Wireless communication management for artificial intelligence application, apparatus, and computer-readable medium
By managing PDU sessions with AI-related information, the method addresses the lack of customization in 5G networks for AI models, improving AI operation efficiency and effectiveness.
Patent Information
- Application Number
- PCT/CN2024/084390
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-03-28
- Publication Date
- 2025-10-02
AI Technical Summary
Current 5G networks lack the capability to provide customized PDU sessions with various AI-related parameters for specific AI models, failing to meet the requirements for AI-related session configuration, monitoring, and prediction.
Implementing a wireless communication method that includes sending and receiving messages between communication nodes with detailed AI-related information such as computing power, response time, model accuracy, and QoS flow identification to manage PDU sessions with per-UE, per-PDU session, or per-QoS flow granularity, facilitating AI/ML operations.
Enables customized management of PDU sessions to meet the specific requirements of AI models, enhancing the efficiency and effectiveness of AI operations within 5G networks.
Smart Images

Figure CN2024084390_02102025_PF_FP_ABST
Abstract
Description
WIRELESS COMMUNICATION MANAGEMENT FOR ARTIFICIAL INTELLIGENCE APPLICATION, APPARATUS, AND COMPUTER-READABLE MEDIUMTECHNICAL FIELD
[0001] This disclosure is generally related to wireless communication, and more particularly to management of wireless communication with implementation of artificial intelligence (AI) technologies.BACKGROUND
[0002] Artificial intelligence (AI) , generally, is a field of research in computer science focusing on the automation of intelligent behavior through machine learning. This field develops and studies methods and software which enable machines to perceive their environment and take actions that maximize their chances of achieving defined goals, with the aim of performing tasks typically associated with human intelligence. Such machines may be called AIs.
[0003] AI has been used together with many different fields, and the development may be used to improve the functionality of wireless communication. The AI can include decision trees, support vector machines, artificial neural networks, ensemble models, generative models, reinforcement leaning models, and / or probabilistic models to name a few.SUMMARY
[0004] This summary is a brief description of certain aspects of this disclosure. It is not intended to limit the scope of this disclosure.
[0005] According to some embodiments of this disclosure, a wireless communication method is disclosed. The method includes sending, by a first communication node to a second communication node, a first message; and receiving, by the first communication node from the second communication node, a second message in response to the first message. At least one of the first message or the second message includes at least one of: computing power value information, computing power offset information, computing response time information, AI computing response time offset information, AI model accuracy information, AI prediction confidence information, parallel processing information, parallel processing offset information, AI or ML (machine learning) functionality information, enhanced offset AI indications, memory information, UE identification information, PDU session identification information; and / or QoS flow identification information.
[0006] According to some embodiments of this disclosure, another wireless communication method is disclosed. The method includes receiving, from a first communication node by a second communication node, a first message; and sending, by the second communication node to the first communication node, a second message in response to the first message. At least one of the first message or the second message includes at least one of: computing power value information, computing power offset information, computing response time information, AI computing response time offset information, AI model accuracy information, AI prediction confidence information, parallel processing information, parallel processing offset information, AI or ML (machine learning) functionality information, enhanced offset AI indications, memory information, UE identification information, PDU session identification information; and / or QoS flow identification information.
[0007] Still another embodiment of this disclosure provides a wireless communication apparatus, including one or more memory units storing one or more programs and one or more processors electrically coupled to the one or more memory units and configured to execute the one or more programs to perform any method or step or their combinations in this disclosure.
[0008] Still another embodiment of this disclosure provides non-transitory computer-readable storage medium, storing one or more programs, the one or more programs being configured to, when performed by at least one processor, cause to perform any method or step or their combinations in this disclosure.
[0009] According to some embodiments of this disclosure, one or more wireless communication methods are further disclosed, the methods include combinations of certain methods, aspects, elements, and steps (either in a generic view or specific view) disclosed in the various embodiments of this disclosure.
[0010] The above and other aspects and their implementations are described in greater detail in the drawings, the descriptions, and the claims.BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Various exemplary embodiments of the present disclosure are described in detail below with reference to the following drawings. The drawings are provided for purposes of illustration only and merely depict exemplary embodiments of the present disclosure to facilitate the understanding of the present disclosure. Therefore, the drawings should not be considered as limiting of the breadth, scope, or applicability of the present disclosure. It should be noted that for clarity and ease of illustration these drawings are not necessarily drawn to scale.
[0012] FIG. 1 shows an exemplary QoS flow architecture;
[0013] FIGS. 2A-2E show various transmission flow diagrams between two different nodes with different types of messages;
[0014] FIG. 3 illustrates a flow chart for failure management of an AI function;
[0015] FIG. 4 illustrates another flow chart for failure management of an AI function;
[0016] FIGS. 5A-5C together illustrate a block diagram of an exemplary wireless communication system.DETAILED DESCRIPTION
[0017] FIG. 1 shows an exemplary QoS flow architecture. The QoS architecture in NG-RAN, both for NR connected to 5GC and for E-UTRA connected to 5GC, is depicted in FIG. 1. At the NAS level, the QoS flow can be the finest granularity of QoS differentiation in a PDU (Protocol Data Unit) session. A QoS flow is identified within a PDU session by a QoS Flow ID (QFI) carried in an encapsulation header over NG-U. For each user equipment (UE) , the core network (5GC) establishes one or more PDU sessions. A PDU session provides end-to-end user plane connectivity between the UE and a specific Data Network (DN) through the User Plane Function (UPF) . A PDU session can support one or more QoS Flows. There can be a one-to-one mapping between QoS flow and QoS profile. In such case, all packets belonging to a specific QoS flow have the same “5G Quality of Service Identifier” .
[0018] Except for NB-IoT and IAB-MT in an SA mode, for each UE, the NG-RAN establishes at least one Data Radio Bearers (DRB) together with an PDU session, and additional DRB (s) for QoS flow (s) of that PDU session can be subsequently configured. If NB-IoT UE supports NG-U data transfer, the NG-RAN may also establish DRBs together with the PDU session. In such case, one PDU session maps to only one DRB. The NG-RAN maps packets belonging to different PDU sessions to different DRBs. Also, NAS level packet filters in the UE and in the 5GC associate UL (uplink) and DL (downlink) packets with QoS flows, and likewise, AS-level mapping rules in the UE and in the NG-RAN associate UL and DL QoS flows with DRBs. NG-RAN and 5GC ensure quality of service (such as reliability and target delay) by mapping packets to appropriate QoS flows and DRBs. There is a 2-step mapping of IP-flows to QoS flows (NAS) and from QoS flows to DRBs (Access Stratum) .
[0019] Artificial Intelligence or Machine Learning (AI / ML) function is wildly used in the human’s daily life and will be also deeply involved in the next generation communication system. Different AI models may have various requirements in the communication network (such as data types, data transmissions, latency, burst arrival data, data throughput, and so on) . At current stage, 5G network cannot provide a customized PDU session with various AI related parameters for a specific AI model to fulfill the requirement on AI related session configuration, monitoring, and prediction.
