one intent for the user device.

[0053] In some embodiments, the first set of at least one intent comprises one or more: objective functions, utility functions, performance metrics, key performance indicators, KPIs, quality of service, QoS, requirements, or quality of experience, QoE, requirements.

[0054] In some embodiments, the objective functions, utility functions, performance metrics, and KPIs comprise one or more of: user throughput, user spectral efficiency, transmitted information per packet, resource utilization or allocation, latency, block error rate, bit error rate, reliability, packet error rate, energy consumption, energy savings, discontinuous transmission, discontinuous reception, or signal quality, and wherein the QoS or QoE requirements comprise one or more of: a throughput requirement, a spectral efficiency requirement, a transmission latency requirement, a block error rate requirement, a packet error rate requirement, a bit error rate requirement, a packet error rate requirement, a reliability requirement, a signal to noise ration requirement, a resource utilization requirement, an energy consumption requirement, an energy savings requirement, a resource utilization requirement, a discontinuous transmission requirement, a discontinuous reception requirement, a transport block size requirement, or a resource element requirement.

[0055] In some embodiments, the communication configuration comprises one or more of: an allocation or configuration of at least one communication resource, an allocation or configuration of link adaptation parameters, an allocation or configuration of information bits to be received or transmitted by the at least one user device, an allocation or configuration of downlink or uplink reference signals, a configuration for discontinuous transmission or reception, or a configuration for energy saving.

[0056] According to another aspect, a user device in a communication network environment for optimizing transmission parameters for the user device is provided. The user device comprises one or more memories comprising instruction data representing a set of instructions; and one or more processors configured to communicate with the one or more memories and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the one or more processors to perform the methods described above20.

[0057] In another aspect there is provided a user device in a communication network environment for optimizing transmission parameters for the user device, the user device being configured to perform the methods described above.

[0058] In yet another aspect, a computer program comprising instructions is provided, which, when executed by processing circuitry, causes the processing circuitry to carry out the methods of the methods described above. In another aspect, a carrier containing a computer program is provided, wherein the carrier comprises one of an electronic signals, optical signal, radio signal or computer readable storage medium. In yet another aspect, a computer program product is provided comprising a non-transitory computer readable medium having stored thereon a computer program. In another aspect, an apparatus is provided, the apparatus comprising: a memory; and processing circuitry coupled to the memory, wherein the apparatus is configured to perform the methods described above.BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate various embodiments.

[0060] FIG. 1 illustrates a NG-RAN node architecture.

[0061] FIG. 2 illustrates a NG-RAN node architecture.

[0062] FIG. 3 illustrates a functional framework for RAN Intelligence.

[0063] FIG. 4 illustrates a functional framework for RAN Intelligence.

[0064] FIG. 5 illustrates a method, according to some embodiments.

[0065] FIG. 6 illustrates a signaling diagram, according to some embodiments.

[0066] FIG. 7 illustrates a signaling diagram, according to some embodiments.

[0067] FIG. 8 is a flow diagram, according to some embodiments.

[0068] FIG. 9 is a flow diagram, according to some embodiments.

[0069] FIG. 10 is a flow diagram, according to some embodiments.

[0070] FIG. 11 illustrates a flow diagram, according to some embodiments.

[0071] FIG. 12 illustrates a graph, according to some embodiments.

[0072] FIG. 13 illustrates a graph, according to some embodiments.

[0073] FIG. 14 illustrates a graph, according to some embodiments.

[0074] FIG. 15 illustrates a flow diagram, according to some embodiments.

[0075] FIG. 16 illustrates a method, according to some embodiments.

[0076] FIG. 17 illustrates a method, according to some embodiments.

[0077] FIG. 18 a block diagram of a network node or a user device, according to some embodiments.DETAILED DESCRIPTION

[0078] As used herein, a network node can be a RAN node, an OAM, a Core Network node, an CAM, a Service Management and Orchestration function (SMO), a Network Management System (NMS), a Non-Real Time RAN Intelligent Controller (Non-RT RIO), a Real-Time RAN Intelligent Controller (RT-RIC), a gNB, eNB, en-gNB, ng-eNB, gNB-CU, gNB-CU-CP, gNB-CU-UP, eNB-CU, eNB-CU-CP, eNB-CU-UP, lAB-node, lAB-donor DU, lAB-donor-CU, IAB-DU, IAB-MT, O-CU, O-CU-CP, O-CU-UP, O-DU, O-RU, O-eNB, a UE.

[0079] The terms model training, model optimizing, model optimization, model updating are herein used interchangeably with the same meaning unless explicitly specified otherwise.

[0080] The terms model change, modify or similar are herein used interchangeably with the same meaning unless explicitly specified otherwise. In particular, these phrases refer to the fact that the type, structure, parameters,connectivity, or other aspect of an AI / ML model may have changed compared to a previous format / configuration of the AI / ML model.

[0081] The terms AI / ML model, AI / ML policy, AI / ML algorithm, as well the terms, model, policy or algorithm are herein used interchangeably with the same meaning unless explicitly specified otherwise.

[0082] References to "network nodes” herein should be understood such that a network node may be a physical node or a function or logical entity of any kind, e.g. a software entity implemented in a data center or a cloud, e.g. using one or more virtual machines, and two network nodes may well be implemented as logical software entities in the same data center or cloud.

[0083] Non-limiting examples of AI / ML algorithms may include supervised learning algorithms, deep learning algorithms, reinforcement learning type of algorithms (such as DQN, A2C, A3C, etc.), contextual multi-armed bandit algorithms, autoregression algorithms, etc., or combinations thereof. Such algorithms may exploit functional approximation models, hereafter referred to as AI / ML models, such as neural networks (e.g., feedforward neural networks, deep neural networks, recurrent neural networks, convolutional neural networks, etc.). Examples of reinforcement learning algorithms may include deep reinforcement learning (such as deep Q-network (DQN), proximal policy optimization (PPO), double Q-learning), actor-critic algorithms (such as Advantage actor-critic algorithms, e.g. A2C or A3C, actor-critic with experience replay, etc.), policy gradient algorithms, off-policy learning algorithms, etc.

[0084] The terms intent, intent function, reward, and reward function are herein used interchangeably with the same meaning unless explicitly specified otherwise. An intent value or a reward value indicate a quantifiable value assumed by an intent or reward. The terms intent vector or reward vector refer to a set of one or more intents, intent functions, rewards or reward functions, each indicating a quantifiable value. In some embodiments, intents may be defined by multiple quantifiable and often competing KPIs. The 3GPP TS 28.312, e.g. v. 18.0.0, for instance, describes the concept of intent as a tool to specify objectives, requirements, and constraints for a service or network management workflow

[0085] FIG. 5 illustrates a method according to some embodiments. Method 500 may be executed by a first network node to optimize transmission parameters for at least first user device in a radio communication system based on a plurality of intents associated to the user device and / or to at least a second user device. Step s501 of the method includes obtaining a set of at least one intent for optimizing communication to or from at least a user device. Step s503 of the method includes obtaining at least an intent preference value associated with the at least one user device based on the corresponding at least one intent for optimizing communication to or from the user device. Step s505 of the method includes determining a communication configuration for the user device based on the at least one intent preference value associated to the user device. Step s507 of the method includes configuring the at least one user device to transmit information to or receive information from the first network node based on the communication configuration optimizing the at least one intent associated to the user device.

