Procedures and signaling designs for additional condition notification from the network to the UE

Abstract

Methods and systems for signaling additional conditions to assist in the training and inference of artificial intelligence (AI) / machine learning (ML) models are discussed herein. For example, a user device (UE) transmits a capability report indicating AI / ML features to a network and receives compatible AI / ML models from the network. The UE also receives information regarding network-side additional conditions used to select and / or design AI / ML models for training. In some cases, the UE also receives auxiliary information identifying AI / ML model features used to assist in selecting and / or designing AI / ML models for training. The UE may display the selected and / or designed AI / ML models to the network. Additionally, the UE may perform inference using the selected AI / ML models based on the received additional conditions. In some cases, the additional conditions are received in an encrypted format.

Description

Technology Field

[0001] The present application generally relates to wireless communication systems comprising wireless communication systems that implement artificial intelligence (AI) / machine learning (ML) models. Background Technology

[0002] Wireless mobile communication technology uses various standards and protocols to transmit data between base stations and wireless communication devices. Wireless communication system standards and protocols include, for example, the 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) (e.g., 4G), 3GPP New Radio (NR) (e.g., 5G), and the Institute of Electrical and Electronics Engineers (IEEE) 802.11 standard for Wireless Local Area Networks (WLANs) (generally, Wi-Fi ® It may include (known to industrial groups).

[0003] As considered by 3GPP, different radio communication system standards and protocols may use various RANs for communication between a base station of a radio access network (RAN) (which may also, sometimes, generally be referred to as a RAN node, network node, or simply node) and a radio communication device known as user equipment (UE). 3GPP RANs may include, for example, GSM (global system for mobile communications), GERAN (enhanced data rates for GSM evolution (EDGE) RAN), UTRAN (Universal Terrestrial Radio Access Network), E-UTRAN (Evolved Universal Terrestrial Radio Access Network), and / or NG-RAN (Next-Generation Radio Access Network).

[0004] Each RAN can perform communication between a base station and a UE using one or more radio access technologies (RATs). For example, GERAN implements GSM and / or EDGE RAT, UTRAN implements UMTS (Universal Mobile Telecommunication System) RAT or other 3GPP RAT, E-UTRAN implements LTE RAT (sometimes referred to simply as LTE), and NG-RAN implements NR RAT (sometimes referred to herein as 5G RAT, 5G NR RAT, or simply NR). In certain deployments, E-UTRAN may also implement NR RAT. In certain deployments, NG-RAN may also implement LTE RAT.

[0005] Base stations used by a RAN can correspond to that RAN. An example of an E-UTRAN base station is an E-UTRAN (Evolved Universal Terrestrial Radio Access Network) Node B (also commonly referred to as Evolved Node B, Enhanced Node B, eNodeB, or eNB). An example of an NG-RAN base station is a Next Generation Node B (sometimes also referred to as g Node B or gNB).

[0006] A RAN provides its communication services to external entities through its connection to a core network (CN). For example, E-UTRAN can utilize the Evolved Packet Core (EPC), whereas NG-RAN can utilize the 5G Core Network (5GC). Brief explanation of the drawing

[0007] To facilitate the identification of discussions regarding any specific element or action, the top digits or numbers within a drawing number refer to the drawing number in which the element was first introduced. FIG. 1 illustrates an example of additional condition signaling between a network and a UE for training an AI / ML model on the UE side, according to the embodiments discussed in this specification. FIG. 2 illustrates an example of additional condition signaling between a network and a UE for UE-side AI / ML inference using function-based life cycle management (LCM) according to embodiments discussed in this specification. FIG. 3 illustrates an example of additional condition signaling between a network and a UE for UE-side AI / ML inference using a model identifier (ID)-based LCM according to the embodiments discussed in this specification. FIG. 4 illustrates a method of the UE according to embodiments of the present specification. FIG. 5 illustrates a method of a base station according to embodiments of the present specification. FIG. 6 illustrates a method of the UE according to embodiments of the present specification. FIG. 7 illustrates a method of a base station according to embodiments of the present specification. FIG. 8 illustrates a method of the UE according to embodiments of the present specification. FIG. 9 illustrates a method of a base station according to embodiments of the present specification. FIG. 10 illustrates a method of the UE according to embodiments of the present specification. FIG. 11 illustrates a method of a base station according to embodiments of the present specification. FIG. 12 illustrates an exemplary architecture of a wireless communication system according to embodiments disclosed in this specification. FIG. 13 illustrates a system for performing signaling between a wireless device and a network device according to embodiments disclosed in this specification. Specific details for implementing the invention

[0008] Various embodiments are described with respect to the UE. However, references to the UE are provided only for exemplary purposes. The exemplary embodiments may be utilized with any electronic component capable of establishing a connection to a network and consist of hardware, software, and / or firmware to exchange information and data with the network. Accordingly, the UE as described herein is used to represent any suitable electronic component.

[0009] In various wireless communication systems, for AI / ML-enabled features / functional groups (FGs), "additional condition(s)" may refer to any modalities assumed for training an AI / ML model but not part of the UE's capability for the AI / ML-enabled features / FGs. It should be noted that under such circumstances, additional condition(s) are not necessarily explicitly specified and may be divided into two categories: network-side additional condition(s) and UE-side additional condition(s).

[0010] Various mechanisms may be considered to support the transmission of additional condition(s) from the network to the UE (and any corresponding signaling). In some cases, AI / ML algorithms may be applicable to various scenarios, locations, configurations, and / or deployments because the AI / ML algorithms can be updated (e.g., by model changes) to reflect various scenarios, locations, and / or deployments. In the case of UE-side AI / ML models, it may be assumed that the UE can notify the RAN regarding the applicability conditions of the AI / ML algorithm(s) available to the UE to support RAN control (e.g., enabling, disabling, and / or switching). UE reporting of AI / ML model applicability may also be further considered.

[0011] In some cases, UE assistance information (UAI) may be used to report additional condition(s).

[0012] Additionally, reactive and preemptive approaches for additional condition reporting may be considered (e.g., the UE reacting to the network configuration, or the UE preemptively notifying the network of updates and / or changes to its supported models and / or functionalities).

[0013] In various wireless communication systems, AI / ML functionality identification and / or functionality-based lifecycle management (LCM) of the UE portion of UE-side models and / or both-side models may be used. Functionality refers to an AI / ML-enabled feature / FG enabled by configuration(s), wherein the configuration(s) are supported based on conditions indicated by UE capabilities. Accordingly, the functionality-based LCM operates based on at least one configuration of the AI / ML-enabled feature / FG or specific configurations of the AI / ML-enabled feature / FG.

[0014] Various signaling mechanisms can be considered to support function-based LCM operations (e.g., to enable, disable, fallback, and / or switch AI / ML functions). Similarly, mechanisms can also be considered regarding whether and / or how to handle additional condition(s) (e.g., scenarios, sites, and datasets) to assist UE-side transparent model operations at the function level (without model identification).

[0015] In the case of AI / ML model identification and Model ID-based LCM of UE-side models and / or the UE portion of both-side models, the Model ID-based LCM operates based on the identified models, wherein the models may be associated with specific configurations and / or conditions related to the UE capabilities of AI / ML enabled features / FGs, and additional condition(s) (e.g., scenarios, sites, and datasets) determined and / or identified between the UE and the network.

[0016] In some wireless communication systems, for the use of UE-side AI / ML model inference, various approaches may be taken to ensure consistency between training and inference regarding network-side additional condition(s) (if identified). In one approach, model identification may be used to achieve alignment between the network-side and the UE-side on / with network-side additional conditions. In a second approach, model training may occur on the network and be transmitted to the UE, where the model is trained under additional conditions. In a third approach, information and / or indication(s) of the network-side additional condition(s) may be provided to the UE. In a fourth approach, consistency may be aided by monitoring the performance of UE-side candidate models / functions (e.g., by the UE and / or the network) to select a model / function. This is not a complete list of approaches, and other approaches (including those for achieving these and / or similar results) are not excluded.

[0017] The embodiments discussed herein provide procedures and signaling designs for additional condition indications. For example, examples of signaling procedures for additional condition(s) transmitted from a network to a UE, and / or for transmitting additional condition(s) from a network to a UE, are discussed herein. Additionally, the embodiments of this specification discuss procedures for UE-side model training and procedures for generating inference in the UE. Examples of such procedures may incorporate a function-based LCM and / or a model ID-based LCM utilizing UE autonomous model switching. It should be understood that the LCM allows for the anticipation of introduced interference or user mobility, which may cause the need to switch to a more applicable AI / ML model, for example.