[0020] According to some embodiments, a wireless communication method is disclosed. The method includes sending, by a first communication node to a second communication node, a first message; and receiving, by the first communication node from the second communication node, a second message in response to the first message. According to some embodiments, another wireless communication method is disclosed. The method includes receiving, from a first communication node by a second communication node, a first message; and sending, by the second communication node to the first communication node, a second message in response to the first message. At least one of the first message or the second message includes at least one of: computing power value information, computing power offset information, computing response time information, AI computing response time offset information, AI model accuracy information, AI prediction confidence information, parallel processing information, parallel processing offset information, AI or ML functionality information, one or more enhanced offset AI indications, memory information, ; UE identification information, PDU session identification information, and / or QoS flow identification information.
[0021] FIGS. 2A-2E show various transmission flow diagrams between two different nodes with different types of messages. As shown in FIG. 2A, the first communication node is core network (CN) , and the second communication node is a RAN (Radio Access Network) node, or a base station, BS. The CN first sends a NGAP message A to the RAN node for an AI PDU session management purpose. This NGAP message may be either a newly-defined NGAP message or an existing message (such as a PDU SESSION RESOURCE SETUP REQUEST message, PDU SESSION RESOURCE MODIFY REQUEST message, or HANDOVER REQUEST message) . The message A may include at least one of the following information: UE identification information; PDU session identification information; and / or QoS flow identification information.
[0022] The UE identification information can include a UE information list. The UE information list may include a UE ID. The UE identification information can be used to identify a specific UE. The PDU session identification information can include a PDU session information list. The PDU session identification information can be used to identify a particular PDU session. The QoS flow identification information can include a QoS flow information list. The QoS flow identification information can be used to identify a particular QoS flow. With the UE identification information, PDU session identification information, and / or QoS flow identification information, the message A can be associated with one or more UEs, PDU sessions, and / or QoS flows. Additionally, the message A can include at least one of computing power value information; computing power offset information; computing response time information; AI (artificial intelligence) computing response time offset information; AI model accuracy information; AI prediction confidence information; parallel processing information; parallel processing offset information; AI or ML functionality information; one or more enhanced offset AI indications; and / or memory information. The detailed of the information will be described in this disclosure.
[0023] In the second step, the RAN node sends message B, such as a NGAP message, in response to message A. The NGAP message B may be either a newly-defined NGAP message or an existing message (such as a PDU SESSION RESOURCE SETUP RESPONSE message, PDU SESSION RESOURCE MODIFY RESPONSE message, or HANDOVER REQUEST ACKNOWLEDGE message) .
[0024] FIG. 2B shows a similar transmission between a central unit of a RAN node (RAN node-CU) and a distributed unit of a RAN node (RAN node-DU) . As shown in FIG. 2B, the first communication node is a RAN node-CU, and the second communication node is a RAN node-DU. The RAN node-CU first sends a F1AP message A to the RAN node-DU for an AI PDU session management purpose. This F1AP message may be either a newly-defined F1AP message or an existing message (such as a UE CONTEXT SETUP REQUEST or UE CONTEXT MODIFICATION REQUEST) . The message A may include at least one of the following information: UE identification information; PDU session identification information; and / or QoS flow identification information.
[0025] The UE identification information can be a UE information list. The UE information list may include a UE ID. The UE identification information can be used to identify a specific UE. The PDU session identification information can include a PDU session information list. The PDU session identification information can be used to identify a particular PDU session. The QoS flow identification information can include a QoS flow information list. The QoS flow identification information can be used to identify a particular QoS flow. With the UE identification information, PDU session identification information, and / or QoS flow identification information, the message A can be associated with one or more UEs, PDU sessions, and / or QoS flows. Additionally, the message A can include at least one of computing power value information; computing power offset information; computing response time information; AI computing response time offset information; AI model accuracy information; AI prediction confidence information; parallel processing information; parallel processing offset information; AI or ML functionality information; one or more enhanced offset AI indications; and / or memory information. The detailed of the information will be described in this disclosure.
[0026] In the second step, the RAN node-DU sends message B, such as an F1AP message, in response to message A. The F1AP message B may be either a newly-defined F1AP message or an existing message (such as a UE CONTEXT SETUP RESPONSE or UE CONTEXT MODIFICATION RESPONSE) .
[0027] FIG. 2C shows a similar transmission between a control plane of a central unit of a RAN node (RAN node-CU-CP) and a user plane of a central unit of a RAN node (RAN node-CU-UP) . As shown in FIG. 2C, the first communication node is a RAN node-CU-CP, and the second communication node is a RAN node-CU-UP. The RAN node-CU-CP first sends a E1AP message A to the RAN node-CU-UP for an AI PDU session management purpose. This E1AP message may be either a newly-defined F1AP message or an existing message (such as a BEARER CONTEXT SETUP REQUEST or BEARER CONTEXT MODIFICATION REQUEST) . The message A may include at least one of the following information: UE identification information; PDU session identification information; and / or QoS flow identification information.
[0028] The UE identification information can be a UE information list. The UE information list may include a UE ID. The UE identification information can be used to identify a specific UE. The PDU session identification information can include a PDU session information list. The PDU session identification information can be used to identify a particular PDU session. The QoS flow identification information can include a QoS flow information list. The QoS flow identification information can be used to identify a particular QoS flow. With the UE identification information, PDU session identification information, and / or QoS flow identification information, the message A can be associated with one or more UEs, PDU sessions, and / or QoS flows. Additionally, the message A can include at least one of computing power value information; computing power offset information; computing response time information; AI computing response time offset information; AI model accuracy information; AI prediction confidence information; parallel processing information; parallel processing offset information; AI or ML functionality information; one or more enhanced offset AI indications; and / or memory information. The detailed of the information will be described in this disclosure.
[0029] In the second step, the RAN node-CU-UP sends a message B, such as an E1AP message, in response to message A. The E1AP message B may be either a newly-defined F1AP message or an existing message (such as a BEARER CONTEXT SETUP RESPONSE, or BEARER CONTEXT MODIFICATION RESPONSE) .
[0030] FIG. 2D shows a similar transmission between two RAN nodes. The RAN node 1 first sends a XnAP message A to the RAN node 2. This XnAP message may be either a newly-defined XnAP message or an existing message (such as a HANDOVER REQUEST, S-NODE ADDITION REQUEST, or S-NODE MODIFICATION REQUEST) . The message A may include at least one of the following information: UE identification information; PDU session identification information; and / or QoS flow identification information.
[0031] The UE identification information can be a UE information list. The UE information list may include a UE ID. The UE identification information can be used to identify a specific UE.The PDU session identification information can include a PDU session information list. The PDU session identification information can be used to identify a particular PDU session. The QoS flow identification information can include a QoS flow information list. The QoS flow identification information can be used to identify a particular QoS flow. With the UE identification information, PDU session identification information, and / or QoS flow identification information, the message A can be associated with one or more UEs, PDU sessions, and / or QoS flows. Additionally, the message A can include at least one of computing power value information; computing power offset information; computing response time information; AI computing response time offset information; AI model accuracy information; AI prediction confidence information; parallel processing information; parallel processing offset information; AI or ML functionality information; one or more enhanced offset AI indications; memory information. The detailed of the information will be described in this disclosure.
[0032] In the second step, the RAN node 2 sends a message B, such as an XnAP message, in response to message A. The XnAP message B may be either a newly-defined XnAP message or an existing message (such as a HANDOVER REQUEST ACKNOWLEDGE, S-NODE MODIFICATION REQUEST ACKNOWLEDGE, or S-NODE ADDITION REQUEST ACKNOWLEDGE) .
