A control plane intelligent agent system and a method for requesting intelligent agent services from a core network through a control plane

Abstract

The application discloses a control plane intelligent agent system and a method for requesting intelligent agent services from a core network through a control plane, wherein the system adds a control plane intelligent agent gateway to an existing 5G core network control plane architecture, and an intelligent agent adaptation layer is arranged between a NAS layer and an A2A layer of a UE, the intelligent agent adaptation layer is used for converting A2A messages of the A2A layer of the UE into intelligent agent adaptation layer messages and delivering the intelligent agent adaptation layer messages to a network side through NAS message encapsulation, the control plane intelligent agent gateway is used for analyzing the intelligent agent adaptation layer messages and reassembling A2A messages, and the reassembled A2A messages are routed to target intelligent agents according to target intelligent agent identifiers. By introducing the control plane intelligent agent gateway and the intelligent agent adaptation layer arranged above the NAS layer, the problem that a control plane protocol layer and an upper intelligent agent application layer protocol stack do not match can be solved without greatly modifying an existing architecture.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and specifically relates to a control plane intelligent agent system and a method for requesting intelligent agent services from the core network through the control plane. Background Technology

[0002] Currently, 6G is driving the transformation of networks from connection-oriented to service-oriented. This transformation places new demands on the design of mobile core networks and architectures, especially the traditional Control Plane (CP). As the embodiment of the control plane, traditional Non-Access Stratum (NAS) protocols undertake control functions such as Mobility Management (MM) and Session Management (SM), and are inherently connection-oriented, centralized, and slow to evolve. This has become a bottleneck restricting the deployment of service-oriented 6G networks, especially for new services represented by AI agents.

[0003] Unlike basic AI assistants that merely respond to commands, AI agents can proactively plan and execute a series of steps, invoke various tools, and learn from their own experience to complete complex tasks. When AI agents and related solutions are introduced into network operations, especially in the control plane, the network will transform from a rigid system managed by static rules into an intelligent platform with self-organizing, self-optimizing, and self-healing capabilities, thereby providing diversified services in complex tasks and resource planning and execution.

[0004] Currently, major manufacturers are directly involved in defining the core network architecture for intelligent agents. Huawei and HiSilicon are driving the "Intelligent Agent Core Network for 6G" theme in the 3GPP SA2 working group. Their participation in these early technical working groups (WTs) aims to ensure that their architecture proposals, especially those involving embedding native AI capabilities in NFs (Network Functions) to achieve intelligent operation, are incorporated into the 6G Release 21 basic specifications. By proactively defining architecture enhancements, Huawei strives to build a framework that makes its data-centric distributed approach superior to other alternative models. However, currently only the framework design exists, treating network elements as intelligent agents and communicating via A2A (Agent to Agent) protocols, but it has several shortcomings:

[0005] Protocol stack mismatch: A2A protocols, such as those based on HTTP / REST, JSON-RPC, SSE / WebSocket, etc., all assume that the underlying transport layer based on the TCP / IP protocol stack provides reliable, connection-oriented, or message-based transmission services. However, control plane protocols such as NAS / RRC (Radio Resource Control) are designed for signal interaction between devices and the network; they are not themselves general-purpose IP (Internet Protocol) transport channels.

[0006] Payload limitations: Control plane messages such as NAS PDUs (Protocol Data Units) are subject to strict size limitations, typically ranging from a few hundred bytes to slightly more than a thousand bytes, which makes them unable to carry larger A2A message bodies.

[0007] Addressing and Routing: Since there are no IP addresses, the network control plane cannot directly route an agent's request to another agent. A completely new addressing and routing mechanism is needed.

[0008] Missing network functions: Traditional network functions in the core network, such as AMF (Access & Mobility Management Function) or MME (Mobility Management Entity), do not understand the A2A protocol; they only handle standard mobility management and session management signaling. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention provides a control plane intelligent agent system and a method for requesting intelligent agent services from the core network through the control plane, solving problems such as protocol stack mismatch, limited payload, inability to directly route, and lack of network functionality.