[0018] Additional condition(s)

[0019] In some embodiments, additional condition(s) may include a combination of information and parameters transparent to the UE, but when displayed to the UE, they may help the UE select an AI / ML model for training and / or inference. For example, information transparent to the UE may include network implementation information, such as dataset information and / or network antenna information. Dataset information may include dataset category, dataset type, dataset size, dataset freshness, etc. Antenna information may include antenna panel information, virtual port mapping, port number(s), antenna polarization type, down-tilt, and / or beam pattern (e.g., synchronization signal block (SSB) / channel state information reference signal (CSI-RS)). In other examples, information transparent to the UE may include network configurations that may not be shared with the UE, such as network bandwidth and / or network frequency. In the third example, information transparent to the UE may include network state information, such as base station, OAM (operations administration and management), core network (CN), and / or location management function (LMF) traffic load and / or resource utilization; base station, OAM, CN, and / or LMF computation load; and / or base station, OAM, CN, and / or LMF congestion level(s). It should be noted that base station, OAM, CN, and LMF may be examples of network entities and may be interchangeable with one another by reference to indicating and utilizing additional condition(s). Additionally, various combinations of the examples discussed may be considered as additional condition(s) transmitted from the network to the UE.

[0020] It should be understood that in some embodiments, additional condition(s) may take various formats. In one example, additional condition(s) may take an open format having explicit content known across various UEs and / or network nodes (e.g., across different vendors for these UEs and / or network nodes).

[0021] In the second example, additional condition(s) may take the format of one or more IDs (e.g., dataset ID(s), network configuration ID(s), network status ID(s)). Such ID(s) may be used to keep network information private and / or exclusive (e.g., to hide network and / or base station implementation details). In some cases, the ID may be a network-assigned ID and / or a manufacturer-assigned ID. In use, the network provides / allows a mapping relationship between the ID and the corresponding dataset in the UE, but the UE may ultimately not know, at a technical level, which network information and / or configuration is mapped to the ID and / or dataset.

[0022] Referring also to the second example, the second example relates to definitions of formats for dataset IDs / model IDs that may be used to facilitate data collection and inference in relation to different use cases (e.g., different network-side conditions for corresponding network signaling). In such a context, it may be understood that the same dataset ID / model ID used in the Radio Resource Control (RRC) configuration for the data collection procedure for training an AI / ML model is also used in the corresponding configuration for the inference procedure using the AI / ML model to ensure compatibility / consistency between the data collection procedure and the inference procedure. Thus, when used in such a way, the dataset ID / model ID may be understood as indicating or representing applicable network-side conditions for the corresponding network signaling in an abstract manner. In some contexts applicable to the second example disclosed herein, the use of "dataset ID" versus "model ID" may be understood as interchangeable. This may correspond, for example, to cases where a single dataset identified by the dataset ID is used to train a corresponding model identified by the model ID. In such a situation, the dataset ID and model ID are logical IDs corresponding to the same scenario (e.g., the same network-side conditions), and therefore a straight swap is possible between using the dataset ID instead of the corresponding model ID (or vice versa).

[0023] In the third example, the additional condition(s) may take the form of a closed format with standardized fields. In some cases, encrypted messages of information regarding more sensitive information (e.g., beam width information) may be used via a public key system and a vendor ID. The UE can recover the information using the vendor ID, the private key, and additional information associated with the vendor ID. Additionally, the private key may be shared with the UE offline (e.g., during negotiations between infrastructure vendor A and UE vendor C). The additional information may be associated with different UE classes (e.g., higher-class UEs can obtain more sensitive information).

[0024] The embodiments disclosed herein discuss the use of additional condition(s) transmitted from a network to the UE for, for example, UE-side AI / ML model training and AI / ML model inference.

[0025] UE-side training procedures

[0026] FIG. 1 illustrates an example of additional condition signaling between a network (106) and a UE (104) for training an AI / ML model on the UE side, according to embodiments discussed in this specification.

[0027] In some embodiments, additional condition(s) signaled from the network (106) to the UE (104) may be used for training the UE-side AI / ML model. To begin, the UE (104) may transmit a UE capability report to the network (106) indicating supported AI / ML features / FGs (108). Subsequently, the network (106) may respond with model configuration information identifying AI / ML models compatible with the supported AI / ML features. Based on the UE capability report, the network (106) may compile a list of compatible model(s) for training according to various alternatives (110).

[0028] In one alternative, the network (106) may configure the UE (104) with a list of compatible model(s) for training transmitted to the UE (104) via an RRC message (110). In some cases, the configured model list is a full list of compatible AI / ML model(s), but the UE (104) may train a subset of the compatible AI / ML model(s) from the full list of AI / ML model(s). This is because the subset of compatible AI / ML models may be those model(s) applicable to the current batch, due to memory constraints, and / or due to UE-based or server-based computational resource constraints. Additionally, model IDs and meta-information (e.g., applicability conditions) may be included in the message. In some examples, such conditions may not be displayed in the message to avoid exclusive information sharing.

[0029] In a second alternative, the network (106) may configure the UE (104) with auxiliary information to have finer control over which AI / ML is selected (110). Non-limiting examples of auxiliary information modalities correspond to CSI compression and / or beam management. In some cases, the auxiliary information may be part of an RRC configuration to trigger a data collection procedure. Whether to design distinct models for different auxiliary information and / or a generalized model applicable to multiple sets of different auxiliary information may depend on the UE (104) implementation. In some such examples, model IDs may not be strictly required.

[0030] Subsequently, the UE (104) may optionally report additional UE-side condition(s) to the network (106), including information within the UE (104) and network transparency information (e.g., memory, speed, location) through UE assistance information (UAI) messages or RRC messages (112).

[0031] Subsequently, the network (106) transmits additional network-side condition(s) to the UE (104), for example, via an RRC message (e.g., RRCReconfiguration) (114). Based on the additional condition(s) signaled from the network (106) and the applicability conditions in the meta-information, the UE (104) can determine which model(s) to train. The UE (104) can collect a dataset of the model by receiving signaling from the network (106) used for training the model. Thus, the UE (104) can perform offline training (116) of the AI / ML model based on the collected dataset. Note that in some alternative embodiments, the offline training (116) of the AI / ML model may take place on the server (102) as facilitated by the UE (104).

[0032] After the AI / ML model training is complete, the UE (104) may notify the network (106), for example, via a UAI message or an uplink (UL) RRC message, which model(s) are ready, which AI / ML models are available in the UE (104) memory, and / or which AI / ML models are available on the server (102) (118). In some cases, the UE (104) may store some model(s) on the server (102) but may not download the models due to memory constraints. Therefore, the network (106) may need to know such information (so that it can account for any remote access delays in terms of latency parameters that can be applied in relation to the inference to be generated by the models).

[0033] UE-side AI / ML model inference functionality-based LCM

[0034] FIG. 2 illustrates an example of additional condition signaling between a network (206) and a UE (204) for UE-side AI / ML inference using a functional-based LCM according to embodiments discussed in this specification.

[0035] In some embodiments, additional condition(s) signaled from the network (206) to the UE (204) may be used for UE-side AI / ML inference using a function-based LCM. To begin, the network (206) may configure functions for inference and / or model switching conditions (208) and transmit the configured functions to the UE (204). Note that in some cases, one such function may be associated with / mapped to multiple model(s), and the UE (204) may autonomously switch between models according to the model switching conditions. In some examples, the model switching conditions may include one or more (or a combination thereof) of a performance monitor metric threshold, network-side additional condition(s), and / or UE-side additional condition(s).

[0036] Subsequently, the transmission of network-side additional conditions from the network (206) to the UE (204) may be triggered according to various alternatives. In one alternative, the network (206) may initiate the transmission of auxiliary information and / or additional condition(s). The UE (204) may select model(s) that correspond to the auxiliary information for training to be used to perform inference. In some cases, the time for transmitting the auxiliary information may depend on the implementation of the network (206).

[0037] In a second alternative, the UE (204) may initiate the transmission of auxiliary information and / or additional condition(s) when a configured event is satisfied (210). Thus, the UE (204) may transmit a UL message to the network to request network-side additional condition(s) and / or auxiliary information. The configured event may be, for example, a monitored inference error exceeding a configured threshold. Additionally, the UL message may take the form of a UL RRC message (e.g., a UAI message), a medium access control-control element (MAC-CE) message, and / or a dedicated scheduling request (SR). In some cases, if AI / ML model inference is latency-sensitive, the network may configure a dedicated SR resource.

[0038] Subsequently, in response to a triggering event, the network (206) displays network-side additional condition(s) to the UE (204), for example, via an RRC message (e.g., RRCReconfiguration) or a MAC-CE message (212). Then, based on the active network-side additional condition(s) received from the network (206) and based on the conditions configured for model switching, the UE (204) determines an AI / ML model for inference and / or switches to that model (214) and collects a corresponding dataset for the AI / ML model. Then, the UE (204) performs AI / ML model inference using the AI / ML model. In some cases, where the AI / ML model is stored on the server (202) and not on the UE (204), the UE (204) downloads the AI / ML model from the server for inference. Optionally, in some instances, the UE (204) can notify the network (206), for example, of the AI / ML model used for inference in the UAI message (216).