[0033] FIG. 2E shows another similar transmission between two RAN nodes. The RAN node 1 first sends a XnAP message A to the RAN node 2. This XnAP message may be either a newly-defined XnAP message or an existing message (such as a RETRIEVE UE CONTEXT REQUEST) .
[0034] In the second step, the RAN node 2 sends a message B, such as an XnAP message, in response to message A. The XnAP message B may be either a newly-defined XnAP message or an existing message (such as a RETRIEVE UE CONTEXT RESPONSE) . The message B may include at least one of the following information: UE identification information; PDU session identification information; and / or QoS flow identification information.
[0035] The UE identification information can be a UE information list. The UE information list may include a UE ID. The UE identification information can be used to identify a specific UE. The PDU session identification information can include a PDU session information list. The PDU session identification information can be used to identify a particular PDU session. The QoS flow identification information can include a QoS flow information list. The QoS flow identification information can be used to identify a particular QoS flow. With the UE identification information, PDU session identification information, and / or QoS flow identification information, the message A can be associated with one or more UEs, PDU sessions, and / or QoS flows. Additionally, the message B can include at least one of computing power value information; computing power offset information; computing response time information; AI computing response time offset information; AI model accuracy information; AI prediction confidence information; parallel processing information; parallel processing offset information; AI or ML functionality information; one or more enhanced offset AI indications; memory information. The detailed of the information will be described in this disclosure.
[0036] To facilitate the management of PDU sessions for the adoption of the AI / ML technologies, the following information can be carried or communicated between two or more communication nodes for the management of PDU sessions for the implementation of the AI / ML application. The newly-introduced parameters in may be configured as per-UE, per-PDU session, or per-QoS flow granularity. The parameters here may be configured as a requirement information or assistance information. That is, the parameters carried by the messages can be used to indicate the requirements for AI / ML operations; alternatively or additionally, the parameters can be used for an informative purpose. When the parameters are used to indicate requirements, the receiving entity fulfills the received configuration. The requirement, based on the configured granularity, may be associated with one or more UEs, PDU sessions, or QoS flows. When the parameters are used to indicate assistance information here, it means the receiving entity may use the received information as a reference. Likewise, the assistance information, based on the configured granularity, may be associated with one or more UEs, PDU sessions, or QoS flows.
[0037] Computing Power Value Information
[0038] Computing power value information includes information used to indicate the computing power related to either one UE, one or multiple PDU sessions, or one or multiple QoS flows. Using the computing power value information, the UE or network may transfer the data, according to one or different AI functions with specific computing power levels, to one or more related PDU sessions or QoS flows.
[0039] In some examples, the computing power value information includes information related to a number of floating-point operations per second (FLOPS) , which is a unit used in the field of AI. Based on the different value ranges, this unit may also be transferred to 10n FLOPS (wherein n is an integer) , such as kFLPOS (or 103 FLOPS, kilo FLOPS) , or MFLOPS (or 106 FLOPS, mega FLOPS) .
[0040] In some examples alternatively or additionally, the computing power value information includes information related to a number of tera operations per second related information (TOPS) , a unit be used in the AI field. Based on the different value ranges, this unit may also be transferred to 10n TOPS (wherein n is an integer. ) , such as kTOPS (or kilo TOPS) , MTOPS (or mega TOPS) , or GOPS (giga TOPS) .
[0041] In some examples alternatively or additionally, the computing power value information includes a number of standard GPUs (graphics processing units) . The computing power may be evaluated by the number of standard GPUs. In some examples alternatively or additionally, the computing power value information includes a number of standard CPUs (Central Processing Units, or Computing Processing Units) .
[0042] In some examples alternatively or additionally, the computing power value information includes level information. The computing power value information may be classified as several levels by using different kinds of value ranges. At least one of the following alternatives may be used. The computing power value information may indicate a value in the range of (N, M) , where N and M are integers. Here, different numbers stand for different levels of computing power. N or M may or may not be included in the level range.
[0043] In some examples alternatively or additionally, the computing power value information includes one code point in an enumerated IE (information element) . Also, several code points may be used to classify different levels of computing power, such as levels of low, medium, and high. In some examples alternatively or additionally, the computing power value information includes a bit string. Each bit in the bit string indicates a level of computing power value.
[0044] Computing Power Offset Information
[0045] Computing power offset information can be used to show the offset of the computing power requirement related to either one UE, one or multiple PDU sessions, or one or multiple QoS flows. By using this computing power value information and / or the configured computing power information, the UE or network may transfer the data, according to one or different AI functions with the computing power requirement, to related PDU sessions or QoS flows.
[0046] For example, the computing power offset information can be indicated by a value M1. Assuming the configured computing power value information is N1, N1 being an integer or floating point number, the computing power requirement of a PDU session or QoS flow may be a range between N1-M1 and N1+M1 with or without the boundary values N1-M1 and N1+M1.
[0047] For example, the computing power offset information can be indicated by a value M1%. Assuming the configured computing power value information is N1, N being an integer or float, the computing power requirement of a PDU session or QoS flow may be a range between N1* (100-M1%) and N1* (100+M1%) with or without the boundary values N1* (100-M1%) and N1* (100+M1%) .
[0048] Alternatively or additionally, the computing power offset information can be indicated by an enumerated information element or a bit string. The computing power offset information can indicate different kinds of offset by using different values in the enumerated IE or bit string.
[0049] Computing Response Time Information
[0050] Additionally or alternatively, (AI) computing response time information can be included in the message. The (AI) computing response time information may be used for the AI model training and / or AI model inference. Table 1 below shows different examples of starting points (starting timing) and end points (stop timing) that can be used to define two boundaries of the response time..
[0051] Table 1
[0052] For example, if the starting point and the end point are chosen from the first row , the response time is counted from the time point when an entity starts to send data to an AI model for AI function (s) to the time when the AI model starts to send the responsive data-such as the responsive data for the AI functions-to the entity. For example, if the starting point and the end point are chosen from the second row, the response time is counted from a time when an entity complete sending data to an AI model for AI function (s) to a time when the entity receives AI response data. For example, if the starting point and the end point are chosen from the fifth row, the response time is counted from a time when the AI model starts its operation, such as training or inference to a time when the AI model completes the operation. For example, if the starting point and the end point are chosen from the sixth row, the response time is represented by a time when the AI model complete the operations, such as AI training or inference. Also, the start times and stop times at different rows in Table 1 can be combined to generate different combination for the definition of the two ends of the response time. For example, the “AI model starts to receive the data for AI function (s) ” at row 3 can serve as the start point, while the “AI model completes working” at row 5 serve as the end point of the “response time. ”
[0053] The computing response time information can be used to show AI function (s) response time by using the time or time range information for either one UE, one or multiple PDU sessions, or one or multiple QoS flows. By using this computing response time information, UE or network may transfer the data, according to different AI functions with a specific AI response time, to related PDU session (s) or QoS flow (s) .