[0086] In some embodiments, an intent associated to a user device may comprise any one or a combination of: one or more of objective function, utility functions, performance metrics, and key performance indicators, or one or more quality of service requirements, quality of experience requirement.

[0087] In some embodiments, a communication configuration for the user device, either in downlink or uplink, may comprise a plurality of transmission parameters such as, for example, one or a combination of: an allocation of at least one communication resources, an allocation of link adaptation parameters, an allocation of information bits, or an allocation or a configuration of downlink or uplink reference signals.

[0088] In some embodiments, the term intent refers to a quantifiable value based on one or more of: objective functions, utility functions, performance metrics, key performance indicators, as well as quality of service requirements, quality of experience requirements, traffic demands, traffic requirements, etc. The communication configuration may represent one or more transmission parameters that the first network node can determine to optimize the user device communication based on the associated communication intents and / or intents preferences.

[0089] In some embodiments, step s505 of determining a communication configuration for the user device based on the at least one intent preference value associated to the user device may comprise determining one or more of: (i) A configuration or allocation of at least one communication resources (e.g., for downlink transmission or uplink transmission), such as at least one physical resource blocks (PRB), at least one physical resource block groups (PRBG), at least one frequency sub-bands, at least one transmission time intervals; (ii) A configuration or allocation of at least one link adaptation parameter (e.g., for downlink transmission or uplink transmission), such as the communication and coding scheme, a modulation order, a transmission rank; (iii) A configuration or allocation of information bits to be received or transmitted by the user device (i.e., for downlink transmission or uplink transmission, respectively); (iv) A configuration or an allocation of downlink or uplink reference signal to be used to communicate with the user device; (v) A configuration for discontinuous transmission or discontinuous reception to optimize, for instance, energy savings of the user device; and / or (vi) A configuration for energy saving, such as a configuration for discontinuous transmission and / or discontinuous reception, a configuration for activating or deactivating a sleeping mode, an instruction to enter an RRCJDLE state.

[0090] FIG. 6 illustrates a signaling diagram, according to some embodiments. In some embodiments, step s501 of obtaining the set of at least one intent for optimizing the communication with at least a user device and / or step s503 of obtaining at least an intent preference value associated with the at least one user device may require the first network node 600 receiving a FIRST MESSAGE 603 from the at least one user device 602, the FIRST MESSSAGE indicating information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device.

[0091] As shown in FIG. 6, where the first network node 600 receives information related to the user device intent directly from the user device 602, the first network node may additionally transmit a SECOND MESSAGE 605 to theuser device, the SECOND MESSAGE requesting information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device.

[0092] At 607, the first network node 600 performs intent-based user communication optimization, as described in further detail below.

[0093] In some embodiments, configuring the at least one user device to transmit information to or receive information from the first network node based on the communication configuration optimizing the at least one intent associated to the user device may include transmitting a THIRD MESSAGE 609 to the user device, the THIRD MESSAGE providing the communication configuration for the user device optimizing one or more intents associated to the user device.

[0094] In some embodiments, the FIRST MESSAGE 603 may be transmitted by the user device as part of an existing or a new Radio Resource Control (RRC) type of message. In one example, the first network node may receive the FIRST MESSAGE or one or more information provided by the FIRST MESSAGE, as part of a UE capability transfer procedure, as defined by the 3GPP NG-RAN or E-UTRAN systems. In one example, the first network node may receive the FIRST MESSAGE or one or more information provided by the FIRST MESSAGE as part of a UECapabilitylnfomation message. As such information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device may be comprised within a UECapabilitylnfomation message. Additionally, the first network node may transmit SECOND MESSAGE 605 or one or more information provided by the SECOND MESSAGE as part of UECapability Enquiry message as defined by the 3GPP NG-RAN or E-UTRAN systems.

[0095] FIG. 7 illustrates a signaling diagram, according to some embodiments. In another signaling variant as shown in FIG. 7, obtaining the set of at least one intent for optimizing the communication with at least a user device and / or obtaining at least an intent preference value associated with the at least one user device may require the first network node 700 receiving a FIRST MESSAGE 703 from a second network node 704, the FIRST MESSSAGE indicating information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with at least one user device 702.

[0096] In some embodiments, the second network node 704 may represent the source node and the first network node 700 may represent the target node of a handover procedure involving the user device 702. In other words, the user device 702 may be moving between a radio cell controlled by the second network node 704 to a radio cell controlled by the first network node 700. In one possible realization of this embodiment, the anchor point for radio resource control (RRC) signaling of the user device may be moved from the second network node to the first network node. In another possible realization of this embodiment, the anchor point for radio resource control (RRC) signaling of the user device may be retained by the second network node. In this case, the first network node may be the control node of a new secondary cell serving the user device, such as in the case of a dual connectivity procedure.

[0097] With this signaling variant, the first network node 700 may receive from the second network node 704 information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with atleast one user device 702. Such communication may, in some cases, involve serving the user device by means of joint transmission or reception capabilities with the second network node or with a third network node. Therefore, the management of the user device intents may need to be coordinated or shared among multiple network nodes. Coordinating management of the user device intents among multiple network nodes may provide the advantage that the risk of suboptimal configuration of the user device in respect to multiple network nodes can be mitigated. A potential advantage of this signaling variant is to assure that the user intents, such as quality of service requirements, are efficiently and seemingly met during mobility events between two network nodes. An additional potential advantage of this signaling variant is to assure that multiple transmission points serving a user device with joint transmission or joint reception types of communications can efficiently meet user intents.

[0098] In case of a handover event, the first network node 700 may receive the FIRST MESSAGE 703 from the second network node as part of a HANDOVER REQUEST message. Therefore, the second network node may provide information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with at least one user device as part of an HANDVER REQUEST message. In an alternative embodiment, the FIRST MESSAGE 703 may be received by the first network node as part of the user device context setup procedure. In this case, the FIRST MESSAGE 703 may be provided by the second network node to the first network node as part of, for instance, the INITIAL UE CONTEXT SETUP REQUEST message of a 3GPP NG-RAN or E-UTRAN system. The first network node, in this case would represent a RAN node while the second network node may represent a node hosting an Access and Mobility Management Function (AMF).

[0099] In the signaling variant illustrated in FIG. 7, wherein the first network node receives information related to the user device intent from the second network node, the first network node 700 may transmit a SECOND MESSAGE 705 to the second network node 704, the SECOND MESSAGE requesting information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device. In some embodiments, the first network node may receive a FOURTH MESSAGE (not shown in FIG. 7) from the second network node, the SECOND MESSAGE indicating a positive acknowledgement (ACK) or a negative acknowledgment (NACK) related if the intent related information requested by the first network node.