[0010] The present invention adopts the following technical solution.

[0011] The first aspect of the present invention provides a control plane intelligent agent system, which adds a control plane intelligent agent gateway to the existing 5G core network control plane architecture, and deploys an intelligent agent adaptation layer between the NAS layer and the A2A layer of the UE. The intelligent agent adaptation layer is used to convert the A2A messages of the UE's A2A layer into intelligent agent adaptation layer messages and encapsulate them through NAS messages and transmit them to the network side. The control plane intelligent agent gateway is used to parse the intelligent agent adaptation layer messages and restore the A2A messages, and route the restored A2A messages to the target intelligent agent according to the target intelligent agent identifier in the A2A messages.

[0012] Optionally, the agent adaptation layer is used to convert the UE's A2A layer A2A messages into agent adaptation layer messages and encapsulate them through NAS messages, including:

[0013] The serialization module is used to serialize A2A messages into binary data;

[0014] The compression module compresses binary data;

[0015] The segmentation and reassembly module is used to divide the compressed A2A message into multiple A2A data fragments, add a header message to each A2A data fragment to obtain the corresponding agent adaptation layer message, and package each agent adaptation layer message into the corresponding NAS protocol data unit to obtain the NAS message.

[0016] Optionally, the agent adaptation layer message can be encapsulated into a NAS message, and the payload container type can be set in the header of the NAS message.

[0017] Optionally, when the control plane intelligent agent gateway parses the intelligent agent adaptation layer messages to obtain the A2A message, it includes:

[0018] The parsing module is used to parse messages from the agent adaptation layer.

[0019] The reassembly module is used to reassemble the parsed agent adaptation layer messages into A2A messages;

[0020] The decompression module is used to decompress A2A messages;

[0021] The deserialization module is used to deserialize the decompressed A2A message to obtain the complete A2A message.

[0022] A second aspect of the present invention provides a method for requesting agent services from a core network via control plane using the above-described control plane agent system, comprising:

[0023] The agent adaptation layer converts the A2A request message generated by the UE's local agent into an agent adaptation layer message and encapsulates it into a NAS message;

[0024] The UE's NAS layer sends NAS messages as uplink NAS transmission messages to the base station via the air interface;

[0025] The base station will forward the uplink NAS transmission message to the AMF in the core network;

[0026] AMF checks the payload container type of the uplink NAS transmission message. If the payload type corresponds to an A2A request, it parses the agent adaptation layer message from the uplink NAS transmission message and forwards it to the control plane agent gateway.

[0027] The control plane agent gateway parses the agent adaptation layer message and restores it to obtain the A2A message. Based on the target agent identifier in the A2A message, it routes the A2A message to the target agent.

[0028] Optionally, the agent adaptation layer converts the A2A request message generated by the UE's local agent into an agent adaptation layer message and encapsulates it into a NAS message, including:

[0029] Serialize the A2A request message into binary data;

[0030] Compress binary data;

[0031] The compressed A2A message is divided into multiple A2A data fragments. A header message is added to each A2A data fragment to obtain the corresponding agent adaptation layer message. Each agent adaptation layer message is then packaged into a corresponding NAS protocol data unit.

[0032] Optionally, the agent adaptation layer message can be encapsulated into a NAS message, and the payload container type can be set to A2A_DATA in the header of the NAS message.

[0033] Optionally, the control plane agent gateway parses the agent adaptation layer messages to obtain A2A messages, including:

[0034] Parse the messages from the agent adaptation layer;

[0035] Reassemble the parsed agent adaptation layer messages into A2A messages;

[0036] Decompress the A2A message;

[0037] The decompressed A2A message is deserialized to obtain the complete A2A message.

[0038] Optionally, the control plane intelligent agent gateway maintains a mapping table of target intelligent agent identifiers to actual network addresses, and routes the parsed A2A messages to the target intelligent agent based on the mapping table.

[0039] Optionally, the target intelligent agent may be located on an application server within the operator's network, on the public internet, or on another UE connected in the same way.