[0039] UE-side AI / ML model inference Model ID-based LCM

[0040] FIG. 3 illustrates an example of additional condition signaling between a network (306) and a UE (304) for UE-side AI / ML inference using a model ID-based LCM according to embodiments discussed in this specification.

[0041] In some embodiments, additional condition(s) signaled from the network (306) to the UE (304) may be used for UE-side AI / ML inference using a model ID-based LCM. To begin, the network (306) may compose a list of model(s) for inference (308). Subsequently, the transmission of active network-side additional conditions from the network to the UE may be triggered similarly to the transmission triggering for UE-side AI / ML model inference using a function-based LCM discussed herein. As discussed in the embodiments of this specification (i.e., regarding the function-based LCM), in one alternative, the network may request the UE (304) to perform inference (310). In a second alternative, the UE may request network-side additional condition(s) for inference (312).

[0042] Accordingly, in response to a triggering event, the network (306) can transmit additional active network-side condition(s) to the UE (304) via, for example, an RRC message (e.g., an RRCReconfiguration message) or MAC-CE (314).

[0043] Based on additional active network-side conditions received from the network, the UE (304) can down-scope a subset of configured AI / ML models (316). That is, the UE (304) can select a subset of AI / ML models from a set of configured AI / ML models for inference. Subsequently, the UE (304) can notify the network (306) of the selected subset of models, for example, via a UL RRC message or a UAI message (318). In response to the notification of the selected subset of models, the network (306) can activate and / or configure a model ID for inference (320) and transmit the activated and / or configured model ID to the UE (304). It should be understood that if the model ID cannot be enabled or configured, the network (306) may indicate to the UE (304) to switch to a different AI / ML model, fall back to previously used AI / ML models, or fall back to previous / legacy procedures.

[0044] Based on the received activation and / or configured model ID, the UE (304) may perform AI / ML model inference using the AI / ML model corresponding to the activation and / or configured model ID (322). It should be understood that the UE (304) may perform AI / ML model interference using the AI / ML model that was switched when the model ID cannot be activated or configured (322).

[0045] FIG. 4 illustrates a method (400) of a UE according to embodiments of the present specification. The method (400) includes the step (402) of transmitting a capability report (402) to a base station indicating one or more supported AI / ML features that enable the UE to perform AI / ML model training. The method (400) further includes the step (404) of receiving a model configuration message (404) from a base station identifying one or more compatible AI / ML models, wherein each of the one or more compatible AI / ML models is compatible with at least one of one or more supported AI / ML features. The method (400) further includes the step (406) of receiving one or more active network-side additional conditions (406) from a base station. The method (400) further includes the step (408) of selecting an AI / ML model selected for training from one or more compatible AI / ML models based on one or more active network-side additional conditions. The method (400) further includes the step (410) of generating a training dataset for the selected AI / ML model based on signaling received from a base station. The method (400) further includes the step (412) of transmitting a training dataset to a server to cause training of a selected AI / ML model on the server. The method (400) further includes the step (414) of transmitting a notification message to a base station indicating that the selected AI / ML model is available to the UE.

[0046] In some embodiments of the method (400), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0047] In some embodiments of the method (400), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0048] In some embodiments of the method (400), one or more additional active network-side conditions include one or more of network bandwidth and network frequency.

[0049] In some embodiments of the method (400), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0050] In some embodiments of the method (400), one or more active network-side additional conditions are received from the UE in an encrypted format, and the method further includes the step of decrypting one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station.

[0051] In some embodiments of the method (400), the step of selecting an AI / ML model for training from one or more compatible AI / ML models is further based on the memory constraints of the UE.

[0052] In some embodiments of the method (400), the step of selecting an AI / ML model for training from one or more compatible AI / ML models is further based on computational constraints of the UE.

[0053] In some embodiments of the method (400), the step of selecting an AI / ML model for training from one or more compatible AI / ML models is additionally based on the computational constraints of the server.

[0054] In some embodiments of the method (400), the model configuration message further includes additional condition compatibility information for one or more compatible AI / ML models, and the step of selecting an AI / ML model selected for training from one or more compatible AI / ML models based on additional active network-side conditions includes the step of comparing the additional active network-side conditions with the additional condition compatibility information for one or more compatible AI / ML models.

[0055] In some embodiments, the method (400) further includes the step of transmitting one or more active UE-side additional conditions to a base station.

[0056] In some embodiments of the method (400), the notification message further indicates that the selected AI / ML model is stored in the UE.

[0057] In some embodiments of the method (400), the notification message further indicates that the selected AI / ML model is stored on the server.

[0058] FIG. 5 illustrates a method (500) of a base station according to embodiments of the present specification. The method (500) includes the step (502) of receiving from a UE a capability report indicating one or more supported AI / ML features that enable the UE to perform AI / ML model training. The method (500) further includes the step (504) of transmitting a model configuration message to the UE that identifies one or more compatible AI / ML models, wherein each of the one or more compatible AI / ML models is compatible with at least one of the one or more supported AI / ML features. The method (500) further includes the step (506) of transmitting one or more active network-side additional conditions to the UE. The method (500) further includes the step (508) of performing signaling with the UE that enables the UE to generate a training dataset for an AI / ML model selected from one or more compatible AI / ML models. The method (500) further includes the step (510) of receiving a notification message from the UE indicating that a selected AI / ML model is available to the UE.

[0059] In some embodiments of the method (500), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0060] In some embodiments of the method (500), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0061] In some embodiments of the method (500), one or more additional active network-side conditions include one or more of network bandwidth and network frequency.

[0062] In some embodiments of the method (500), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0063] In some embodiments, the method (500) further comprises the step of encrypting one or more active network-side additional conditions in an encrypted format using an encryption key corresponding to a base station, and the one or more active network-side additional conditions are transmitted by the base station in an encrypted format.

[0064] In some embodiments of the method (500), the model configuration message further includes additional condition compatibility information for one or more compatible AI / ML models.

[0065] In some embodiments, the method (500) further includes the step of receiving one or more active UE-side additional conditions from the UE.

[0066] In some embodiments of the method (500), the notification message further indicates that the selected AI / ML model is stored in the UE.

[0067] In some embodiments of the method (500), the notification message further indicates that the selected AI / ML model is stored on the server.

[0068] FIG. 6 illustrates a method (600) of a UE according to embodiments of the present specification. The method (600) includes the step (602) of transmitting to a base station a capability report indicating one or more supported AI / ML features that enable the UE to perform AI / ML model training. The method (600) further includes the step (604) of receiving from the base station auxiliary information identifying one or more active AI / ML features, wherein each of the one or more active AI / ML features includes one of one or more supported AI / ML features. The method (600) further includes the step (606) of receiving one or more active network-side additional conditions from the base station. The method (600) further includes the step (608) of designing an AI / ML model for training based on one or more active AI / ML features and one or more active network-side additional conditions. The method (600) further includes the step (610) of generating a training dataset for an AI / ML model based on signaling received from the base station. The method (600) further includes the step (612) of transmitting a training dataset to a server to cause training of an AI / ML model on the server. The method (600) further includes the step (614) of transmitting a notification message to a base station indicating that the AI / ML model is available to the UE.

[0069] In some embodiments of the method (600), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0070] In some embodiments of the method (600), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0071] In some embodiments of the method (600), one or more additional active network-side conditions include one or more of network bandwidth and network frequency.

[0072] In some embodiments of the method (600), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0073] In some embodiments of the method (600), one or more active network-side additional conditions are received from the UE in an encrypted format, and the method further includes the step of decrypting one or more active network-side additional conditions using a decryption key corresponding to the base station.

[0074] In some embodiments of the method (600), one or more active AI / ML features include one or more of CSI compression features and beam management features.

[0075] In some embodiments, the method (600) further includes the step of transmitting one or more active UE-side additional conditions to a base station.

[0076] In some embodiments of the method (600), the notification message further indicates that the selected AI / ML model is stored in the UE.

[0077] In some embodiments of the method (600), the notification message further indicates that the selected AI / ML model is stored on the server.

[0078] FIG. 7 illustrates a method (700) of a base station according to embodiments of the present specification. The method (700) includes the step (702) of receiving from a UE a capability report indicating one or more supported AI / ML features that enable the UE to perform AI / ML model training. The method (700) further includes the step (704) of transmitting auxiliary information identifying one or more active AI / ML features to the UE, wherein each of the one or more active AI / ML features includes one of the one or more supported AI / ML features. The method (700) further includes the step (706) of transmitting one or more active network-side additional conditions to the UE. The method (700) further includes the step (708) of performing signaling with the UE, which enables the UE to generate a training dataset for an AI / ML model designed in the UE based on one or more active AI / ML features and one or more active network-side additional conditions. The method (700) further includes the step (710) of receiving a notification message from the UE indicating that an AI / ML model is available to the UE.