[0100] At 707, the first network node 700 performs intent-based user communication optimization, as described in further detail below.

[0101] In some embodiments, configuring the at least one user device to transmit information to or receive information from the first network node based on the communication configuration optimizing the at least one intent associated to the user device may comprise the first network node 700 transmitting a THIRD MESSAGE 709 to second network node 704 or to the user device 702, the THIRD MESSAGE providing the communication configuration for the user device optimizing one or more intents associated to the user device. In one example, the second network node may additionally forward the THIRD MESSAGE or information received within the THRID MESSAGE to the user device.

[0102] In some embodiments, the first network node may receive the FIRST MESSAGE 603, 703, from the second network node and / or the at least one user device whose transmission or reception parameters need to be optimized to meet one or more intents. In one embodiment, the information provided by the FIRST MESSAGE relates to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device. The FIRST MESSAGE may comprise one or more of: at least one intent for optimizing the communication with the user device and / or at least an intent preference value associated with the at least one intent indicated for optimizing the communication with the user device.

[0103] In some embodiments, an intent indicated by the FIRST MESSAGE associated to a user device may comprise a combination of one or more of objective function, utility functions, performance metrics, and key performance indicators and / or one or more quality of service requirements, quality of experience requirement.

[0104] Non-limiting examples of objective functions, utility functions, performance metrics, and key performance indicators indicated as intents in the FIRST MESSAGE may comprise one or more of:

[0105] User throughput. Non-limiting examples may include: (a) An indication to maximize user throughput; (b) An indication to maximize peak user throughput; (c) An indication to maximize mean user throughput; (d) An indication to maximize minimum user throughput; (e) An indication to of a range of user throughput to be provided; and / or (f) An indication to of a minimum user throughput to be provided.

[0106] User spectral efficiency. Non-limiting examples may include: (a) An indication to maximize spectral efficiency; (b) An indication to maximize peak spectral efficiency; (c) An indication to maximize mean spectral efficiency; (d) An indication to maximize minimum spectral efficiency; (e) An indication to of a range of user spectral efficiency be provided; (f) An indication to of a minimum user spectral efficiency to be provided.

[0107] Transmitted information per packet. Non-limiting examples may include an indication of the transport block size defined by the 3GPP LTE or NG-RAN systems, MAC-PDU size, PDCP-PDU size, etc.

[0108] Resource utilization or allocation. In non-limiting examples, resource utilization or allocation may be expressed, e.g., as a function of one or more of: (a) Number of resource elements (RE) per packet transmission, such as a resource element of an OFDM-based time-frequency resource grid of a 3GPP E-UTRAN or NG-RAN systems, consisting of a subcarrier in frequency domain and an OFDM symbol in time domain; (b) Aggregate number of resource elements (Res) allocated or used per packet transmission, where the aggregation includes number of resource elements used in multiple transmissions of the same packet; (c) Number of physical resource blocks (PRBs) or physical resource block groups (PRBG) allocated or used per packet transmission as defined, for instance, by an OFDM-based time-frequency resource grid of a 3GPP E-UTRAN or NG-RAN systems; and / or (d) Aggregate number of resource elements per packet transmission, where the aggregation includes number of resource elements used in multiple transmissions of the same packet.

[0109] Latency. In non-limiting examples, latency may be expressed, e.g., as a function of one or more of: (a) latency per packet transmission; (b) latency per packet successful delivery; (c) cross-layer end-to-end latency; and / or (d) data service latency, etc.[001 10] Block error rate (BLER). In non-limiting examples, BLER may be expressed, e.g., as a function of one or more of: (a) BLER per packet transmission, and / or (b) statistical BLER, such as average, variance, or standard deviation over a plurality of packets[001 1 1] Bit error rate (BER). In non-limiting examples, BER may be expressed, e.g., as a function of one or more of: (a) BER per packet transmission; and / or (b) Statistical BER, such as average, variance, or standard deviation over a plurality of packets.[001 12] Reliability. In non-limiting examples, reliability may be expressed, e.g., as a function of one or more of: (a) reliability per packet transmission; (b) Statistical reliability, such as average, variance, or standard deviation over a plurality of packets; (c) cross-layer end-to-end reliability; (d) data service reliability, etc.[001 13] Additional examples may include packet error rate, energy consumption, energy savings, discontinuous transmission, discontinuous reception, and / or signal quality, such as signal to noise and interference ratio or signal to noise ratio.[001 14] Non-limiting examples of quality-of-service requirements, quality of experience requirement indicated as intents in the FIRST MESSAGE may comprise one or more of: one or more throughput requirement, such as a minimum throughput requirement; one or more spectral efficiency requirement, such as minimum spectral efficiency requirement; one or more transmission latency requirement, such as a maximum transmission latency requirement; one or more block error rate requirements, such as a maximum block error rate requirement; one or more packet error rate requirements, such as a maximum packet error requirement; one or more bit error rate requirements, such as a maximum bit error rate requirement; one or more packet error rate requirements, such as a maximum packet error requirement; one or more reliability requirement, such as a minimum reliability requirement; one or more signal to noise and interference ratio requirements, such as a minimum signal to noise and interference ratio; one or more signal to noise ration requirements, such as a minimum signal to noise ratio requirement; one or more resource utilization requirements, such as a minimum resource utilization requirement; one or more energy consumption requirements, such as a maximum energy consumption requirement; one or more energy savings requirements, such as a minimum energy savings requirement; one or more resource utilization requirements, such as a minimum resource utilization requirement; one or more discontinuous transmission or discontinuous reception requirements, such as a minimum or a maximum discontinuous transmission requirement or discontinuous reception requirement; one or more transport block size requirements, such as a minimum or maximum transport block size requirement; and / or one or more number of resource elements requirements, such as a minimum or a maximum number of resource elements requirements.[001 15] In some embodiments, the THIRD MESSAGE 609, 709 provides a communication configuration for the user device optimizing one or more intents associated to the user device. The communication configuration maycomprise one or more of: A configuration or allocation of at least one communication resources (e.g., for downlink transmission or uplink transmission), such as at least one physical resource blocks (PRB), at least one physical resource block groups (PRBG), at least one frequency sub-bands, at least one transmission time intervals; a configuration or allocation of at least one link adaptation parameter (e.g., for downlink transmission or uplink transmission), such as the communication and coding scheme, a modulation order, a transmission rank; a configuration or allocation of information bits to be received or transmitted by the user device (i.e. , for downlink transmission or uplink transmission, respectively); a configuration or an allocation of downlink or uplink reference signal to be used to communicate with the user device; a configuration for discontinuous transmission or discontinuous reception to optimize, for instance, energy savings of the user device; and / or a configuration for energy saving, such as a configuration for discontinuous transmission and / or discontinuous reception, a configuration for activating or deactivating a sleeping mode, an instruction to enter an RRCJDLE state.[001 16] In some embodiments, obtaining at least an intent preference value associated with the at least one user device may comprise one or more of: (I) Receiving the FIRST MESSAGE from the user device or from a second network node, the FIRST MESSAGE comprising at least one intent for optimizing the communication with the user device; and / or (ii) Determining one or more intent preference values associated to one or more intents based on at least one intent for optimizing the communication with the user device.[001 17] In embodiments where the first network node obtains a set of A? equal to or greater than one intent(s) associated to optimizing the communication to or from the user device, obtaining at least an intent preference value associated with the user devices intents may include determining a set of N equal to or greater than one intent preference values, wherein each intent preference values represents the relative preference of the user device for the associated intent function.[001 18] FIG. 8 is a flow diagram, according to some embodiments. In some embodiments as illustrated in FIG. 8, obtaining at least an intent preference value associated with the user device(s) intents may be realized by means of a function 800, hereafter referred to as intent manager, mapping one or more user intents to one or mode intent preference values. In the special case where the mapping is realized by means of an identify function, such as an identity matrix, the intent preference values corresponds to the user communication intent values. In a first embodiment, the intent manager 800 mapping the user intents to intent preference values is realized by using parametrized user performance obtained from testing results of the first AI / ML model used by the first network node to determine a communication configuration for the at least one user device. In a second embodiment, the intent manager 800 is realized by using a second AI / ML model trained to map user intents to intent preference values.[001 19] In some embodiments, the intent manager 800 may be realized by mapping one or more user intents to one or more user performance parametrized with respect to one or more intent preference values. For instance, user performance associated to different transmission parameters configurations may be evaluated for several values of each of the intent preference. In one example, user performance parameterized by different intent preference valuesmay be obtained by testing an AI / ML model (used interchangeably herein with a "first AI / ML model”) over a range of values that can be assumed for the intent preference values. In some embodiments, further described below, the AI / ML model is used to determine a communication configuration for a user device or a group of user devices. Examples of testing results obtained by said AI / ML model that the first network node may use for mapping user intents to intent preference values are further described below.