[0040] Optionally, the control plane intelligent agent gateway can be registered with the NRF as a standard NF. The AMF or SMF can query the NRF to determine the corresponding control plane intelligent agent gateway and interact with the control plane intelligent agent gateway through a standard service interface to achieve A2A message forwarding.

[0041] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0042] Protocol decoupling: This invention introduces the A2A-GF (A2A Gateway Function, control plane intelligent agent gateway) network element and an intelligent agent adaptation layer that interacts with the A2A-GF, located above the UE (User Equipment) NAS layer. By setting the NAS payload container type, the NAS protocol is extended, while the A2A protocol itself does not need to be changed. All the complexities of the underlying transmission are shielded by the adaptation layer and A2A-GF, which solves the problem of mismatch between the control plane network layer protocol and the upper-layer intelligent agent application layer protocol stack without significantly modifying the existing architecture.

[0043] Service-oriented integration: In 5G networks, the new A2A-GF can be smoothly integrated into the existing architecture as a service-oriented network element without modifying existing network element connections, reducing deployment costs. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0045] Figure 1 This is a schematic diagram of an intelligent agent adaptation layer above the NAS layer in a control plane provided by an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of a control plane protocol stack provided in an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of an AAL layer protocol stack design provided in an embodiment of the present invention;

[0048] Figure 4 This is a schematic diagram of a UE-side message workflow provided in an embodiment of the present invention;

[0049] Figure 5 This is a schematic diagram of a network-side message workflow provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0051] like Figure 1 As shown, the existing 5G system adopts an architecture design that separates control plane and user plane functions. The control plane protocol stack consists of the application layer, NAS layer, RRC layer (Radio Resource Control) layer and common underlying radio protocol stack from top to bottom. Among them, control plane services can be accessed by application client 1 through control plane invocation. The NAS layer is the carrier layer for signaling interaction between UE and core network control plane.

[0052] The user plane protocol stack consists of the application layer, the IP layer (Internet Protocol), the SDAP (Service Data Adaptation Protocol) layer, and a common underlying radio protocol layer shared with the control plane. The SDAP layer and the IP layer are responsible for the encapsulation and forwarding of user data. The user plane services of application client 2 can be directly accessed through this user plane protocol stack.

[0053] The common underlying wireless protocol layer includes PDCP (Packet Data Convergence Protocol), RLC (Radio Link Control), MAC (Medium Access Control), and PHY (Physical Layer). These four layers are shared by the control plane and the user plane and are responsible for data encryption, segmentation / reassembly, scheduling, and physical transmission over the air interface.

[0054] like Figure 1 , Figure 2 As shown, this invention provides a control plane intelligent agent system. A control plane intelligent agent gateway is added to the existing 5G core network control plane architecture. An intelligent agent adaptation layer is deployed between the NAS layer and the A2A layer of the UE. The intelligent agent adaptation layer is used to convert A2A messages from the UE's A2A layer into NAS messages and transmit them to the control plane intelligent agent gateway. The control plane intelligent agent gateway is used to parse the intelligent agent adaptation layer messages and reconstruct the A2A messages. Based on the target intelligent agent identifier in the A2A message, the parsed A2A message is routed to the target intelligent agent.

[0055] like Figure 2As shown in the embodiment of the present invention, a control plane protocol stack is provided, wherein the communication protocol stack between the user equipment and the 5G RAN (Radio Access Network) / AN (Access Network) is, from bottom to top, PHY, MAC, RLC, PDCP, and RRC. The communication protocol stack between the 5G RAN / AN and the 5G core network AMF / SMF (Session Management Function) network elements is, from bottom to top, PHY, MAC, IP, SCTP (Stream Control Transmission Protocol), and NGAP (Next Generation Application Protocol). The communication protocol stack between the 5G core network AMF / SMF network elements and the A2A-GF is, from bottom to top, PHY, MAC, IP, and the AMF service interface layer Namf or the SMF service interface layer Nsmf or NEF (Network Exposure Function), or the N4 interface layer between the 5G SMF network elements and the user plane gateway. The communication protocol stack between A2A-GF and the agent server, from bottom to top, consists of the PHY, MAC, IP, and A2A layer protocol stacks.