[0054] The computing response time information can include a time information having a common unit used to a time or time range. Based on the different values, this unit may also be indicated by at least one of the following units, including hours, minutes, seconds, milliseconds, microseconds, or nanoseconds. Alternatively or additionally, the computing response time information be classified as several levels by using different kinds of value ranges. For example, the computing response time information can include one value in the range of (N, M) , N and M being integers. Different numbers here stand for different levels of AI response time, and N or M may or may not be included in the level range. The computing response time information can be represented by one code point in an enumerated IE. Additionally, several code points may be used to classify different levels of the AI response time. The computing response time information can be represented by one or more bits in a bit string. Several bits in the bit string can be used to classify different levels (such as low, medium, or high) of the AI response time.
[0055] AI Computing Response Time Offset Information
[0056] The AI computing response time offset information is used to show the offset of the AI response time requirement related to either one UE, one or multiple PDU sessions, or one or multiple QoS flows. By using the AI computing response time offset and / or the configured AI response time information, UE or network may transfer the data, according to one or different AI functions with an AI response time requirement, to related PDU session (s) or QoS flow (s) .
[0057] For example, the AI computing response time offset information can be indicated by a value M2. Assuming the configured AI computing response time information is N2, N2 being an integer or floating point number, the AI computing response time requirement of a PDU session or QoS flow may be a range between N2-M2 and N2+M2 with or without the boundary values N2-M2 and N2+M2.
[0058] For example, the AI computing response time offset information can be indicated by a value M2%. Assuming the configured computing power value information is N2, N2 being an integer or a floating point number, the AI computing response time requirement of a PDU session or QoS flow may be a range between N2* (100-M2%) and N2* (100+M2%) with or without the boundary values N2* (100-M2%) and N2* (100+M2%) .
[0059] Alternatively or additionally, the AI computing response time offset information can be indicated by an enumerated information element or a bit string. The computing power offset information can indicate different kinds of offsets by using different values in the enumerated IE or bit string.
[0060] AI Model Accuracy Information
[0061] The AI model accuracy information is used to show the accuracy information of an AI function result for either one UE, one or multiple PDU sessions or one or multiple QoS flows. By using the AI model accuracy information, the UE or network may transfer the data, according to different AI functions with specific result accuracy information of an AI function, to related PDU sessions or QoS flows.
[0062] For example, the AI model accuracy information can be represented by N3%, N3 being an integer or a floating point number.
[0063] Alternatively or additionally, the AI model accuracy information can be indicated by a range (N3%, M3%) , N3 and M3 each being an integer or a floating point number. The accuracy of an AI model is then between N3 and M3. The value of N3 and M3 may or may not be included in the accuracy requirement.
[0064] Alternatively or additionally, the AI computing response time offset information can be indicated by a range N3%±M3%, N3 and M3 each being an integer or a floating point number. The accuracy of an AI model is then between (N3-M3) %and (N3+M3) %. The value of N3-M3 and N3+M3 may or may not be included in the accuracy requirement.
[0065] The AI model accuracy information can be represented by one code point in an enumerated IE. Additionally, several code points may be used to classify different levels of the AI accuracy information. The AI model accuracy information can be represented by one or more bits in a bit string. Several bits in the bit string can be used to classify different levels (such as low, medium, or high) of the AI model accuracy information.
[0066] AI Prediction Confidence Information
[0067] The AI prediction confidence information is used to indicate the confidence information of an AI model related to either one UE, one or multiple PDU sessions or one or more multiple QoS flows. The AI related data for one or multiple AI models may be transmitted via such PDU session (s) or one or multiple QoS flow (s) if the AI model confidence is equal or not lower than the confidence requirement.
[0068] For example, the AI prediction confidence information can be represented by N4%, N4 being an integer or a floating point number.
[0069] Alternatively or additionally, the AI prediction confidence information can be indicated by a range (N4%, M4%) , N4 and M4 each being an integer or a floating point number. The AI prediction confidence of an AI model is then between N4 and M4. The value of N4 and M4 may or may not be included in the AI prediction confidence requirement.
[0070] Alternatively or additionally, the AI prediction confidence information can be indicated by a range N4%±M4%, N4 and M4 each being an integer or a floating point number. The AI prediction confidence of an AI model is then between (N4-M4) %and (N4+M4) %. The value of N4-M4 and N4+M4 may or may not be included in the AI prediction confidence requirement.
[0071] The AI prediction confidence information can be represented by one code point in an enumerated IE. Additionally, several code points may be used to classify different levels of the AI prediction confidence information. The AI prediction confidence information can be represented by one or more bits in a bit string. Several bits in the bit string can be used to classify different levels (such as low, medium, or high) of the AI prediction confidence information.
[0072] Parallel Processing Information
[0073] The parallel processing information is used to inform the capability or requirement of how many AI models can be performed or run in parallel by one service via one or multiple PDU session (s) or one or multiple QoS flow (s) . The AI related data for one or multiple AI models may be transmitted from UE / NW via such PDU sessions or QoS flows if this service fulfills the parallel processing requirement. The parallel processing information may also be used to inform the capacity or requirement of how may AI models can performed or run by one UE in parallel.
[0074] For example, the parallel processing information can be indicated by a value M5, M5 being an integer. Assuming the configured parallel processing information is N5, N5 being an integer, the parallel processing requirement of a service may be a range between N5-M5 and N5+M5 with or without the boundary values N5-M5 and N5+M5.
[0075] For example, the parallel processing information can be indicated by a value M5%, M5%being an integer. Assuming the configured parallel processing information is N5%, N5 being an integer, the parallel processing requirement of a service may be a range between N5*(100-M5%) and N5* (100+M5%) with or without the boundary values N5* (100-M5%) and N5* (100+M5%) .
[0076] The parallel processing information can also be indicated by different kind of offsets by using different values in an enumerated IE or a bit string.
[0077] AI / ML Functionality Information
[0078] The AI / ML functionality information is used to indicate which kind (s) of AI model can transmit or is transmitting data via one or multiple PDU session (s) or one or multiple QoS flow (s) . By the AI / ML functionality information, the UE or network may only transfer the data for certain indicated AI model (s) via the one or multiple PDU session (s) or one or multiple QoS flow (s) .
[0079] Alternatively or additionally, the AI / ML functionality information may also be used to indicate which kind (s) of AI model cannot transmit data via one or multiple PDU session (s) or one or multiple QoS flow (s) . By the AI / ML functionality information, the UE or network shall not transfer the data for certain indicated AI model (s) via the one or multiple PDU session (s) or one or multiple QoS flow (s) . Alternatively or additionally the AI / ML functionality information may also be used to inform UE which kinds of AI model can or cannot transmit data from or to this UE.
[0080] For example, the AI models indicated in the AI / ML functionality information can be classified by the AI models’ purposes, such as prediction purpose or performance evaluation purpose. Alternatively or additionally, the AI models indicated in the AI / ML functionality information can be classified by their function fields, such as the function fields of power saving, mobility, RAN overload, QoE, QoS, MDT, IoT, positioning, NTN, UAV, Redcap, NW slicing, and / or UE trajectory. Alternatively or additionally, the AI models indicated in the AI / ML functionality information can be classified by their model types, such as a long short-term memory network or a convolutional neural network. Alternatively or additionally, the AI models indicated in the AI / ML functionality information can be classified by their AI model identifications (ID) . For example, different AI model IDs may be configured to various AI models deployed in a network.