[0120] In one example, an intent preference value a>j associated to an intent function ij of the user device may represent the preference of the user device for the intent function i with respect to the N - 1 intent functions associated to the user device.

[0121] For example, assuming the user device is associated with two intent functions equal to throughput and i2equal reliability of communication, different pairs of preference values a>1and a>2associated toand i2, respectively, may result in different choice of communication configurations.

[0122] If a)1< o>2(e g., rui = 0.2 and a>2= 0.8), the user device prefers a reliable communication over high throughput. Therefore, while with good channel conditions the user device may be configured to reliably communicate even with high throughput parameters (e.g., with a large modulation and coding scheme index and / or high rank), with bad channel conditions the user device will be configured to communicated with reliable communication parameters, possibly at the expense of throughput (e.g., with a low modulation and coding scheme index and / or low rank).

[0123] If a)1> o>2(e.g.,= 0.8 and a>2= 0.2), the user device prefers a high data rate communication over reliability. Therefore, regardless of the channel condition, the user device may always be configured to communicate with high throughput parameters, possibly at the expense of a reliable communication (e.g., with a large modulation and coding scheme index and / or high rank).

[0124] In a non-limiting implementation, the set of N intent preference values may be represented as an N- dimensional preference vector M associated to a user device based on the user device intents. In one example, each component of the preference vector M is real-valued, i.e., M e RN, with each component a>j e M expressing the relative preference of the user device for the corresponding intent function it is associated with.

[0125] In some embodiments, obtaining at least an intent preference value associated with the at least one user device may comprise one or more of determining one or more intent preference values associated to one or more intents of the user device based on: (a) one or more intents of the user device and / or (b) one or more intents of a least a second user device. In this case, the preference of a user device for a certain intent is evaluated with respect to the preference of at least a second user device with respect to the same intent or to other intents.

[0126] In a non-limiting example, the network node may obtain or determine: (I) a first intent preference value a)1for a first user device associated to a first intent function(e.g., throughput), and (ii) a second intent preference value a>2for a second user device associated to a second intent function i2(e.g., reliability).

[0127] In one realization of this example, the intent preference values t, and a>2may represent the relative preference or importance that the first network node should consider when allocating radio resources for the first user device and the second user device. In one example, wherein the first network node may concurrently allocate resources for the first user device and the second user device, the first network node may: (i) prioritize the allocation of radio resources to the first user device if> a>2or (ii) prioritize the allocation of radio resources to the second user device

[0128] In some embodiments, the first network node determines a communication configuration for the user device based on a multi-objective reinforcement learning algorithm based on the at least intent preference value associated to the user device and an AI / ML model. In some embodiments, the first network node executes a reinforcement learning algorithm based on a K-dimensional reward function vector r e RK, with K equal to or greater than one, wherein one or more reward components r may correspond to one or more intents lJ tsuch as objective functions, utility functions, performance metrics, and key performance indicators and / or one or more quality of service requirements or quality of experience requirements indicated as intents in the FIRST MESSAGE.

[0129] In some embodiments, each reward component r corresponds to one intent ij indicated by the FIRST MESSAGE, thereby the dimension K of the reward function vector equals the dimension of the set of N intents indicated by the FIRST MESSAGE, i.e., K equals to N. For example, the reward component may be the same as an intent, such as throughput, and there is a one-to-one mapping between the reward component and the intent.

[0130] In some embodiments, at least a component r7is associated to at least two intents ij indicated by the FIRST MESSAGE, thereby the dimension K of the reward function vector is less than the dimension of the set of N intents indicated by the FIRST MESSAGE, i.e., K less than N. For example, the reward component may implicitly represent an intent, such as using transport block size (TBS) instead of throughput.

[0131] In some embodiments, the negative value of one or more reward components is used instead of the corresponding positive value. This enables to minimize, instead of maximizing, the intents associated to reward components used with negative values.

[0132] In some embodiments, the first network node executes a reinforcement learning algorithm comprising an AI / ML model and a multi-dimensional reward function vector associated to a packet transmission to or form the user device consisting of: (i) an indication of the number of information bits carried by a packet, such as the transport block size (TBS) defined by the 3GPP E-UTRAN or NG-RAN systems and (ii) the negative value of an indication of the aggregate resource utilization for delivering packet delivery, potentially over multiple transmission attempts. These negative values may be one or more of: (a) the negative value of the total number of resource elements N^, (b) the negative value of the total number of physical resource blocks NpRB, (c) the negative value of the total number of physical resource blocks groups NpBBG, or (d) the negative value of the total number resource sub-bands N^.

[0133] In one example, the multi-dimensional reward function vector associated to a packet transmission to or form the user device may be expressed as shown in equation (1) below: h transmission attempt transmission attempt

[0134] Where s: represents a set of one or more information elements indicating or characterizing the state of the first network node, and / or the state of the user device and / or the state of the radio environment relevant to the communication between the first network node and the user device. Non-limiting examples of this information may comprise: (a) information characterizing the radio link, in uplink and / or in downlink, between the network node and the user device, (b) information characterizing the radio interference, in uplink or in downlink, that the user device may experience, (c) information characterizing the resource utilization, load, or capacity in at least the serving cell of the user device, (d) information characterizing the resource utilization, load, or capacity in at least an interfering cell of the user device, (e) the type of traffic required by the user device, and / or (f) the positioning of the user device with respect to the serving cell and / or at least an interfering cell.