[0056] It should be noted that Namf, Nsmf, NEF, and N4 are four different interface protocols used by A2A-GF to communicate with 5G core network elements AMF / SMF in three different deployment scenarios. When A2A-GF is deployed on the core network control plane, it communicates with AMF / SMF through the Namf / Nsmf interface inside the core network; when A2A-GF is deployed outside the core network, it communicates with AMF / SMF indirectly through the NEF interface; when A2A-GF is deployed on the core network data plane, for example, when it shares a site with a UPF (User Plane Function), it can communicate with AMF / SMF through the N4 interface.

[0057] It should be noted that PHY, MAC, RRC, PDCP, RRC, NAS, SCTP, NG-AP, Namf, NEF, N4, A2A, etc. are all technical terms in the existing communication technology field, and their functions will not be described in detail here.

[0058] This invention specifically includes:

[0059] I. Definition of the intelligent agent adaptation layer.

[0060] like Figure 2As shown, an AAL (Agent Adaptation Layer) is defined / introduced above the NAS layer. It will perform the following functions:

[0061] Message serialization and compression: Compression methods such as serialization are used to reduce the payload of the A2A protocol.

[0062] Segmentation and Reassembly: Large A2A messages are segmented into multiple smaller data blocks and then reassembled at the receiving end.

[0063] AID (Agent ID) addressing: Identifies the source agent and target agent in A2A messages.

[0064] like Figure 4 A schematic diagram of an AAL layer protocol stack design provided for an embodiment of the present invention is shown below, and its main workflow is as follows:

[0065] 1) Message serialization and compression:

[0066] Serialization: Serializes the complex structures of A2A protocols (such as JSON and XML) into a more compact binary format. Examples include: Protobuf (Protocol Buffers), CBOR (Concise Binary Object Representation), or custom TLV (Type-Length-Value) formats. This is crucial for reducing payload size.

[0067] Compression: Compressing the serialized binary data to further reduce its size.

[0068] 2) Segmentation and Reorganization:

[0069] Due to the size limitations of NAS PDUs, larger A2A messages must be split into multiple smaller data blocks.

[0070] The adaptation layer needs to add header information to each shard, including sequence number, total number of shards, transaction ID, etc., to obtain the corresponding agent adaptation layer message so that the receiver can correctly reassemble them.

[0071] The embodiments disclosed herein can solve the payload limitation problem through serialization, compression, and segmentation.

[0072] 3) Address and Identification:

[0073] Design a non-IP Agent ID (AID) addressing scheme. This AID can be carried in the A2A message to identify the source agent and the target agent.

[0074] An AID can be a globally unique string, similar to a URI (Uniform Resource Identifier). The core network needs to be able to parse this AID and find the target Agent.

[0075] like Figure 2 As shown, if application client 1 needs to request intelligent agent-related services from the core network, it needs to connect to the AAL layer through a control plane call (CP call). If it is just other ordinary services, it can connect through the IP layer of UP (UserPlane). Both application client 1 and application client 2 retain this traditional UP plane connection.

[0076] 2. Extend the NAS protocol to include A2A protocol payloads.

[0077] Specifically, a new NAS payload container type is defined to inform the AMF / SMF that the NAS message contains A2A data requiring special handling, rather than regular signaling. This new NAS payload container type is defined to encapsulate A2A protocols within the NAS layer as NAS payloads.

[0078] With this new NAS load container type, AMF / SMF can identify NAS messages related to A2A, parse out agent adaptation layer messages from them, and forward them to the corresponding agent gateway function for further parsing.

[0079] In order for the AMF / SMF to know that the NAS message contains not ordinary communication commands, but A2A data that requires special processing, a special payload container type needs to be specified in the NAS message.

[0080] When the AMF / SMF sees this type, it knows that it should not parse the agent adaptation layer message payload itself, but should forward it to a network function entity that specializes in handling A2A messages.