[0081] Enhanced Offset AI Indications
[0082] The enhanced offset AI indications include an indicator used to inform the UE or network that if one or multiple PDU session (s) or one or multiple QoS flow (s) is used to transmit AI related data. The PDU session related parameters and / or QoS parameters, such as data rate, throughput, latency, reliability, and / or energy efficiency, may be enhanced for a dedicated level, value, or percentage for better performance boosting from the configured parameters for common data (other than data for AI) transmission. Detail enhancement may be configured or based on different implementations.
[0083] For example, a specific enhancement M6 (M being, for example, a number, a level, or a percentage) for different classified parameters may be configured to one or multiple PDU session (s) or one or multiple QoS flow (s) . For example, the enhancement for involved PDU session and / or QoS flow related parameters may be based on implementations without further standardization.
[0084] Alternatively or additionally, a specific enhancement M6 (M6 being, for example, a number, a level, or a percentage) for the related parameters for PDU sessions and / or QoS flows may be classified by different IE types. For involved integer related IEs, with the enhanced offset indication for AI #1, the previous configured parameters for common data (other than AI related data) transmission may be enhanced for M6%or M6 value. M6 is integer. For example, at least one of the following IEs may be involved, including: Expected Activity Period, Expected Idle Period, UE Aggregate Maximum Bit Rate Downlink, UE Aggregate Maximum Bit Rate Uplink, UE Slice Maximum Bit Rate Downlink, UE Slice Maximum Bit Rate Uplink, CN Packet Delay Budget DL, CN Packet Delay Budget UL, Priority level, Packet Delay Budget, Packet Error Rate, Max Data Burst Volume, Max Flow Bit Rate DL, Max Flow Bit Rate UL, Guaranteed Flow Bit Rate DL, Guaranteed Flow Bit Rate UL, Max Packet Loss Rate DL, Max Packet Loss Rate UL, and / or QoS Monitoring Reporting Frequency.
[0085] Alternatively or additionally, for level-related IEs, according to the enhanced offset indication for AI #1, the previous configured parameters for common data (other than AI related data) transmission may be enhanced for a M6 level. M6 is an integer. At least one of the following IEs may be involved: Priority Level, RAN Paging Priority
[0086] Alternatively or additionally, a specific enhancement M7 (where M7 may be a value, a level, or a percentage) for an indicated parameter type may be configured to one or multiple PDU session (s) or one or multiple QoS flow (s) . The enhancement for the AI data transmission may be configured with different parameter types, including data rate, throughput, latency, reliability, and / or energy efficiency. Correspondingly, the enhanced offset AI indications include enhancement degree information (the level, the value, or the percentage) and the information of the involved parameter types, such as the data rate, the throughput, the latency, the reliability, and / or the energy efficiency.
[0087] Alternatively or additionally, a specific enhancement M8 (where M8 may be a value, a level, or a percentage) for the indicated parameter (s) may be configured to one / multiple PDU session (s) or one / multiple QoS flow (s) . The indicated parameters may be enhanced for M8. Correspondingly, the enhanced offset AI indications include enhancement degree information (the level, the value, or the percentage) and the information of the involved parameter types. Abut which at least one of the following IEs for PDU sessions or QoS flows may be informed, including Expected Activity Period, Expected Idle Period, UE Aggregate Maximum Bit Rate Downlink, UE Aggregate Maximum Bit Rate Uplink, UE Slice Maximum Bit Rate Downlink, UE Slice Maximum Bit Rate Uplink, CN Packet Delay Budget DL, CN Packet Delay Budget UL, Priority level, Packet Delay Budget, Packet Error Rate, Max Data Burst Volume, Max Flow Bit Rate DL, Max Flow Bit Rate UL, Guaranteed Flow Bit Rate DL, Guaranteed Flow Bit Rate UL, Max Packet Loss Rate DL, Max Packet Loss Rate UL, QoS Monitoring Reporting Frequency, Priority Level, and / or RAN Paging
[0088] Priority.
[0089] Memory Information
[0090] The memory information informs the entity of the memory or buffer or storage that is related to one or multiple AI models for one UE, PDU session (s) and / or QoS flow (s) . For example, the memory information includes AI model classification information and / or memory resource information. AI model classification information includes AI model information used to identity different AI models or different AI model types. The detailed classification information can be found in the AI / ML functionality information IE. The memory resource information is used to indicate the resource information for the memory. The unit used here for the memory resource information may be at least one of the following units: bit, byte, kilobyte, megabyte, gigabyte, terabyte, petabyte, exabyte, zettabyte, and / or yottabyte. Additionally or alternatively, the memory information may include at least one of: a resource limit that can be used for UE, PDU sessions and / or QoS flows, the allocated resource for UE, PDU sessions and / or QoS flows, maximum resource limit that can be used for one or multiple AI models for UE, PDU sessions and / or QoS flows, or the allocated resource for one or multiple AI models for UE, PDU sessions and / or QoS flows.
[0091] Failure Management
[0092] FIG. 3 illustrates a flow chart for failure management of an AI function. FIG. 4 illustrates another flow chart for failure management of an AI function. The examples can be used to handle the failure scenarios for the other embodiments disclosed in this disclosure. In Step 1 of FIG. 3 and FIG. 4, entity 1 (such as a CN, a RAN node CU, a RAN node CU-CP, or a RAN node 1) sends a message A to entity 2 (such as a RAN node, a RAN node DU, a RAN node CU-UP, or a RAN node 2) with AI PDU session management information. The detailed information in message A has been explained elsewhere in this disclosure, and for the sack of conciseness, the detail is omitted here.
[0093] In FIG. 3, due to some reason, entity 2 is not above to perform at least one of the following processes, including: setting up required resource for the AI function for a UE, providing required resource to PDU session (s) , and / or providing required resource to QoS flow (s) , which leads to the occurrence of failure. In response, entity 2 may send a failure message (such as the message B) to entity 1. The message B may include a failure indicator and additional detailed failure information. The failure indicator may indicate the procedure failure and may further indicate the reason of the failure. The detailed failure information may indicate which kinds of configuration cannot be fulfilled. Additionally or alternatively, each of the computing power value information, computing power offset information, computing response time information, AI computing response time offset information, AI model accuracy information, AI prediction confidence information, parallel processing information, parallel processing offset information, AI or ML functionality information, enhanced offset AI indications, and memory information can be given a specific indicator, used to indicate the corresponding failure reasons associated with the configuration.
[0094] Additionally, either newly-defined message or existing message can be used as the message A or message B here. For example, if entity 1 is CN and entity 2 is a RAN node, the existing message B may be: Handover Failure message. If entity 1 is a RAN node CU and entity 2 is a RAN node DU, the existing message B may be UE Context Setup Failure message or UE Context Modification Failure message. If entity 1 is a RAN node CU-CP and entity 2 is a RAN node CU-UP, the existing message B may be a BEARER CONTEXT SETUP FAILURE message or BEARER CONTEXT MODIFICATION FAILURE message. If entity 1 is a RAN node 1 and entity 2 is a RAN node 2, the existing message B may be a HANDOVER PREPARATION FAILURE message, S-NODE MODIFICATION REQUEST REJECT message, or S-NODE ADDITION REQUEST REJECT message.