[0135] a: represents one or more communication configuration parameters, such as transmission or reception parameters, determined by the first network node for communicating with the user device based on one or more intents. The communication configuration parameters may comprise one or more of: (a) an allocation of at least one communication resources for downlink transmission or uplink transmission, (b) an allocation of link adaptation parameters for downlink transmission or uplink transmission, (c) an allocation of information bits for downlink transmission or uplink transmission, (d) an allocation or a configuration of downlink or uplink reference signal to be used to communicate with the user device.

[0136] TBS represents the transport block size, or number of transmission bits, carried by a packet transmission to the user device.

[0137] represents the total number of resource elements used for communicating with the user device until the nthtransmission attempt of the packet associated to the reward function.

[0138] Therefore, the first network node may use a 2-dimensional reward function with values: (I) zero information bits with -N^ resource elements used, if a negative acknowledgement (NACK) is received at the nthpacket transmission attempt; and (ii) TBS information bits with -N^ resource elements used, if a positive acknowledgement (ACK) is received at the nthpacket transmission attempt.

[0139] In some embodiments, the first network node may additionally obtain an N-dimensional preference vector a) associated to a user device based on the user device intents, wherein the preference vector M has as many dimensions as the number of reward components in r, i.e. M e RK, with each component a>j e M expressing the relative preference of the user device for the corresponding reward component r7e r. The first network node maydetermine a preference function fM(r) representing or parametrizing the relative importance of the different reward components (i.e., objectives) with respect to the K-dimensional preference (importance) vector M e RK, associated to at least a user device. In embodiments where each reward component r correspond to one intent ij indicated by the FIRST MESSAGE, each component a>j of the preference vector M also expresses the relative preference of the user device for the corresponding intent ij.

[0140] In embodiments, the preference function / ^(r) represents a linear combination of the weighted relative importance of the reward components as fM

[0141] In one embodiment of the method, the preference vector components must fulfill the condition that their cumulative sum equals one, i.e. j= 1-

[0142] In one embodiment of the method where the first network node uses a two-dimensional reward vector, e.g., r e R2. the first network node may obtain a preference vector associated to a user device expressed as function of a single scalar parameter taking values in the continuous range of values between zero and one, i.e., e [0, 1], In this case, preference vector M = [ru,, a>2]Tcan be expressed as = [to, 1 - >]7’, i.e.,= a>, a)2= 1 — at. Furthermore, the first network node may determine a preference function / ^(r) representing a linear combination of the weighted relative importance of the reward components associated to the user device as

[0143] In some embodiments, where the first network node uses a 2-dimensional reward function vector associated to a packet transmission of a user device comprising a function of the transport block size TBS and the negative cumulative resource element utilization for multiple transmission of the user packet, the first network node may determine a preference function / ^(r) representing a linear combination of the weighted relative importance of the reward components associated to the user device a / ^(r) = o>TBS - (1 - oj)N^ai.

[0144] In some embodiments as illustrated in FIG. 9, the first network node determines a communication configuration for a user device or a group of user devices based on executing an AI / ML model 902. FIG. 9 is a flow diagram according to some embodiments. An intent manager 900 may obtain as input one or more communication intents, and output one or more intents preference values. The AI / ML model 902 may take as input the intents preference values, and potentially a user and / or network state, as input, and output a user device communication configuration. In some embodiments, the network state comprises a state of the first network node.

[0145] The AI / ML model 902 may require one or more input information in the group of: (I) a set of one or more information elements indicating or characterizing the state of the first network node, and / or the state of the user device and / or the state of the radio environment relevant to the communication between the first network node and the user device; (ii) an indication of at least one communication intent to be optimized for the at least one user device or group of user devices; and / or (ill) an indication of at least one communication intent preference value to be optimized for the at least one user device or group of user devices. In one example, each communication intent preference valuerepresents the preference of the user device for an associated communication intent with respect to other communication intents associated to the user device. In one example, each communication intent preference value represents the preference of the user device for an associated communication intent with respect to other communication intents associated to a second user device or a group of user devices

[0146] In some embodiments, the AI / ML model 902 output may directly provide a communication configuration for a user device or a group of user devices. In another embodiment the AI / ML model output may indirectly provide a communication configuration for a user device or a group of user devices. In one example, the AI / ML model output may be used to determine one or more communication configuration parameters for a user device or a group of user devices. Therefore, the AI / ML model output, in some embodiments, may provide an indication of the importance or correctness or the optimality of one communication configuration for the user device with respect to other communication configurations.

[0147] The following describes a non-limiting example wherein the first network node adopts a multiobjective reinforcement learning algorithm using an AI / ML model 902 to optimize link adaptation transmission parameters based on one or more intents or intents preference values associated to a user device. In this example, a 2-dimensional reward design is defined based on: (1) TBS: the number of information bits (e.g., transport block size), carried by a packet transmission to the user device; andthe total number of resource elements used for communicating with the user device until the nthtransmission attempt of the packet associated to the reward function defined in a previous embodiment as follows: h transmission attempt transmission attempt

[0148] Furthermore, in this example, the communication configuration parameters determined by the first network node for the user device based on the at least one intent preference value associated to the user device comprise one or more link adaptation parameters for downlink transmission or uplink transmission in the group of: (i) a modulation and coding scheme (MCS) index as defined, for instance, by the 3GPP technical specifications for E- UTRAN or NG-RAN systems, (ii) a modulation order, such as QPSK, 16QAM, 64QAM, 256QAM, 1024 QAM, etc., (iii) a transmission rank, or number of spatial streams, or number of spatial layers, and / or (iv) an allocation of information bits, such as a transport block size as defined, for instance, by the 3GPP technical specifications for E-UTRAN or NG- RAN systems.

[0149] The communication configuration parameters represent one or more choices that the first network node can determine to optimize the user device communication intent preferences. In this example, the first network node optimized the communication with the user device with an intent-driven link adaptation.

[0150] In a non-limiting implementation of this example, the first network node may obtain or determine a preference value associated each component of the reward vector, namely T =and r2= TBS, based on a single scalar parameter at taking values in the continuous range of values between zero and one, i.e., at e [0, 1], as a)1= o>, O>2= 1 - t , respectively.

[0151] The first network node may therefore determine different link adaptation parameters optimizing different intents, such as transmission reliability, throughput, or spectral efficiency by determining the preference parameter at that provides the relative preference of the user device for the first or the second component of the reward function.