[0081] Specifically, AAL-DU (Data Unit) needs to be encapsulated in this dedicated payload container, NASPayload Container. This is NAS-level encapsulation and is not part of the AAL layer header, but together they constitute the complete payload transmitted at the NAS layer.

[0082] It is understandable that the NAS layer transmits the payload as: NAS header + message type (A2A_DATA) + AAL-DU content, where the AAL header is already included in the AAL-DU.

[0083] This application does not require the invention of a completely new air interface protocol, but rather defines a new type of NAS payload container and utilizes existing NAS messages as "containers" to carry our A2A data.

[0084] III. Control plane intelligent agent gateway.

[0085] Specifically, a new network element function has been defined / introduced: A2A-GF (A2A Gateway Function, Control Plane Intelligent Agent Gateway), used for:

[0086] Terminate and parse agent adaptation layer messages that contain A2A payloads.

[0087] Reassemble / decompress / deserialize the A2A payload to recover the complete A2A message.

[0088] Parse the target AID and, with the help of different underlying protocols / interfaces, route the recovered A2A message to the final target agent.

[0089] like Figure 3 As shown, a new network function, A2A-GF, is introduced in the core network, especially in 5G's SBA (Service-Based Architecture). Its main functions are:

[0090] 1) Termination and parsing of messages in the agent adaptation layer:

[0091] AMF / SMF parses the agent adaptation layer message from the received A2A NAS payload and forwards it to A2A-GF.

[0092] A2A-GF is the endpoint of this non-IP control plane channel, and it is responsible for extracting the A2A data payload from the agent adaptation layer messages.

[0093] 2) Reverse operation of the adaptation layer:

[0094] The received data fragments are reassembled to restore a complete A2A message.

[0095] The message is decompressed and deserialized to restore it to an application-layer readable A2A protocol format (such as JSON).

[0096] 3) Addressing and routing:

[0097] Parse the target Agent ID (AID) in the A2A message.

[0098] Maintain a mapping table or service discovery mechanism between AIDs and actual network addresses (which can be IP addresses + ports, or service names within another core network). In this way, A2A-GF can find the target agent based on the mapping table or service discovery mechanism.

[0099] The restored A2A message is routed to the final target agent. This target agent may be located on an application server within the operator's network, on the public internet, or on another UE connected in the same way, in which case the A2A-GF is required as a relay.

[0100] In the 5G SBA architecture, A2A-GF can be used as a standard NF to register its services with the NRF (NF Repository Function).

[0101] AMF / SMF can discover the A2A-GF by querying the NRF and interact with it through standard service interfaces such as Nsmf and Namf to forward A2A data. In this way, other existing core networks such as AMF or SMF can find the A2A-GF through the NRF. Namf and Nsmf are the service interfaces provided by AMF and SMF to the outside world in the 5G service-oriented architecture, respectively.

[0102] Combination Figure 4 , Figure 5 As shown, the present invention provides a method for requesting agent services from the core network via control plane using the above-described control plane agent system, comprising the following steps:

[0103] The agent adaptation layer packages the A2A request messages generated by the UE's local agent into NAS messages;

[0104] The UE's NAS layer sends NAS messages as uplink NAS transmission messages to the base station via the air interface;

[0105] The base station will forward the uplink NAS transmission message to the AMF in the core network;

[0106] AMF checks the payload container type of the uplink NAS transmission message. If the payload type corresponds to an A2A request, it parses the agent adaptation layer message from the uplink NAS transmission message and forwards it to the control plane agent gateway.

[0107] The control plane intelligent agent gateway parses the uplink intelligent agent adaptation layer message and restores it to obtain the A2A message. Based on the target intelligent agent identifier in the A2A message, the parsed A2A message is routed to the target intelligent agent.

[0108] Specifically, the entire workflow of a message flow is as follows:

[0109] 1. The UE test message workflow is as follows:

[0110] like Figure 4 As shown, a local agent on the UE side generates an A2A request.

[0111] The A2A adaptation layer (AAL) serializes this request (which may be in JSON format) into a protocol buffer (Protobuf) format, and then splits it into two fragments. Protobuf is an efficient binary serialization format.