[0095] Likewise, in the Step 2 of FIG. 4, entity 2 responds with the message B to entity 1. The detailed information in message B has been explained elsewhere in this disclosure, and for the sack of conciseness, the detail is omitted here. In this case, entity 2 can only partially perform at least one of the following processes, including providing required resource to partial PDU session (s) and / or providing required resource to partial QoS flow (s) . For each involved PDU session which cannot be provided with configured resources, at least one of the following information may be informed in message B, including PDU session information and failure information. The PDU session information is used to indicate which PDU session cannot be set up with required configuration. The failure information indicates which kinds of configuration cannot be fulfilled. Additionally or alternatively, each of the computing power value information, computing power offset information, computing response time information, AI computing response time offset information, AI model accuracy information, AI prediction confidence information, parallel processing information, parallel processing offset information, AI or ML (machine learning) functionality information, enhanced offset AI indications, and memory information can be given a specific indicator, used to indicate the corresponding failure reasons associated with the configuration.
[0096] Additionally or alternatively, for each involved QoS flow which cannot be provided with configured resources, at least one of the following information may be informed in message B, including PDU session information, QoS flow information, and failure information. The details of the PDU session information and failure information have been described above. The QoS flow information is used to indicate the QoS flow information that cannot be set up with required configuration for the indicated PDU session. The failure information can be associated with each associated QoS flow.
[0097] FIGS. 5A-5C together illustrate a block diagram of an exemplary wireless communication system 20, in accordance with some embodiments of this disclosure. The system 20 may perform the methods / steps and their combinations or sub-combinations disclosed in this disclosure. The system 20 may include components and elements configured to support operating features that need not be described in detail herein.
[0098] The system 20 may include at least one base station (BS) 110 (or a RAN node) , at least one user equipment (UE) 120, and at least one core network (CN) 130, including the NF, RNDF, AF, AMF, CREF, or other functions. The BS 110 includes a BS transceiver or transceiver module / circuitry 112, a BS antenna system 116, a BS memory or memory module / circuitry 114, a BS processor or processor module / circuitry 113, and a network interface 111. The components of BS 110 may be electrically coupled and in communication with one another as necessary via a data communication bus 190. Likewise, the UE 120 includes a UE transceiver or transceiver module / circuitry 122, a UE antenna system 126, a UE memory or memory module / circuitry 124, a UE processor or processor module / circuitry 123, and an I / O interface 121. The components of the UE 120 may be electrically coupled and in communication with one another as necessary via a data communication bus 190. The UE 120 communicates with the one or more BSs 110 via communication channels therebetween, which can be any wireless channel or other medium known in the art suitable for transmission of data as described herein.
[0099] The CN 130 includes at least one CN transceiver or transceiver module / circuitry 132, at least one CN antenna system 136, at least one CN memory or memory module / circuitry 134, at least one CN processor or processor module / circuitry 133, and at least one network interface 131. The CN can be formed by a distributed system, including multiple devices 130. The components of the CN 130 may be electrically coupled and in communication with one another as necessary via a data communication bus 190. The CN can communicate with one or more application servers and one or more base stations (RAN node) via wired or wireless communication.
[0100] The processor module / circuitry 113, 123, 133 may be implemented, or realized, with a general-purpose processor, a content addressable memory, a digital signal processor, an application specific integrated circuit, a field programmable gate array, any suitable programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof, designed to perform the functions described herein. In this manner, a processor module / circuitry may be realized as a microprocessor, a controller, a microcontroller, a state machine, or the like. A processor module / circuitry may also be implemented as a combination of computing devices, e.g., a combination of a digital signal processor and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a digital signal processor core, or any other such configuration.
[0101] Furthermore, the steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in firmware, in a software module performed by the processor module / circuitry 113, 123, 133, respectively, or in any practical combination thereof. The memory module / circuitry 114, 124, 134 may be realized as RAM memory, flash memory, EEPROM memory, registers, ROM memory, EPROM memory, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. In this regard, the memory module / circuitry 114, 124, 134 may be coupled to the processor module / circuitry 113, 123, 133 respectively, such that the processors module / circuitry 113, 123, 133 can read information from, and write information to, memory modules 114, 124, 134 respectively. The memory module / circuitry 114, 124, 134 may also be integrated into their respective processor module / circuitry 113, 123, 133. In some embodiments, the memory module / circuitry 114, 124, 134 may each include a cache memory for storing temporary variables or other intermediate information during execution of instructions to be performed by the processor module / circuitry 113, 123, 133 respectively. The memory module / circuitry 114, 124, 134 may also each include non-volatile memory for storing instructions to be performed by the processor module / circuitry 113, 123, 133, respectively.
[0102] Various exemplary embodiments of the present disclosure are described herein with reference to the accompanying figures to enable a person of ordinary skill in the art to make and use the present disclosure. The present disclosure is not limited to the exemplary embodiments and applications described and illustrated herein. Additionally, the specific order and / or hierarchy of steps in the methods disclosed herein are merely exemplary approaches. Based upon design preferences, the specific order or hierarchy of steps of the disclosed methods or processes can be re-arranged while remaining within the scope of the present disclosure. Thus, those of ordinary skill in the art would understand that the methods and techniques disclosed herein present various steps or acts in exemplary order (s) , and the present disclosure is not limited to the specific order or hierarchy presented unless expressly stated otherwise.
[0103] This disclosure is intended to cover any conceivable variations, uses, combination, or adaptive changes of this disclosure following the general principles of this disclosure, and includes well-known knowledge and conventional technical means in the art and undisclosed in this application.
[0104] It is to be understood that this disclosure is not limited to the precise structures or operation described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from the scope of this application. The scope of this application is subject only to the appended claims.
[0105] The methods, devices, processing, circuitry, and logic described above may be implemented in many different ways and in many different combinations of hardware and software. For example, all or parts of the implementations may be circuitry that includes an instruction processor or controller, such as a Central Processing Unit (CPU) , microcontroller, or a microprocessor; or as an Application Specific Integrated Circuit (ASIC) , Programmable Logic Device (PLD) , or Field Programmable Gate Array (FPGA) ; or as circuitry that includes discrete logic or other circuit components, including analog circuit components, digital circuit components or both; or any combination thereof. The circuitry may include discrete interconnected hardware components or may be combined on a single integrated circuit die, distributed among multiple integrated circuit dies, or implemented in a Multiple Chip Module (MCM) of multiple integrated circuit dies in a common package, as examples.
[0106] Accordingly, the circuitry may store or access instructions for execution, or may implement its functionality in hardware alone. The instructions may be stored in a tangible storage medium that is other than a transitory signal, such as a flash memory, a Random Access Memory (RAM) , a Read Only Memory (ROM) , an Erasable Programmable Read Only Memory (EPROM) ; or on a magnetic or optical disc, such as a Compact Disc Read Only Memory (CDROM) , Hard Disk Drive (HDD) , or other magnetic or optical disk; or in or on another machine-readable medium. A product, such as a computer program product, may include a storage medium and instructions stored in or on the medium, and the instructions when performed by the circuitry in a device may cause the device to implement any of the processing described above or illustrated in the drawings.