[0152] FIG. 10 is a flow diagram according to some embodiments. FIG. 10 shows a non-limiting realization of an embodiment wherein the first network node executes an AI / ML model 1002 as part of a multi-objective reinforcement learning algorithm to determine at least a communication configuration parameter for a user device based to intent preference values,and a>2respectively, parameterized by a scalar at e [0, 1], As intent manager 1000 may take as input intent 1 and intent 2, and output intent preference valuesand a>2, respectively, parameterized by a scalar at e [0, 1], in order to determine a user device communication configuration.

[0153] In some embodiments, the first network node may determine one or more communication configuration parameter for link adaptation, such as modulation order, coding rate, transport block size, by: (1) determining one or more intent preference value by mapping the required user intents to intent preference values using parametrized user performance obtained from testing results of the first AI / ML model used by the first network node to determine the user device's configuration communication parameters; and (2) determining the configuration communication parameters for the user device by executing the AI / ML model based at least on the user device and network state and the intent preference values associated to the user device.

[0154] FIG. 11 illustrates a flow diagram, according to some embodiments. FIG. 11 shows a non-limiting example wherein the intent manager 1100 mapping the required user intents to intent preference values is realized by using parametrized user performance obtained from testing results of the AI / ML model 1102 used by the first network node to determine the user device's configuration communication parameters. In a second non-limiting example, not shown in FIG. 11 , the intent manager 1100 mapping the required user intents to intent preference values is realized by using a second AI / ML model 1102.

[0155] FIGs. 12 and 13 illustrates a graph, according to some embodiments. FIGs. 12-13 shows a nonlimiting example of user performance parametrized by different intent preference values at for optimizing transmission parameters of downlink link adaptation. In this example, FIG. 12 illustrates user throughput and spectral efficiency, while FIG. 13 shows block error rate and latency (e.g., average number of transmissions required per packet) obtained by testing the AI / ML model over multiple ranges of the intent preference values can take at in the interval [0, 1], The example shown in FIGs. 12-13 shows three operating points: co=O maximizing reliability and minimizing latency andresource utilization; co=O.2 maximizing throughput, thus balancing latency and spectral efficiency; w=1 maximizing spectral efficiency.

[0156] In a first example, when the intent information associated to the user device requires to maximize transmission reliability or to minimize the packet transmission latency, the first network node may: (1) determine to use an intent preference value at close to zero; and (2) determine one or more link adaptation parameters for the user device optimizing reliability or latency based on the user device and network state and the determined preference value.

[0157] By selecting a preference value at « 0 (i.e., approximately zero), the first network node determines link adaptation parameters minimizing the aggregate resource utilization for delivering a packet to the user device for this example.

[0158] In a second example, when the intent information associated to the user device requires to maximize transmission spectral efficiency (e.g., regardless of reliability and latency), the first network node may: (1) determine to use an intent preference value at close to one; and (2) determine one or more link adaptation parameters for the user device optimizing reliability or latency based on the user device and network state and the determined preference value.

[0159] By selecting a preference value at « 1 (i.e., approximately one), the first network node determines link adaptation parameters maximizing the transmission spectral efficiency, although at the expense of higher block error rate and latency (i.e., larger number of retransmissions).

[0160] In a third example, when the intent information associated to the user device requires to maximize the user throughput, the first network node may: (1) determine to use an intent preference value at close to 0.2; and (2) determine one or more link adaptation parameters for the user device optimizing reliability or latency based on the user device and network state and the determined preference value.

[0161] FIGs. 12-13 show that by selecting a preference value at « 0.2 (i.e., approximately 0.2), the first network node determines link adaptation parameters maximizing the user device throughput.

[0162] FIG. 14 is a graph, according to some embodiments. FIG. 14 illustrates the same examples by showing the Pareto frontier of the two reward components, i.e. transport block size and total number of resource elements per packet, that obtained by testing the AI / ML model trained optimizing link adaptation parameters using different intent preference values at in the interval [0, 1],

[0163] FIG. 15 illustrates a flow diagram, according to some embodiments. In some embodiments, the first network node 1500 may additionally: (1) obtain 1501 at least one training data sample associated for training an AI / ML model or algorithm for optimizing at least communication configuration for user devices, and (2) train an AI / ML model or algorithm for optimizing at least communication configuration for user devices based on the at least one training data sample or transmit a FOURTH MESSAGE 1503 to the second network node or to a third network node 1506 the atleast one training data sample associated for training an AI / ML model or algorithm for optimizing at least communication configuration for user devices.

[0164] Therefore, some embodiments, the first network node 1500 may train an AI / ML model or algorithm for optimizing at least communication configuration for user devices. In another possible realization of the method, illustrated in FIG. 15, the second network node or a third network node 1506 may train 1505 an AI / ML model or algorithm for optimizing at least communication configuration for user devices used by the first network node. In this case, the first network node may additionally receiving a FIFTH MESSAGE 1507 from the second network node or from the third network node, the FIFTH MESSAGE comprising at least an AI / ML model or algorithm for optimizing at least a communication configuration for user devices. In some embodiments, the first network node 1500 may receive a SIXTH MESSAGE 1509 comprising a training data request from the second or third network node 1506.

[0165] In some embodiments, a user device in a radio communication network may execute a method for optimizing communication with a first network node based on a plurality of intents associated to the user device, the method comprising one or more steps of: (1) transmitting a FIRST MESSAGE to the first network node, the FIRST MESSSAGE indicating information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device, and / or (2) receiving a THIRD MESSAGE from the first network node or a FOURTH MESSAHGE from a second network node, where the THIRD MESSAGE or the FOURTH MESSAHE PROVIDE a communication configuration for the user device optimizing one or more intents associated to the user device. The method of the user device may additionally comprise the step of receiving a SECOND MESSAGE from the first network node, the SECOND MESSAGE requesting information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device. In one example, the user device may transmit the FIRST MESSAGE in response to receiving the SECOND MESSAGE.

[0166] In some embodiments, a method is executed by a second network node in a radio communication network for optimizing the communication between a user device and the first network node based on a plurality of intents associated to the user device. The method may include the step of transmitting a FIRST MESSAGE to the first network node, the FIRST MESSSAGE indicating information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device. The method of the second network node may additionally comprise the step of receiving a SECOND MESSAGE from the first network node, the SECOND MESSAGE requesting information related to at least one intent for optimizing the communication (i.e., transmission and / or reception) with the least one user device. In this case, the second network node may transmit the FIRST MESSAGE in response to receiving the SECOND MESSAGE. In one embodiment, the second network node may additionally: (i) receive a THIRD MESSAGE from the first network node, the THIRD MESSAGE providing the communication configuration for the user device optimizing one or more intents associated to the user device; and / or (ii) transmit a FOURTH MESSAGE to the user device, the FOURTH MESSAGE providing the communication configuration for the user device, or part thereof obtained from the first network node.

[0167] In some embodiments, the first network node and / or the second network node can be different RAN nodes (e.g. two gNBs, or two eNBs, or two en-gNBs, or two ng-eNBs).