[0079] In some embodiments of the method (700), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0080] In some embodiments of the method (700), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0081] In some embodiments of the method (700), one or more active network-side additional conditions include one or more of network bandwidth and network frequency.

[0082] In some embodiments of the method (700), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0083] In some embodiments, the method (700) further comprises the step of encrypting one or more active network-side additional conditions in an encrypted format using an encryption key corresponding to a base station, and the one or more active network-side additional conditions are transmitted by the base station in an encrypted format.

[0084] In some embodiments of the method (700), one or more active AI / ML features include one or more of CSI compression features and beam management features.

[0085] In some embodiments, the method (700) further includes the step of receiving one or more active UE-side additional conditions from the UE.

[0086] In some embodiments of the method (700), the notification message further indicates that the selected AI / ML model is stored in the UE.

[0087] In some embodiments of the method (700), the notification message further indicates that the selected AI / ML model is stored on the server.

[0088] FIG. 8 illustrates a method (800) of a UE according to embodiments of the present specification. The method (800) includes the step (802) of receiving a model configuration message from a base station indicating one or more supported AI / ML functionalities for inference and one or more model switch conditions. The method (800) further includes the step (804) of receiving one or more active network-side additional conditions from a base station. The method (800) further includes the step (806) of selecting an AI / ML model from one or more compatible AI / ML models based on one or more active network-side additional conditions and one or more model switch conditions. The method (800) further includes the step (808) of receiving a signaling from a base station containing a dataset for use with the selected AI / ML model. The method (800) further includes the step (810) of performing inference using the selected AI / ML model and the received dataset corresponding to the selected AI / ML model.

[0089] In some embodiments of the method (800), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0090] In some embodiments of the method (800), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0091] In some embodiments of the method (800), one or more active network-side additional conditions include one or more of network bandwidth and network frequency.

[0092] In some embodiments of the method (800), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0093] In some embodiments of the method (800), one or more active network-side additional conditions are received from the UE in an encrypted format, and the method further includes the step of decrypting one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station.

[0094] In some embodiments, the method (800) further comprises the steps of determining that a configured event is satisfied, and transmitting a UL request message to a base station indicating a request for one or more active network-side additional conditions based on the determination that the configured event is satisfied. In some such embodiments, the UL request message includes one or more of a UL RRC message, a MAC-CE message, and a dedicated SR message. In some other such embodiments, the configured event is that a monitored inference error is greater than or equal to a configured threshold.

[0095] In some embodiments, the method (800) further includes the step of transmitting a notification message to a base station identifying a selected AI / ML model used to perform inference.

[0096] In some embodiments of the method (800), one or more model switch conditions include one or more of a performance monitor metric threshold, a network-side additional condition, and a UE-side additional condition.

[0097] In some embodiments of the method (800), one or more active network-side additional conditions are received from the base station in one or more of the RRC message and MAC-CE message.

[0098] FIG. 9 illustrates a method (900) of a base station according to embodiments of the present specification. The method (900) includes the step (902) of transmitting a model configuration message to a UE indicating one or more supported AI / ML functions for inference and one or more model switch conditions. The method (900) further includes the step (904) of transmitting one or more active network-side additional conditions to the UE. The method (900) further includes the step (906) of transmitting a signaling to the UE containing one or more datasets corresponding to one or more supported AI / ML functions.

[0099] In some embodiments of the method (900), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0100] In some embodiments of the method (900), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0101] In some embodiments of the method (900), one or more additional active network-side conditions include one or more of network bandwidth and network frequency.

[0102] In some embodiments of the method (900), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0103] In some embodiments, the method (900) further includes the step of receiving a notification message from the UE, comprising a selected AI / ML model used to perform inference in the UE.

[0104] In some embodiments, the method (900) further comprises the step of encrypting one or more active network-side additional conditions in an encrypted format using an encryption key corresponding to a base station, and the one or more active network-side additional conditions are transmitted by the base station in an encrypted format.

[0105] In some embodiments, the method (900) further comprises the step of receiving a UL request message from a UE requesting one or more active network-side additional conditions, and one or more active network-side additional conditions are transmitted to the UE based on the UL request message. In some such embodiments, the UL request message received from the UE includes one or more of an RRC message, a MAC-CE message, and a dedicated SR message.

[0106] In some embodiments of the method (900), one or more model switch conditions include one or more of a performance monitor metric threshold, a network-side additional condition, and a UE-side additional condition.

[0107] In some embodiments of the method (900), one or more active network-side additional conditions include one or more of an RRC message and a MAC-CE message.

[0108] FIG. 10 illustrates a method (1000) of a UE according to embodiments of the present specification. The method (1000) includes the step (1002) of receiving a model configuration message from a base station indicating one or more compatible AI / ML models for inference. The method (1000) further includes the step (1004) of receiving one or more active network-side additional conditions from a base station. The method (1000) further includes the step (1006) of identifying a subset of AI / ML models from one or more compatible AI / ML models for inference based on one or more active network-side additional conditions. The method (1000) further includes the step (1008) of transmitting a notification message containing the subset of AI / ML models to a base station. The method (1000) further includes the step (1010) of receiving an AI / ML model ID for a selected AI / ML model to be used for inference from a base station. The method (1000) further includes a step (1012) of performing inference using a selected AI / ML model.

[0109] In some embodiments of the method (1000), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0110] In some embodiments of the method (1000), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0111] In some embodiments of the method (1000), one or more active network-side additional conditions include one or more of network bandwidth and network frequency.

[0112] In some embodiments of the method (1000), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0113] In some embodiments of the method (1000), one or more active network-side additional conditions are received from the UE in an encrypted format, and the method further includes the step of decrypting one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station.

[0114] In some embodiments, the method (1000) further includes the step of sending a notification message containing a subset of AI / ML models to a server.

[0115] In some embodiments, the method (1000) further comprises the step of transmitting a request for one or more active network-side additional conditions to a base station, and the one or more active network-side additional conditions are received in response to the request for one or more active network-side additional conditions.

[0116] In some embodiments of the method (1000), one or more active network-side additional conditions include one or more of an RRC message and a MAC-CE message.

[0117] In some embodiments of the method (1000), the notification message includes a UL RRC message.

[0118] FIG. 11 illustrates a method (1100) of a base station according to embodiments of the present specification. The method (1100) includes the step (1102) of transmitting a model configuration message to a UE indicating one or more compatible AI / ML for inference. The method (1100) further includes the step (1104) of transmitting one or more active network-side additional conditions to the UE. The method (1100) further includes the step (1106) of receiving a notification message from the UE identifying a subset of AI / ML models, wherein the subset of AI / ML models is a subset of one or more compatible AI / ML models. The method (1100) further includes the step (1108) of identifying a selected AI / ML module for use in inference based on the notification message. The method (1100) further includes the step (1110) of transmitting a display message to the UE containing an AI / ML model ID for the selected AI / ML model for use in inference.

[0119] In some embodiments of the method (1100), one or more active network-side additional conditions include dataset information. In some such embodiments, the dataset information includes one or more of dataset category, dataset type, dataset size, and dataset freshness.

[0120] In some embodiments of the method (1100), one or more active network-side additional conditions include network antenna information. In some such embodiments, the network antenna information includes one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern.

[0121] In some embodiments of the method (1100), one or more additional active network-side conditions include one or more of network bandwidth and network frequency.

[0122] In some embodiments of the method (1100), one or more active network-side additional conditions include network state information. In some such embodiments, the network state information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level.

[0123] In some embodiments, the method (1100) further comprises the step of encrypting one or more active network-side additional conditions in an encrypted format using an encryption key corresponding to a base station, and the one or more active network-side additional conditions are transmitted by the base station in an encrypted format.

[0124] In some embodiments, the method (1100) further comprises the step of receiving a UL request message from a UE requesting one or more active network-side additional conditions, and one or more active network-side additional conditions are transmitted to the UE based on the UL request message.

[0125] In some embodiments of the method (1100), one or more active network-side additional conditions include one or more of an RRC message and a MAC-CE message.

[0126] In some embodiments of the method (1100), the notification message includes a UL RRC message.

[0127] FIG. 12 illustrates an exemplary architecture of a wireless communication system (1200) according to embodiments disclosed herein. The following description is provided for an exemplary wireless communication system (1200) operating with LTE system standards and / or 5G or NR system standards as provided by 3GPP technical specifications.