[0107] The implementations may be distributed. For instance, the circuitry may include multiple distinct system components, such as multiple processors and memories, and may span multiple distributed processing systems. Parameters, databases, and other data structures may be separately stored and managed, may be incorporated into a single memory or database, may be logically and physically organized in many different ways, and may be implemented in many different ways. Example implementations include linked lists, program variables, hash tables, arrays, records (e.g., database records) , objects, and implicit storage mechanisms. Instructions may form parts (e.g., subroutines or other code sections) of a single program, may form multiple separate programs, may be distributed across multiple memories and processors, and may be implemented in many different ways. Example implementations include stand-alone programs, and as part of a library, such as a shared library like a Dynamic Link Library (DLL) . The library, for example, may contain shared data and one or more shared programs that include instructions that perform any of the processing described above or illustrated in the drawings, when performed by the circuitry.
[0108] In some examples, each unit, subunit, and / or module of the system may include a logical component. Each logical component may be hardware or a combination of hardware and software. For example, each logical component may include an application specific integrated circuit (ASIC) , a Field Programmable Gate Array (FPGA) , a digital logic circuit, an analog circuit, a combination of discrete circuits, gates, or any other type of hardware or combination thereof. Alternatively or in addition, each logical component may include memory hardware, such as a portion of the memory, for example, that includes instructions executable with the processor or other processors to implement one or more of the features of the logical components. When any one of the logical components includes the portion of the memory that includes instructions executable with the processor, the logical component may or may not include the processor. In some examples, each logical component may just be the portion of the memory or other physical memory that includes instructions executable with the processor or other processor to implement the features of the corresponding logical component without the logical component including any other hardware. Because each logical component includes at least some hardware even when the included hardware includes software, each logical component may be interchangeably referred to as a hardware logical component.
[0109] A second action may be said to be “in response to” a first action independent of whether the second action results directly or indirectly from the first action. The second action may occur at a substantially later time than the first action and still be in response to the first action. Similarly, the second action may be said to be in response to the first action even if intervening actions take place between the first action and the second action, and even if one or more of the intervening actions directly cause the second action to be performed. For example, a second action may be in response to a first action if the first action sets a flag and a third action later initiates the second action whenever the flag is set.
[0110] To clarify the use of and to hereby provide notice to the public, the phrases “at least one of , , …and <N>” or “at least one of , , …<N>, or combinations thereof” or “, , …and / or <N>” or “at least one of , , …or <N>” are defined by the Applicant in the broadest sense, superseding any other implied definitions hereinbefore or hereinafter unless expressly asserted by the Applicant to the contrary, to mean one or more elements selected from the group comprising A, B, …and N. In other words, the phrases mean any combination of one or more of the elements A, B, …or N including any one element alone or the one element in combination with one or more of the other elements which may also include, in combination, additional elements not listed.
Claims
1.A wireless communication method, comprising:sending, by a first communication node to a second communication node, a first message; andreceiving, by the first communication node from the second communication node, a second message in response to the first message, wherein at least one of the first message or the second message includes at least one of:computing power value information;computing power offset information;computing response time information;AI (artificial intelligence) computing response time offset information;AI model accuracy information;AI prediction confidence information;parallel processing information;parallel processing offset information;AI or ML (machine learning) functionality information;one or more enhanced offset AI indications;memory information;UE identification information;PDU session identification information; and / orQoS flow identification information.2.The method according to claim 1, wherein:the computing power value information is configured to indicate a computing power;the computing power offset information is configured to indicate a computing power offset relative to a computing power baseline;the computing response time information is configured to indicate a response time of AI model training or AI model inference;the AI computing response time offset information is configured to indicate a response time offset relative to a response time baseline of AI model training or AI model inference;the AI model accuracy information is configured to indicate the accuracy level of an AI output;the AI prediction confidence information is configured to indicate a confidence level of an AI output;the parallel processing information is configured to indicate a parallel processing capability and / or requirement of multiple AI models;the parallel processing offset information is configured to indicate a parallel processing capability and / or requirement offset relative to a parallel processing information baseline of multiple AI models;the AI or ML functionality information is configured to indicate one or more AI model types, of which data can be transmitted or is transmitting;the one or more enhanced offset AI indications are configured to indicated performance enhancement for AI-related transmission; and / orthe memory information is configured to indicate memory information for an AI-related operation.3.The method according to claim 2, wherein:the computing power value information is configured to indicate the computing power by at least one of: a number of floating point operations or tera operations per second, a number of standard GPUs or CPUs, a bit string, or a computing level.4.The method according to claim 2, wherein:the computing power offset information is configured to indicate the computing power offset relative to the computing power baseline by representation of N1±M1, N1* (100±M1%) , an enumerated information element, or a bit string, wherein N1 represents the computing power baseline and M1 represents the computing power offset.5.The method according to claim 2, wherein:the response time of AI model is counted between two selected time points of the following events:a start of sending data to an AI model;a start of sending responsive data by an AI model;a completion of sending data to an AI model;a time when responsive AI data from an AI model is received;a start of receiving data by an AI model;a start of receiving data for AI function (s) by an AI model;a completion of receiving data by an AI modela completion of receiving data for AI functions from an entity; anda start of training / inference of an AI model.6.The method according to claim 2, wherein:the AI computing response time offset information is configured to indicate the response time offset relative to a response time baseline of the AI model training or AI model inference by representation of N2±M2, N2* (100±M2%) , an enumerated information element, or a bit string, wherein N2 represents the response time baseline and M2 represents the response time offset.7.The method according to claim 2, wherein:the AI model accuracy information is configured to indicate the accuracy level of an AI output by a representation of N3%, N3%~M3%, N3%± M3%, an enumerated information element, or a bit string, N3 and M3 being an integer or a floating point number.8.The method according to claim 2, wherein:the AI prediction confidence information is configured to indicate the confidence level of an AI output by a representation of N4%, N4%~M4%, N4%± M4%, an enumerated information element, or a bit string, N4 and M4 being an integer or a floating point number.9.The method according to claim 2, wherein:the parallel processing information is configured to indicate the parallel processing capability and / or requirement of multiple AI models by:a one-bit indicator, indicating parallel processing characteristics of service whose data are allowed to be transmitted;a parameter, indicating a number of parallel processes of the multiple AI models;an enumerated information element; ora range of a number of parallel processes of the multiple AI models.10.The method according to claim 2, wherein:the parallel processing offset information is configured to indicate the parallel processing capability and / or requirement offset relative to the parallel processing information baseline of multiple AI models by representation of N5±M5, N5* (100±M5%) , an enumerated information element, or a bit string, wherein N5 represents the parallel processing information baseline and M5 represents the parallel processing capability and / or requirement offset.11.The method according to claim 2, wherein:the one or more enhanced offset AI indications include at least one of: enhancement level information or enhancement type information.12.The method according to claim 2, wherein:the memory information includes at least one of: AI model classification information or memory resource information.13.The method according to claim 2, wherein the computing power value information, the computing power offset information, the computing response time information, the AI computing response time offset information, the AI model accuracy information, the AI prediction confidence information, the parallel processing information, the parallel processing offset information, the AI or ML functionality information, the one or more enhanced offset AI indications, and / or the memory information is associated with one or more pieces of UE, one or more PDU sessions, or