[0168] In some embodiments, the first network node and / or the second network node can be different nodes / functions of a same RAN node (e.g. a gNB-CU-CP and a gNB-DU, or a gNB-CU-CP and a gNB-CU-UP).

[0169] In some embodiments, the first network node can be a RAN node (e.g. a gNBs, or a eNBs, or a en- gNBs, or a ng-eNBs) and the second network node can be component / nodes / functions of a second RAN node (e.g. gNB-CU-CP).

[0170] In some embodiments, the first network node and / or the second network node can pertain to the same Radio Access Technology (e.g. E-UTRAN, NG-RAN, WiFi, etc.) or to different Radio Access Technologies (e.g. one to NR and the other to E-UTRAN or WiFi).

[0171] In some embodiments, the first network node and / or the second network node can pertain to the same RAN system (e.g. E-UTRAN, NG-RAN, WiFi, etc.) or to different RAN systems (e.g. one to NG-RAN and the other to E-UTRAN).

[0172] In some embodiments, the first network node and the second network node may be connected via a direct signaling connection (e.g. two gNB via XnAP), or an indirect signaling connection (e.g. an e-NB and a gNB via S1AP, NGAP and one or more Core Network nodes, e.g. an MME and an AMF).

[0173] In some embodiments, the first network node can be a management system, such as the CAM system or the SMC, while the second network node can consist of a RAN node or function.

[0174] In some embodiments, the first network node can be a RAN node or function while the second network node can be a management system, such as the CAM or the SMC.

[0175] In some embodiments, the first network node can be a core network node or function, such a 5GC function, while the second network node can consist of a RAN node or function.

[0176] In some embodiments, the first network node can be a RAN node or function while the second network node can be a core network node or function, such a 5GC function.

[0177] FIG. 16 illustrates a method, according to some embodiments. In some embodiments, method 1600 is a computer-implemented method, executed by a first network node in a communication network environment, for optimizing transmission parameters for at least one user device.

[0178] Step s1601 of the method includes obtaining a first set of at least one intent for at least one user device.

[0179] Step s1603 of the method includes obtaining at least one intent preference value associated with the at least one user device based on the obtained first set of at least one intent.

[0180] Step s1605 of the method includes determining, using a first AI / ML model, a communication configuration for the at least one user device based on the at least one intent preference value.

[0181] Step s1607 of the method includes causing, based on the communication configuration, a change to a configuration of the at least one user device.

[0182] FIG. 17 illustrates a method, according to some embodiments. Method 1700 is a computer- implemented method executed by a first user device in a communication network environment for optimizing transmission parameters for the user device.

[0183] Step s1701 of the method includes generating a first message, the first message comprising a first set of at least one intent for the user device.

[0184] Step s1703 of the method includes transmitting the first message towards a first network node.

[0185] Step s1707 of the method includes obtaining a communication configuration.

[0186] Step s1707 of the method includes updating one or more transmission parameters based on the communication configuration.

[0187] FIG. 18 is a block diagram of an apparatus 1800, according to some embodiments. Apparatus 1800 may be a network node or a user device. As shown in FIG. 18, the apparatus may comprise: processing circuitry (PC) 1802, which may include one or more processors (P) 1855 (e.g., one or more general purpose microprocessors and / or one or more other processors, such as an application specific integrated circuit (ASIC), field-programmable gate arrays (FPGAs), and the like); communication circuitry 1848, comprising a transmitter (Tx) 1845 and a receiver (Rx) 1847 for enabling the device to transmit data and receive data (e.g., wirelessly transmit / receive data) over network 1810; and a local storage unit (a.k.a., "data storage system”) 1808, which may include one or more non-volatile storage devices and / or one or more volatile storage devices. In embodiments where PC 1802 includes a programmable processor, a computer program product (CPP) 1841 may be provided. CPP 1841 includes a computer readable medium (CRM) 1842 storing a computer program (CP) 1843 comprising computer readable instructions (CRI) 1844. CRM 1842 may be a non-transitory computer readable medium, such as, magnetic media (e.g., a hard disk), optical media, memory devices (e.g., random access memory, flash memory), and the like. In some embodiments, the CR1 1844 of computer program 1843 is configured such that when executed by PC 1802, the CRI causes the apparatus to perform steps described herein (e.g., steps described herein with reference to the flow charts). In other embodiments, the apparatus may be configured to perform steps described herein without the need for code. That is, for example, PC 902 may consist merely of one or more ASICs. Hence, the features of the embodiments described herein may be implemented in hardware and / or software.

[0188] While various embodiments are described herein, it should be understood that they have been presented by way of example only, and not limitation. Thus, the breadth and scope of this disclosure should not be limited by any of the above described embodiments. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the disclosure unless otherwise indicated herein or otherwise clearly contradicted by context.

[0189] Additionally, while the processes described above and illustrated in the drawings are shown as a sequence of steps, this was done solely for the sake of illustration. Accordingly, it is contemplated that some steps may be added, some steps may be omitted, the order of the steps may be re-arranged, and some steps may be performed in parallel.

[0190] ABBREVIATIONS

[0191] AMF 5G Access and Mobility Management Function

[0192] ON Core Network

[0193] gNB-CU gNB Central Unit (CU)

[0194] gNB-DU gNB Distributed Unit (DU)

Claims

CLAIMS1 . A computer-implemented method (1600), executed by a first network node (600, 700, 1500, 1800) in a communication network environment, for optimizing transmission parameters for at least one user device (602, 702), the method comprising: obtaining (s1601) a first set of at least one intent for at least one user device; obtaining (s1603) at least one intent preference value associated with the at least one user device based on the obtained first set of at least one intent; determining (s1605), using a first AI / ML model (902, 1002, 1102), a communication configuration for the at least one user device based on the at least one intent preference value; and causing (s1607), based on the communication configuration, a change to a configuration of the at least one user device.

2. The method of claim 1 , wherein the first set of at least one intent comprises one or more: objective functions, utility functions, performance metrics, key performance indicators, KPIs, quality of service, QoS, requirements, or quality of experience, QoE, requirements.

3. The method of any of claims 1 or 2, wherein the objective functions, utility functions, performance metrics, and KPIs comprise one or more of: user throughput, user spectral efficiency, transmitted information per packet, resource utilization or allocation, latency, block error rate, bit error rate, reliability, packet error rate, energy consumption, energy savings, discontinuous transmission,discontinuous reception, or signal quality, and wherein the QoS or QoE requirements comprise one or more of: a throughput requirement, a spectral efficiency requirement, a transmission latency requirement, a block error rate requirement, a packet error rate requirement, a bit error rate requirement, a packet error rate requirement, a reliability requirement, a signal to noise ration requirement, a resource utilization requirement, an energy consumption requirement, an energy savings requirement, a resource utilization requirement, a discontinuous transmission requirement, a discontinuous reception requirement, a transport block size requirement, or a resource element requirement.