[0128] As illustrated by FIG. 12, the wireless communication system (1200) includes a UE (1202) and a UE (1204) (although any number of UEs may be used). In this example, the UE (1202) and the UE (1204) are exemplified as smartphones (e.g., handheld touchscreen mobile computing devices capable of connecting to one or more cellular networks), but may also include any mobile or non-mobile computing device configured for wireless communication.

[0129] UE (1202) and UE (1204) may be configured to be coupled to communicate with RAN (1206). In embodiments, RAN (1206) may be NG-RAN, E-UTRAN, etc. UE (1202) and UE (1204) utilize connections (or channels) with RAN (1206) (respectively illustrated as connection (1208) and connection (1210)), each of which includes a physical communication interface. RAN (1206) may include one or more base stations (e.g., base station (1212) and base station (1214)) that enable connection (1208) and connection (1210).

[0130] In this example, the connection (1208) and the connection (1210) are air interfaces to enable such communication coupling and may match RAT(s) used by a RAN (1206), such as LTE and / or NR.

[0131] In some embodiments, the UE (1202) and the UE (1204) may also directly exchange communication data through the sidelink interface (1216). The UE (1204) is illustrated as being configured to access an access point (illustrated as an AP (1218)) through a connection (1220). For example, the connection (1220) may include a local wireless connection, such as a connection conforming to any IEEE 802.11 protocol, where the AP (1218) is a Wi-Fi ®A router may be included. In this example, the AP (1218) may be connected to another network (e.g., the Internet) without going through the CN (1224).

[0132] In some embodiments, UE (1202) and UE (1204) may be configured to communicate with each other or with base station (1212) and / or base station (1214) using OFDM communication signals over a multicarrier communication channel according to various communication techniques, such as but not limited to OFDMA (orthogonal frequency division multiplexing) communication techniques (e.g., for downlink communications) or SC-FDMA (single carrier frequency division multiple access) communication techniques (e.g., for uplink and ProSe or sidelink communications), but not limited thereto, but the scope of embodiments is not limited in this respect. OFDM signals may include multiple orthogonal subcarriers.

[0133] In some embodiments, all or part of the base station (1212) or base station (1214) may be implemented as one or more software entities running on server computers as part of a virtual network. Additionally, or in other embodiments, the base station (1212) or base station (1214) may be configured to communicate with each other through an interface (1222). In embodiments where the wireless communication system (1200) is an LTE system (e.g., when CN (1224) is an EPC), the interface (1222) may be an X2 interface. The X2 interface may be defined between two or more base stations (e.g., two or more eNBs, etc.) connected to the EPC, and / or between two eNBs connected to the EPC. In embodiments where the wireless communication system (1200) is an NR system (e.g., when CN (1224) is a 5GC), the interface (1222) may be an Xn interface. The Xn interface is defined between two or more base stations (e.g., two or more gNBs, etc.) connected to 5GC, between a base station (1212) (e.g., gNB) connected to 5GC and an eNB, and / or between two eNBs connected to 5GC (e.g., CN (1224)).

[0134] The RAN (1206) is illustrated as being communicably coupled to the CN (1224). The CN (1224) may include one or more network elements (1226), which are configured to provide various data and telecommunication services to customers / subscribers (e.g., users of the UE (1202) and UE (1204)) connected to the CN (1224) through the RAN (1206). The components of the CN (1224) may be implemented in one physical device or separate physical devices, including components for reading and executing instructions from a machine-readable or computer-readable medium (e.g., a non-transient machine-readable storage medium).

[0135] In embodiments, CN (1224) may be an EPC, and RAN (1206) may be connected to CN (1224) through S1 interface (1228). In embodiments, S1 interface (1228) may be divided into two parts: an S1 user plane (S1-U) interface that carries traffic data between base station (1212) or base station (1214) and a serving gateway (S-GW), and an S1-MME interface that is a signaling interface between base station (1212) or base station (1214) and mobility management entities (MMEs).

[0136] In embodiments, CN (1224) may be 5GC, and RAN (1206) may be connected to CN (1224) through NG interface (1228). In embodiments, NG interface (1228) may be divided into two parts: NG user plane (NG-U) interface, which carries traffic data between base station (1212) or base station (1214) and user plane function (UPF), and S1 control plane (NG-C) interface, which is a signaling interface between base station (1212) or base station (1214) and access and mobility management function (AMF).

[0137] Generally, the application server (1230) may be an element that provides applications using internet protocol (IP) bearer resources together with the CN (1224) (e.g., packet-switched data services). The application server (1230) may also be configured to support one or more communication services (e.g., VoIP sessions, group communication sessions, etc.) for the UE (1202) and UE (1204) through the CN (1224). The application server (1230) may communicate with the CN (1224) through an IP communication interface (1232).

[0138] FIG. 13 illustrates a system (1300) for performing signaling (1334) between a wireless device (1302) and a network device (1318) according to embodiments disclosed herein. The system (1300) may be part of a wireless communication system as described herein. The wireless device (1302) may be, for example, a UE of a wireless communication system. The network device (1318) may be, for example, a base station (e.g., eNB or gNB) of a wireless communication system.

[0139] A wireless device (1302) may include one or more processor(s) (1304). The processor(s) (1304) may execute instructions that cause various operations of the wireless device (1302) to be performed as described herein. The processor(s) (1304) may include one or more baseband processors implemented using, for example, a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), a controller, a field programmable gate array (FPGA) device, other hardware devices, firmware devices, or any combination thereof configured to perform the operations described herein.

[0140] The wireless device (1302) may include a memory (1306). The memory (1306) may be a non-transient computer-readable storage medium that stores instructions (1308) (these may include instructions executed by, for example, the processor(s) (1304). The instructions (1308) may also be referred to as program code or computer programs. The memory (1306) may also store data used by the processor(s) (1304) and results calculated by them.

[0141] The wireless device (1302) may include one or more transceiver(s) (1310) that may include a radio frequency (RF) transmitter circuit and / or receiver circuit that use the antenna(s) (1312) of the wireless device (1302) to facilitate signaling (e.g., signaling (1334)) to and / or from the wireless device (1302) to other devices (e.g., network device (1318)) according to corresponding RATs.

[0142] The wireless device (1302) may include one or more antenna(s) (1312) (e.g., one, two, four or more). In the case of embodiments having multiple antenna(s) (1312), the wireless device (1302) may leverage the spatial diversity of such multiple antenna(s) (1312) to transmit and / or receive multiple different data streams over the same time and frequency resources. This behavior may be referred to, for example, as multiple input multiple output (MIMO) behavior (referring to multiple antennas used in each of the transmitting device and the receiving device that enable this aspect). MIMO transmissions by a wireless device (1302) can be achieved by precoding (or digital beamforming) applied in the wireless device (1302) which multiplexes data streams across antenna(s) (1312) according to known or assumed channel characteristics, so that each data stream is received at a desired location in the spatial domain (e.g., the location of the receiver associated with the data stream) with an appropriate signal strength relative to other streams. Certain embodiments may use single user MIMO (SU-MIMO) methods (where all data streams are directed to a single receiver) and / or multi-user MIMO (MU-MIMO) methods (where individual data streams may be directed to individual (different) receivers at different locations within the spatial domain).

[0143] In some embodiments having multiple antennas, the wireless device (1302) can implement analog beamforming techniques, thereby adjusting the phases of signals transmitted by the antenna(s) (1312) relative to each other so that the (joint) transmission of the antenna(s) (1312) can be directed (this is sometimes referred to as beam steering).

[0144] A wireless device (1302) may include one or more interface(s) (1314). The interface(s) (1314) may be used to provide input to the wireless device (1302) or to provide output from it. For example, a wireless device (1302) which is a UE may include interface(s) (1314), such as microphones, speakers, touchscreens, buttons, etc., to allow input and / or output to the UE by a user of the UE. Other interfaces of such a UE may consist of transmitters, receivers, and other circuitry (other than the transceiver(s) (1310) / antenna(s) (1312) already described) that allow communication between the UE and other devices, and may be composed of known protocols (e.g., Wi-Fi ® , Bluetooth ® It can operate depending on the etc.

[0145] The wireless device (1302) may include an additional condition signaling module (1316). The additional condition signaling module (1316) may be implemented through hardware, software, or a combination thereof. For example, the additional condition signaling module (1316) may be implemented as a processor, circuit, and / or instructions (1308) stored in memory (1306) and executed by the processor(s) (1304). In some examples, the additional condition signaling module (1316) may be integrated within the processor(s) (1304) and / or transceiver(s) (1310). For example, the additional condition signaling module (1316) may be implemented by a combination of software components (e.g., executed by a DSP or general processor) and hardware components (e.g., logic gates and circuits) within the processor(s) (1304) or transceiver(s) (1310).