one or more QoS flows.14.The method according to claim 1, wherein the second message includes a failure indicator, indicating:a request corresponding to the first message cannot be fulfilled;a resource corresponding to the first message cannot be fulfilled; and / ora resource for PDU sessions or QoS flows corresponding to the first messages cannot be fulfilled.15.The method according to claim 1, wherein the second message includes a failure indicator, indicating:PDU session failure information, indicating a PDU session, of which a configuration cannot be set up according to the first message;failure information, indicating the configuration that cannot be set up;QoS flow failure information, indicating a QoS flow, of which a configuration cannot be set up according to the first message.16.The method according to claim 1, wherein:the first communication node is core network, the second communication node is a RAN node, and the first and second messages are NGAP messages;the first communication node is a central unit of a RAN node, the second communication node is a distributed unite of the RAN node, and the first and second messages are F1AP messages;the first communication node is a control plane of a central unit of a RAN node, the second communication node is a user plane of the central unit of the RAN node, and the first and second messages are E1AP messages; orthe first communication node is a first RAN node, the second communication node is a second RAN node, and the first and second messages are XnAP messages.17.A wireless communication method, comprising:receiving, from a first communication node by a second communication node, a first message; andsending, by the second communication node to the first communication node, a second message in response to the first message, wherein at least one of the first message or the second message includes at least one of:computing power value information;computing power offset information;computing response time information;AI (artificial intelligence) computing response time offset information;AI model accuracy information;AI prediction confidence information;parallel processing information;parallel processing offset information;AI or ML (machine learning) functionality information;one or more enhanced offset AI indications;memory information;UE identification information;PDU session identification information; and / orQoS flow identification information.18.The method according to claim 17, wherein:the computing power value information is configured to indicate a computing power;the computing power offset information is configured to indicate a computing power offset relative to a computing power baseline;the computing response time information is configured to indicate a response time of AI model training or AI model inference;the AI computing response time offset information is configured to indicate a response time offset relative to a response time baseline of AI model training or AI model inference;the AI model accuracy information is configured to indicate the accuracy level of an AI output;the AI prediction confidence information is configured to indicate a confidence level of an AI output;the parallel processing information is configured to indicate a parallel processing capability and / or requirement of multiple AI models;the parallel processing offset information is configured to indicate a parallel processing capability and / or requirement offset relative to a parallel processing information baseline of multiple AI models;the AI or ML functionality information is configured to indicate one or more AI model types, of which data can be transmitted or is transmitting;the one or more enhanced offset AI indications are configured to indicated performance enhancement for AI-related transmission; and / orthe memory information is configured to indicate memory information for an AI-related operation.19.The method according to claim 18, wherein:the computing power value information is configured to indicate the computing power by at least one of: a number of floating point operations or tera operations per second, a number of standard GPUs or CPUs, a bit string, or a computing level.20.The method according to claim 18, wherein:the computing power offset information is configured to indicate the computing power offset relative to the computing power baseline by representation of N1±M1, N1* (100±M1%) , an enumerated information element, or a bit string, wherein N1 represents the computing power baseline and M1 represents the computing power offset.21.The method according to claim 18, wherein:the response time of AI model is counted between two selected time points of the following events:a start of sending data to an AI model;a start of sending responsive data by an AI model;a completion of sending data to an AI model;a time when responsive AI data from an AI model is received;a start of receiving data by an AI model;a start of receiving data for AI function (s) by an AI model;a completion of receiving data by an AI modela completion of receiving data for AI functions from an entity; anda start of training / inference of an AI model.22.The method according to claim 18, wherein:the AI computing response time offset information is configured to indicate the response time offset relative to a response time baseline of the AI model training or AI model inference by representation of N2±M2, N2* (100±M2%) , an enumerated information element, or a bit string, wherein N2 represents the response time baseline and M2 represents the response time offset.23.The method according to claim 18, wherein:the AI model accuracy information is configured to indicate the accuracy level of an AI output by a representation of N3%, N3%~M3%, N3%± M3%, an enumerated information element, or a bit string, N3 and M3 being an integer or a floating point number.24.The method according to claim 2, wherein:the AI prediction confidence information is configured to indicate the confidence level of an AI output by a representation of N4%, N4%~M4%, N4%± M4%, an enumerated information element, or a bit string, N4 and M4 being an integer or a floating point number.25.The method according to claim 18, wherein:the parallel processing information is configured to indicate the parallel processing capability and / or requirement of multiple AI models by:a one-bit indicator, indicating parallel processing characteristics of service whose data are allowed to be transmitted;a parameter, indicating a number of parallel processes of the multiple AI models;an enumerated information element; ora range of a number of parallel processes of the multiple AI models.26.The method according to claim 18, wherein:the parallel processing offset information is configured to indicate the parallel processing capability and / or requirement offset relative to the parallel processing information baseline of multiple AI models by representation of N5±M5, N5* (100±M5%) , an enumerated information element, or a bit string, wherein N5 represents the parallel processing information baseline and M5 represents the parallel processing capability and / or requirement offset.27.The method according to claim 18, wherein:the one or more enhanced offset AI indications include at least one of: enhancement level information or enhancement type information.28.The method according to claim 18, wherein:the memory information includes at least one of: AI model classification information or memory resource information.29.The method according to claim 18, wherein the computing power value information, the computing power offset information, the computing response time information, the AI computing response time offset information, the AI model accuracy information, the AI prediction confidence information, the parallel processing information, the parallel processing offset information, the AI or ML functionality information, the one or more enhanced offset AI indications, and / or the memory information is associated with one or more pieces of UE, one or more PDU sessions, or one or more QoS flows.30.The method according to claim 17, wherein the second message includes a failure indicator, indicating:a request corresponding to the first message cannot be fulfilled;a resource corresponding to the first message cannot be fulfilled; and / ora resource for PDU sessions or QoS flows corresponding to the first messages cannot be fulfilled.31.The method according to claim 17, wherein the second message includes a failure indicator, indicating:PDU session failure information, indicating a PDU session, of which a configuration cannot be set up according to the first message;failure information, indicating the configuration that cannot be set up;QoS flow failure information, indicating a QoS flow, of which a configuration cannot be set up according to the first message.32.The method according to claim 17, wherein:the first communication node is core network, the second communication node is a RAN node, and the first and second messages are NGAP messages;the first communication node is a central unit of a RAN node, the second communication node is a distributed unite of the RAN node, and the first and second messages are F1AP messages;the first communication node is a control plane of a central unit of a RAN node, the second communication node is a user plane of the central unit of the RAN node, and the first and second messages are E1AP messages; orthe first communication node is a first RAN node, the second communication node is a second RAN node, and the first and second messages are XnAP messages.33.A wireless communication apparatus, comprising memory circuitry storing one or more programs and one or more processors electrically coupled to the memory circuitry and configured to execute the one or more programs to perform any one of the methods or their combinations or sub-combinations of claims 1 to 32.34.A non-transitory computer-readable storage medium, storing one or more programs, the one or more programs being configured to, when executed by at least one processor, cause to perform any one of the methods or their combinations or sub-combinations of claims 1 to 32.
Citation Information
Patent Citations
Packet delay determination method, terminal device and network device
CN116889012A
Policy enhancement to support group application function (AF) session from artificial intelligence / machine learning (AIML) provider AF with required quality of service (QOS)
US20230199868A1
User equipment machine learning service continuity
US20230422117A1
Method and apparatus for performing handover based on ai model in a wireless communication system
WO2022186458A1