4. The method of any of claims 1-3, wherein the communication configuration comprises one or more of: an allocation or configuration of at least one communication resource, an allocation or configuration of link adaptation parameters, an allocation or configuration of information bits to be received or transmitted by the at least one user device, an allocation or configuration of downlink or uplink reference signals, a configuration for discontinuous transmission or reception, or a configuration for energy saving.

5. The method of any of claims 1-4, wherein the first set of at least one intent is contained in a first message (603, 703) transmitted by the at least one user device or a second network node.

6. The method of claim 5, further comprising: transmitting a second message (605, 705) towards the at least one user device or the second network node, the second message comprising a request for the first set of at least one intent.

7. The method of claims 5 or 6, further comprising: transmitting a third message (609, 709) towards the at least one user device, the third message comprising the communication configuration.

8. The method of any of claims 1-7, wherein the obtaining the at least one intent preference value comprises: mapping the first set of at least one intent to the at least one intent preference value.

9. The method of claim 8, wherein the mapping comprises: mapping the first set of at least one intent to one or more user performance metrics parameterized with respect to one or more intent preference values.

10. The method of claim 9, wherein the mapping is performed by: obtaining the user performance metrics parameterized with respect to one or more intent preference values by testing the first AI / ML model over a range of values assumed for the one or more intent preference values.

11. The method of any of claims 8-10, wherein the mapping is performed by a second AI / ML model.

12. The method of any of claims 1-10, wherein the first set of at least one intent comprises a first intent and a second intent, and wherein the at least one intent preference value comprises a first intent preference value associated with the first intent and a second intent preference value associated with the second intent.

13. The method of claim 12, wherein the first intent and the second intent correspond to a first user device.

14. The method of claim 12, wherein the first intent corresponds to a first user device and the second intent corresponds to a second user device different than the first user device.

15. The method of any of claims 1 -14, wherein the first set of at least one intent comprises at least one intent relating to transmission reliability, throughput, spectral efficiency, block error rate, or latency.

16. The method of any of claims 12-15, wherein the first intent corresponds to a number of information bits carried by a packet transmission to or from the at least one user device, the second intent corresponds to a total number of resource elements used for communicating with the at least one user device until an nthtransmission attempt of the packet, and the communication configuration comprises one or more link adaptation parameters for downlink transmission or uplink transmission.

17. The method of any of claims 1-16, wherein each component / of the at least one intent is comprised in an N-dimensional reward function vector r and wherein each component wj of the at least one intent preference value is associated with a respective component / ; of the reward function vector r.

18. The method of claim 17, wherein the determining using the first AI / ML model comprises providing to the first AI / ML model at least one of: a set of one or more information elements indicating or characterizing a state of the first network node, a state of the at least one user device, a state of the communication network environment relevant to communication between the first network node and the at least one user device, or a combination thereof, an indication of the at least one intent, an indication of the at least intent preference value to be optimized for the at least one user device, or a combination thereof.

19. The method of any of claims 1-18, further comprising: obtaining at least one training data sample associated to the first AI / ML model or the second AI / ML model.

20. The method of claim 19, further comprising training the first AI / ML model based on the at least one training data sample; or transmitting a fourth message (1503) to a third network node comprising the at least one training data sample.

21. The method of claim 20, further comprising: receiving a fifth message (1507), the fifth message comprising the first AI / ML model.

22. A node (600, 700, 1500, 1800) in a communication network environment for optimizing transmission parameters for at least one user device (602, 702), the node comprising: one or more memories (1842) comprising instruction data (1844) representing a set of instructions; and one or more processors (1855) configured to communicate with the one or more memories and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the one or more processors to perform the methods of any one of claims 1 -21 .

23. A computer program (1843) comprising instructions (1844) which, when executed by processing circuitry (1802), causes the processing circuitry to carry out the methods of any one of claims 1-21.

24. A carrier containing a computer program (1843) according to claim 23, wherein the carrier comprises one of an electronic signals, optical signal, radio signal or computer readable storage medium (1842).

25. A computer program product (1841) comprising a non-transitory computer readable medium having stored thereon a computer program according to claim 23.

26. An apparatus (1800), the apparatus comprising: a memory (1842); and processing circuitry (1802) coupled to the memory, wherein the apparatus is configured to perform the methods of any one of claims 1 -21 .

27. A computer-implemented method (1700), executed by a first user device (602, 702, 1800) in a communication network environment for optimizing transmission parameters for the user device, the method comprising: generating (s1701) a first message (603, 703), the first message comprising a first set of at least one intent for the user device; transmitting (s1703) the first message towards a first network node; obtaining (s1705) a communication configuration; and updating (s1707) one or more transmission parameters based on the communication configuration.

28. The method of claim 27, further comprising receiving a second message (605, 705), the second message comprising a request for the first set of at least one intent for the user device.

29. The method of claims 27 or 28, wherein the first set of at least one intent comprises one or more: objective functions, utility functions, performance metrics, key performance indicators, KPIs, quality of service, QoS, requirements, or quality of experience, QoE, requirements.

30. The method of any of claims 27-29, wherein the objective functions, utility functions, performance metrics, and KPIs comprise one or more of: user throughput, user spectral efficiency, transmitted information per packet, resource utilization or allocation, latency, block error rate, bit error rate, reliability, packet error rate, energy consumption, energy savings, discontinuous transmission, discontinuous reception, or signal quality, and wherein the QoS or QoE requirements comprise one or more of: a throughput requirement, a spectral efficiency requirement, a transmission latency requirement, a block error rate requirement, a packet error rate requirement, a bit error rate requirement, a packet error rate requirement, a reliability requirement, a signal to noise ration requirement,a resource utilization requirement, an energy consumption requirement, an energy savings requirement, a resource utilization requirement, a discontinuous transmission requirement, a discontinuous reception requirement, a transport block size requirement, or a resource element requirement.31 . The method of any of claims 27-30, wherein the communication configuration comprises one or more of: an allocation or configuration of at least one communication resource, an allocation or configuration of link adaptation parameters, an allocation or configuration of information bits to be received or transmitted by the at least one user device, an allocation or configuration of downlink or uplink reference signals, a configuration for discontinuous transmission or reception, or a configuration for energy saving.

32. A user device (602, 702, 1800) in a communication network environment for optimizing transmission parameters for the user device, the user device comprising: one or more memories (1842) comprising instruction data (1844) representing a set of instructions; and one or more processors (1855) configured to communicate with the one or more memories and to execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the one or more processors to perform the methods of any one of claims 27-31 .

33. A computer program (1843) comprising instructions (1844) which, when executed by processing circuitry (1802), causes the processing circuitry to carry out the methods of any one of claims 27-31 .

34. A carrier containing a computer program (1843) according to claim 33, wherein the carrier comprises one of an electronic signals, optical signal, radio signal or computer readable storage medium (1842).

35. A computer program product (1841) comprising a non-transitory computer readable medium having stored thereon a computer program according to claim 33.

36. An apparatus (1800), the apparatus comprising: a memory (1842); and processing circuitry (1802) coupled to the memory, wherein the apparatus is configured to perform the methods of any one of claims 27-31 .

Images