[0146] The additional condition signaling module (1316) may be used in various embodiments of the present disclosure, e.g., any one of FIG. 1, FIG. 2, and / or FIG. 3. The additional condition signaling module (1316) is configured to receive one or more additional conditions used to select an AI / ML model for training from a set of compatible AI / ML models. In some cases, an AI / ML module selected based on the received additional conditions may be used for inference.

[0147] A network device (1318) may include one or more processor(s) (1320). The processor(s) (1320) may execute instructions to perform various operations of the network device (1318) as described herein. The processor(s) (1320) may include one or more baseband processors implemented using, for example, a CPU, DSP, ASIC, controller, FPGA device, other hardware device, firmware device, or any combination thereof configured to perform the operations described herein.

[0148] The network device (1318) may include memory (1322). The memory (1322) may be a non-transient computer-readable storage medium that stores instructions (1324) (which may include instructions executed by, for example, the processor(s) (1320). The instructions (1324) may also be referred to as program code or computer programs. The memory (1322) may also store data used by the processor(s) (1320) and results calculated by them.

[0149] The network device (1318) may include one or more transceiver(s) (1326) that may include an RF transmitter circuit and / or receiver circuit that use the antenna(s) (1328) of the network device (1318) to facilitate signaling (e.g., signaling (1334)) to and / or from the network device (1318) to other devices (e.g., wireless device (1302)) according to the corresponding RATs.

[0150] The network device (1318) may include one or more antenna(s) (1328) (e.g., one, two, four or more). In embodiments having multiple antenna(s) (1328), the network device (1318) may perform MIMO, digital beamforming, analog beamforming, beam steering, etc. as described.

[0151] A network device (1318) may include one or more interface(s) (1330). The interface(s) (1330) may be used to provide input to the network device (1318) or to provide output from it. For example, a network device (1318) that is a base station may include interface(s) (1330) composed of transmitters, receivers, and other circuitry (e.g., other than the transceiver(s) (1326) / antenna(s) (1328) already described) that enable the base station to communicate with other equipment in a core network and / or enable the base station to communicate with external networks, computers, databases, etc. for the purposes of operations, management, and maintenance of the base station or other equipment operablely connected thereto.

[0152] The network device (1318) may include an additional condition signaling module (1332). The additional condition signaling module (1332) may be implemented through hardware, software, or a combination thereof. For example, the additional condition signaling module (1332) may be implemented as a processor, circuit, and / or instructions (1324) stored in memory (1322) and executed by the processor(s) (1320). In some examples, the additional condition signaling module (1332) may be integrated within the processor(s) (1320) and / or transceiver(s) (1326). For example, the additional condition signaling module (1332) may be implemented by a combination of software components (e.g., executed by a DSP or general processor) and hardware components (e.g., logic gates and circuits) within the processor(s) (1320) or transceiver(s) (1326).

[0153] The additional condition signaling module (1332) may be used in various embodiments of the present disclosure, e.g., any one of FIG. 1, FIG. 2, and / or FIG. 3. The additional condition signaling module (1332) is configured to transmit one or more additional conditions to the UE that are used to select an AI / ML model for training from a set of compatible AI / ML models. In some cases, an AI / ML module selected based on the transmitted additional conditions may be used for inference.

[0154] The embodiments considered herein include a device comprising means for performing one or more elements of any one of method (400), method (600), method (800) and method (1000). Such a device may be, for example, a device of a UE (e.g., a wireless device (1302) which is a UE as described herein).

[0155] The embodiments considered herein include one or more non-transient computer-readable media comprising instructions, the instructions causing an electronic device to perform one or more elements of any one of method (400), method (600), method (800), and method (1000) upon execution of instructions by one or more processors of the electronic device. Such non-transient computer-readable media may be, for example, the memory of a UE (e.g., the memory (1306) of a wireless device (1302) which is a UE as described herein).

[0156] The embodiments considered herein include a device comprising logic, modules, or circuits for performing one or more elements of any one of method (400), method (600), method (800), and method (1000). Such a device may be, for example, a device of a UE (e.g., a wireless device (1302) which is a UE as described herein).

[0157] The embodiments considered herein include a device comprising one or more processors and one or more computer-readable media comprising instructions, wherein the instructions, when executed by one or more processors, cause one or more processors to perform one or more elements of any one of method (400), method (600), method (800), and method (1000). Such a device may be, for example, a device of a UE (e.g., a wireless device (1302) which is a UE as described herein).

[0158] The embodiments considered in this specification include a signal that is related to one or more elements of any one of method (400), method (600), method (800) and method (1000) or as described therein.

[0159] The embodiments considered herein include a computer program or computer program product comprising instructions, and execution of the program by a processor causes the processor to perform one or more elements of any one of method (400), method (600), method (800) and method (1000). The processor may be a processor of a UE (e.g., processor(s) (1304) of a wireless device (1302) that is a UE as described herein). These instructions may be located on and / or in the processor, for example, in the memory of a UE (e.g., memory (1306) of a wireless device (1302) that is a UE as described herein).

[0160] The embodiments considered herein include a device comprising means for performing one or more elements of any one of method (500), method (700), method (900) and method (1100). Such a device may be, for example, a device of a base station (e.g., a network device (1318) which is a base station as described herein).

[0161] The embodiments considered herein include one or more non-transient computer-readable media comprising instructions, the instructions causing an electronic device to perform one or more elements of any one of method (500), method (700), method (900), and method (1100) when the instructions are executed by one or more processors of the electronic device. Such non-transient computer-readable media may be, for example, the memory of a base station (e.g., the memory (1322) of a network device (1318) that is a base station as described herein).

[0162] The embodiments considered herein include a device comprising logic, modules, or circuits for performing one or more elements of any one of method (500), method (700), method (900), and method (1100). Such a device may be, for example, a device of a base station (e.g., a network device (1318) which is a base station as described herein).

[0163] The embodiments considered herein include a device comprising one or more processors and one or more computer-readable media comprising instructions, wherein the instructions, when executed by one or more processors, cause one or more processors to perform one or more elements of any one of method (500), method (700), method (900), and method (1100). Such a device may be, for example, a device of a base station (e.g., a network device (1318) which is a base station as described herein).

[0164] The embodiments considered in this specification include a signal that is related to one or more elements of any one of method (500), method (700), method (900) and method (1100) or as described therein.

[0165] The embodiments considered herein include a computer program or computer program product comprising instructions, and the execution of the program by a processing element causes the processing element to perform one or more elements of any one of method (500), method (700), method (900) and method (1100). The processor may be a processor of a base station (e.g., processor(s) (1320) of a network device (1318) that is a base station as described herein). These instructions may be located, for example, on the memory of a base station (e.g., memory (1322) of a network device (1318) that is a base station as described herein) and / or within the processor.

[0166] For one or more embodiments, at least one of the components described in one or more of the prior art drawings may be configured to perform one or more operations, techniques, processes, and / or methods as described herein. For example, a baseband processor as described herein in relation to one or more of the prior art drawings may be configured to operate according to one or more of the examples described herein. For other examples, circuitry associated with a UE, base station, network element, etc. as described above in relation to one or more of the prior art drawings may be configured to operate according to one or more of the examples described herein.

[0167] Any of the embodiments described above may be combined with any other embodiments (or combinations of embodiments) unless otherwise clearly indicated. The foregoing description of one or more embodiments is for illustrative and descriptive purposes only and is not intended to be comprehensive or to limit the scope of the embodiments to the exact form disclosed. Modifications and variations may be obtained from the practice of possible or various embodiments in consideration of the above teachings.

[0168] The embodiments and implementations of the systems and methods described herein may include various operations that can be implemented as machine-executable instructions to be executed by a computer system. The computer system may include one or more general-purpose or special-purpose computers (or other electronic devices). The computer system may include hardware components that include specific logic for performing operations, or may include a combination of hardware, software, and / or firmware.

[0169] It should be recognized that the systems described herein include descriptions of specific embodiments. These embodiments may be combined into single systems, partially combined into other systems, divided into multiple systems, or divided or combined in other ways. Additionally, it is considered that parameters, attributes, aspects, etc. of one embodiment may be used in other embodiments. It is recognized that parameters, attributes, aspects, etc. are described in only one or more embodiments for clarity, and that parameters, attributes, aspects, etc. may be combined with or replaced by parameters, attributes, aspects, etc. of other embodiments unless specifically disclaimed herein.

[0170] It is well understood that the use of personally identifiable information must follow privacy policies and practices that are generally recognized as meeting or exceeding industry or government requirements for maintaining user privacy. In particular, personally identifiable information data must be managed and processed to minimize the risk of unintended or unauthorized access or use, and the nature of authorized use must be clearly indicated to users.

[0171] Although the foregoing has been described in some detail for the sake of clarity, it will be apparent that specific changes and modifications may be made without departing from the principles of the invention. It should be noted that there are many alternative ways to implement both the processes and apparatuses described herein. Accordingly, the embodiments shall be regarded as illustrative rather than restrictive, and the description is not limited to the details given herein but may be modified within the scope and equivalents of the appended claims.

Claims

Claim 1 A method of user equipment (UE), comprising: transmitting to a base station a capability report indicating one or more supported AI / ML features capable of performing artificial intelligence (AI) / machine learning (ML) model training; receiving from the base station a model configuration message identifying one or more compatible AI / ML models, wherein each of the one or more compatible AI / ML models is compatible with at least one of the one or more supported AI / ML features; receiving from the base station one or more active network-side additional conditions; selecting an AI / ML model selected for training from the one or more compatible AI / ML models based on the one or more active network-side additional conditions; generating a training dataset for the selected AI / ML model based on signaling received from the base station; transmitting the training dataset to a server to cause training of the selected AI / ML model at the server; and transmitting to the base station a notification message indicating that the selected AI / ML model is available to the UE. Claim 2 A method according to claim 1, wherein one or more additional active network-side conditions include dataset information. Claim 3 A method according to paragraph 2, wherein the dataset information includes one or more of a dataset category, a dataset type, a dataset size, and dataset freshness. Claim 4 A method according to claim 1, wherein one or more additional active network-side conditions include network antenna information. Claim 5 In paragraph 4, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 6 A method according to claim 1, wherein the one or more additional active network-side conditions include one or more of network bandwidth and network frequency. Claim 7 A method according to claim 1, wherein one or more additional active network-side conditions include network state information. Claim 8 In claim 7, the method wherein the network status information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 9 A method according to claim 1, wherein one or more active network-side additional conditions are received in the UE in an encrypted format, and the method further comprises the step of decrypting the one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station. Claim 10 A method according to claim 1, wherein the step of selecting an AI / ML model for training from one or more compatible AI / ML models is additionally based on the memory constraints of the UE. Claim 11 A method according to claim 1, wherein the step of selecting an AI / ML model selected for training from one or more compatible AI / ML models is additionally based on computational constraints of the UE. Claim 12 A method according to claim 1, wherein the step of selecting an AI / ML model for training from one or more compatible AI / ML models is additionally based on computational constraints of the server. Claim 13 A method according to claim 1, wherein the model configuration message further includes additional condition compatibility information for one or more compatible AI / ML models, and the step of selecting an AI / ML model selected for training from one or more compatible AI / ML models based on additional active network-side conditions includes the step of comparing the additional active network-side conditions with the additional condition compatibility information for one or more compatible AI / ML models. Claim 14 A method according to claim 1, further comprising the step of transmitting one or more additional active UE-side conditions to the base station. Claim 15 A method according to claim 1, wherein the notification message further indicates that the selected AI / ML model is stored in the UE. Claim 16 A method according to claim 1, wherein the notification message further indicates that the selected AI / ML model is stored on the server. Claim 17 A method of a base station comprising: receiving from a user device (UE) a capability report indicating one or more supported AI / ML features that enable the UE to perform artificial intelligence (AI) / machine learning (ML) model training; transmitting to the UE a model configuration message identifying one or more compatible AI / ML models, wherein each of the one or more compatible AI / ML models is compatible with at least one of the one or more supported AI / ML features; transmitting one or more active network-side additional conditions to the UE; performing signaling with the UE to enable the UE to generate a training dataset for a selected AI / ML model among the one or more compatible AI / ML models; and receiving from the UE a notification message indicating that the selected AI / ML model is available to the UE. Claim 18 In paragraph 17, the above one or more additional active network-side conditions include dataset information, a method. Claim 19 In paragraph 18, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 20 In paragraph 17, the above one or more additional active network-side conditions include network antenna information, a method. Claim 21 In paragraph 20, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 22 In paragraph 17, the above one or more additional active network-side conditions include one or more of network bandwidth and network frequency, a method. Claim 23 In paragraph 17, the above one or more additional active network-side conditions include network state information, a method. Claim 24 In paragraph 23, the method wherein the network status information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 25 A method according to claim 17, further comprising the step of encrypting one or more active network-side additional conditions into an encrypted format using an encryption key corresponding to the base station, wherein the one or more active network-side additional conditions are transmitted by the base station in the encrypted format. Claim 26 A method according to claim 17, wherein the model configuration message further includes additional condition compatibility information for one or more compatible AI / ML models. Claim 27 A method according to claim 17, further comprising the step of receiving one or more additional active UE-side conditions from the above UE. Claim 28 In paragraph 17, the above notification message further indicates that the selected AI / ML model is stored in the UE. Claim 29 In paragraph 17, the above notification message further indicates that the selected AI / ML model is stored on a server, a method. Claim 30 A method of a user device (UE), comprising: transmitting to a base station a capability report indicating one or more supported AI / ML features capable of performing artificial intelligence (AI) / machine learning (ML) model training; receiving from the base station auxiliary information identifying one or more active AI / ML features, wherein each of the one or more active AI / ML features comprises one of the one or more supported AI / ML features; receiving from the base station one or more active network-side additional conditions; designing an AI / ML model for training based on the one or more active AI / ML features and the one or more active network-side additional conditions; generating a training dataset for the AI / ML model based on signaling received from the base station; transmitting the training dataset to a server to cause training of the AI / ML model at the server; and transmitting to the base station a notification message indicating that the AI / ML model is available for use by the UE. Claim 31 In paragraph 30, the above one or more additional active network-side conditions include dataset information, a method. Claim 32 In paragraph 31, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 33 In paragraph 30, the above one or more additional active network-side conditions include network antenna information, a method. Claim 34 In paragraph 33, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 35 In paragraph 30, the above one or more additional active network-side conditions include one or more of network bandwidth and network frequency, a method. Claim 36 In paragraph 30, the above one or more additional active network-side conditions include network state information, a method. Claim 37 In paragraph 36, the method wherein the network status information comprises one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 38 A method according to claim 30, wherein one or more active network-side additional conditions are received in the UE in an encrypted format, and the method further comprises the step of decrypting the one or more active network-side additional conditions using a decryption key corresponding to the base station. Claim 39 In paragraph 30, the method wherein one or more active AI / ML features include one or more of channel state information (CSI) compression features and beam management features. Claim 40 A method according to claim 30, further comprising the step of transmitting one or more additional active UE-side conditions to the base station. Claim 41 In paragraph 30, the method further indicates that the notification message indicates that the selected AI / ML model is stored in the UE. Claim 42 In paragraph 30, the above notification message further indicates that the selected AI / ML model is stored on the server. Claim 43 A method of a base station comprising: receiving from a user device (UE) a capability report indicating one or more supported AI / ML features that enable the UE to perform training of an artificial intelligence (AI) / machine learning (ML) model; transmitting to the UE auxiliary information identifying the one or more active AI / ML features, wherein each of the one or more active AI / ML features includes one of the one or more supported AI / ML features; transmitting one or more active network-side additional conditions to the UE; performing signaling with the UE to enable the UE to generate a training dataset for an AI / ML model designed in the UE based on the one or more active AI / ML features and the one or more active network-side additional conditions; and receiving from the UE a notification message indicating that the AI / ML model is available to the UE. Claim 44 In paragraph 43, the above one or more additional active network-side conditions include dataset information, a method. Claim 45 In paragraph 44, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 46 In paragraph 43, the above one or more additional active network-side conditions include network antenna information, a method. Claim 47 In paragraph 44, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 48 In paragraph 43, the above one or more additional active network-side conditions include one or more of network bandwidth and network frequency, a method. Claim 49 In paragraph 43, the above one or more additional active network-side conditions include network state information, a method. Claim 50 In paragraph 49, the method wherein the network status information comprises one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 51 A method according to claim 43, further comprising the step of encrypting one or more active network-side additional conditions into an encrypted format using an encryption key corresponding to the base station, wherein the one or more active network-side additional conditions are transmitted by the base station in the encrypted format. Claim 52 In paragraph 43, the method wherein one or more active AI / ML features include one or more of channel state information (CSI) compression features; and beam management features. Claim 53 A method according to claim 43, further comprising the step of receiving one or more additional active UE-side conditions from the above UE. Claim 54 In paragraph 43, the method wherein the notification message further indicates that the selected AI / ML model is stored in the UE. Claim 55 In paragraph 43, the above notification message further indicates that the selected AI / ML model is stored on a server, a method. Claim 56 A method of a user device (UE), comprising: receiving a model configuration message from a base station indicating one or more supported artificial intelligence (AI) / machine learning (ML) functions for inference and one or more model switch conditions; receiving one or more active network-side additional conditions from the base station; selecting an AI / ML model from one or more compatible AI / ML models based on the one or more active network-side additional conditions and the one or more model switch conditions; receiving a signaling from the base station comprising a dataset for use with the selected AI / ML model; and performing inference using the selected AI / ML model and the received dataset corresponding to the selected AI / ML model. Claim 57 In paragraph 56, the above one or more additional active network-side conditions include dataset information, a method. Claim 58 In paragraph 57, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 59 In paragraph 56, the above one or more additional active network-side conditions include network antenna information, a method. Claim 60 In claim 59, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 61 In paragraph 56, the above one or more additional active network-side conditions include one or more of network bandwidth and network frequency, a method. Claim 62 In paragraph 56, the above one or more additional active network-side conditions include network state information, a method. Claim 63 In paragraph 62, the method wherein the network status information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 64 In claim 56, the method further comprises the step of receiving one or more active network-side additional conditions in an encrypted format at the UE, and decrypting the one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station. Claim 65 A method according to claim 56, further comprising the step of determining that a configured event is satisfied, and the step of transmitting an uplink (UL) request message to the base station indicating a request for one or more additional active network-side conditions based on the determination that the configured event is satisfied. Claim 66 A method according to paragraph 65, wherein the UL request message comprises one or more of a UL radio resource control (RRC) message, a medium access control-control element (MAC-CE) message, and a dedicated scheduling request (SR) message. Claim 67 In paragraph 65, the above-mentioned configured event is a method in which the monitored inference error is greater than or equal to a configured threshold. Claim 68 A method according to claim 56, further comprising the step of transmitting a notification message to the base station identifying the selected AI / ML model used to perform the inference. Claim 69 In paragraph 56, the method wherein the one or more model switch conditions include one or more of a performance monitor metric threshold, a network-side additional condition, and a UE-side additional condition. Claim 70 In paragraph 56, the method wherein one or more active network-side additional conditions are received from the base station in one or more of a radio resource control (RRC) message and a media access control-control element (MAC-CE) message. Claim 71 A method of a base station comprising: transmitting a model configuration message to a user device (UE) indicating one or more supported artificial intelligence (AI) / machine learning (ML) functions for inference and one or more model switch conditions; transmitting one or more active network-side additional conditions to the UE; and transmitting a signaling to the UE comprising one or more datasets corresponding to the one or more supported AI / ML functions. Claim 72 In paragraph 71, the method wherein one or more additional active network-side conditions include dataset information. Claim 73 In paragraph 72, the method wherein the dataset information includes one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 74 In paragraph 71, the method wherein one or more additional active network-side conditions include network antenna information. Claim 75 In paragraph 74, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 76 In paragraph 71, the method wherein one or more additional active network-side conditions include one or more of network bandwidth and network frequency. Claim 77 In paragraph 71, the method wherein one or more additional active network-side conditions include network state information. Claim 78 In paragraph 77, the method wherein the network status information comprises one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 79 A method according to claim 71, further comprising the step of receiving a notification message from the UE, the message including a selected AI / ML model used to perform inference in the UE. Claim 80 A method according to claim 71, further comprising the step of encrypting one or more active network-side additional conditions into an encrypted format using an encryption key corresponding to the base station, wherein the one or more active network-side additional conditions are transmitted by the base station in the encrypted format. Claim 81 A method according to claim 71, further comprising the step of receiving an uplink (UL) request message from the UE requesting one or more active network-side additional conditions, wherein the one or more active network-side additional conditions are transmitted to the UE based on the UL request message. Claim 82 A method according to claim 81, wherein the UL request message received from the UE comprises one or more of a radio resource control (RRC) message, a media access control-control element (MAC-CE) message, and a dedicated scheduling request (SR) message. Claim 83 In paragraph 71, the method wherein the one or more model switch conditions include one or more of a performance monitor metric threshold, a network-side additional condition, and a UE-side additional condition. Claim 84 In paragraph 71, the method wherein one or more active network-side additional conditions include one or more of a radio resource control (RRC) message and a media access control-control element (MAC-CE) message. Claim 85 A method of a user device (UE), comprising: receiving a model configuration message from a base station indicating one or more compatible artificial intelligence (AI) / machine learning (ML) models for inference; receiving one or more active network-side additional conditions from the base station; identifying a subset of AI / ML models from the one or more compatible AI / ML models for inference based on the one or more active network-side additional conditions; transmitting a notification message containing the subset of AI / ML models to the base station; receiving an AI / ML model identifier (ID) for a selected AI / ML model to be used for inference from the base station; and performing the inference using the selected AI / ML model. Claim 86 In paragraph 85, the above one or more additional active network-side conditions include dataset information, a method. Claim 87 In paragraph 86, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 88 In paragraph 85, the above one or more additional active network-side conditions include network antenna information, a method. Claim 89 In claim 88, the method wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 90 In paragraph 85, the above one or more additional active network-side conditions include one or more of network bandwidth and network frequency, a method. Claim 91 In paragraph 85, the above one or more additional active network-side conditions include network state information, a method. Claim 92 In claim 91, the method wherein the network status information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 93 A method according to claim 85, wherein one or more active network-side additional conditions are received in the UE in an encrypted format, and the method further comprises the step of decrypting the one or more active network-side additional conditions from the encrypted format using a decryption key corresponding to the base station. Claim 94 A method according to claim 85, further comprising the step of transmitting the notification message, which includes a subset of the AI / ML models, to a server. Claim 95 A method according to claim 85, further comprising the step of transmitting a request for one or more additional active network-side conditions to the base station, wherein the one or more additional active network-side conditions are received in response to the request for the one or more additional active network-side conditions. Claim 96 In paragraph 85, the method wherein one or more active network-side additional conditions include one or more of a radio resource control (RRC) message and a media access control-control element (MAC-CE) message. Claim 97 In paragraph 85, the method wherein the notification message includes an uplink (UL) radio resource control (RRC) message. Claim 98 A method of a base station comprising: transmitting a model configuration message to a user device (UE) indicating one or more compatible artificial intelligence (AI) / machine learning (ML) for inference; transmitting one or more active network-side additional conditions to the UE; receiving a notification message from the UE identifying a subset of AI / ML models, wherein the subset of AI / ML models is a subset of one or more compatible AI / ML models; identifying a selected AI / ML model to be used for inference based on the notification message; and transmitting a display message to the UE including an AI / ML model identifier (ID) for the selected AI / ML model to be used for inference. Claim 99 In paragraph 98, the method wherein one or more additional active network-side conditions include dataset information. Claim 100 In claim 99, the method wherein the dataset information comprises one or more of a dataset category, a dataset type, a dataset size, and dataset recency. Claim 101 In paragraph 98, the method wherein one or more additional active network-side conditions include network antenna information. Claim 102 A method according to claim 101, wherein the network antenna information comprises one or more of antenna panel information, virtual mapping, port number, antenna polarization type, antenna downtilt angle, and beam pattern. Claim 103 In paragraph 98, the method wherein one or more additional active network-side conditions include one or more of network bandwidth and network frequency. Claim 104 In paragraph 98, the above one or more additional active network-side conditions include network state information, a method. Claim 105 A method according to claim 104, wherein the network status information includes one or more of network traffic load, network resource utilization, network computation load, and network congestion level. Claim 106 A method according to claim 98, further comprising the step of encrypting one or more active network-side additional conditions into an encrypted format using an encryption key corresponding to the base station, wherein the one or more active network-side additional conditions are transmitted by the base station in the encrypted format. Claim 107 A method according to claim 98, further comprising the step of receiving an uplink (UL) request message from the UE requesting one or more active network-side additional conditions, wherein the one or more active network-side additional conditions are transmitted to the UE based on the UL request message. Claim 108 In paragraph 98, the method wherein one or more active network-side additional conditions include one or more of a radio resource control (RRC) message and a media access control-control element (MAC-CE) message. Claim 109 In paragraph 98, the method wherein the notification message includes an uplink (UL) radio resource control (RRC) message. Claim 110 An apparatus comprising means for performing the method of any one of claims 1 to 109. Claim 111 Computer-readable media comprising instructions, wherein the instructions enable an electronic device to perform the method of any one of claims 1 to 109 when the instructions are executed by one or more processors of the electronic device. Claim 112 A device comprising logic, modules, or circuits for performing the method of any one of claims 1 to 109. Claim 113 A baseband processor for user equipment (UE) configured to perform one or more elements of any one of paragraphs 1 through 16, paragraphs 30 through 42, paragraphs 56 through 70, and paragraphs 85 through 97. Claim 114 A baseband processor for a base station configured to perform one or more elements of any one of paragraphs 17 through 29, paragraphs 43 through 55, paragraphs 71 through 84, and paragraphs 98 through 109.

Images