[0112] The adaptation layer packages each fragment into a NAS PDU and marks it as Payload Type = A2A_DATA, indicating that the payload container type is A2A_DATA.

[0113] The UE's NAS layer transmits these two NAS PDUs as uplink NAS transmission messages to the base station via the air interface through the lower-layer RRC / PDCP / RLC / MAC / PHY protocol stack. The base station is responsible for transparently transmitting the NAS message to the AMF / SMF.

[0114] 2. Air Interface and RAN:

[0115] The 5G base station gNB transparently transmits these signaling messages to the AMF of the core network, and the gNB itself does not care about the contents of the NAS PDU.

[0116] 3. The core network side message workflow is as follows:

[0117] like Figure 5 As shown, when AMF receives an uplink NAS transmission message and checks its Payload Type, it finds that it is A2A_DATA.

[0118] Knowing that it cannot process this data, AMF parses the agent adaptation layer message from the contents of this NASPDU according to a pre-configured strategy or by querying via NRF and forwards it to A2A-GF.

[0119] A2A-GF receives data fragments of two or more agent adaptation layers from AMF.

[0120] The adaptation layer within A2A-GF reassembles the fragments into a complete binary message based on the transaction ID and sequence number in the fragment header.

[0121] Next, it deserializes the message, restoring it to the original JSON format A2A request.

[0122] The A2A-GF routing engine parses the target AID in the request, queries its service registry, and finds the agent server where the service resides.

[0123] A2A-GF sends an HTTP POST request to the target service over a standard IP network.

[0124] This invention adds an intelligent agent layer or intelligent agent adaptation layer on top of the traditional NAS layer. This layer helps user devices express potential service requests to the mobile core network through control. Furthermore, a new logical entity, the A2A gateway function, is defined on the network side. As a decentralized network function, it is responsible for decapsulating and encapsulating AAL loads and locating appropriate intelligent agent servers on the network side through addressing and routing mechanisms.

[0125] It should be noted that the current 6G standardization work is still in the early research stage. There is no unified conclusion in the industry regarding the networking form, functional division and signaling system of the 6G core network. This invention is based on the 5G architecture and extends it backward compatiblely. Without changing the 3GPP standard protocol stack, it introduces A2A communication capabilities by extending the NAS (Non-Access Stratum) payload container mechanism, and explores the evolution path of the intelligent agent native core network for 6G.

[0126] This invention retains the 5G core network NF architecture and its service interfaces, and adds A2A-GF to the core network. A2A_DATA type service data is carried through the NASPayload Container, and the intelligent agent message encapsulation and parsing after Protobuf encoding is completed via AAL. This approach ensures compatibility with existing 5G terminals and base stations, while also reserving space for a smooth evolution towards the standardization of native intelligent agent communication capabilities in the 6G stage.

[0127] As research into the 6G architecture deepens, related functions are expected to be further decentralized to standard network nodes (NFs) or carried by newly defined 6G dedicated network functions, forming a next-generation core network system that natively supports distributed artificial intelligence, intrinsic intelligence, and global awareness.

[0128] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0129] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0130] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0131] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0132] All parts not covered in this invention are the same as or can be implemented using existing technologies.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A control surface intelligent agent system, characterized in that, A control plane intelligent agent gateway is added to the existing 5G core network control plane architecture. An intelligent agent adaptation layer is deployed between the NAS layer and the A2A layer of the UE. The intelligent agent adaptation layer is used to convert the A2A messages of the UE's A2A layer into intelligent agent adaptation layer messages and encapsulate them through NAS messages before passing them to the network side. The control plane intelligent agent gateway is used to parse the intelligent agent adaptation layer messages and reassemble them to obtain A2A messages. Based on the target intelligent agent identifier in the A2A message, the reconstructed A2A message is routed to the target intelligent agent.

2. The control surface intelligent agent system according to claim 1, characterized in that, The agent adaptation layer is used to convert A2A messages from the UE's A2A layer into agent adaptation layer messages and encapsulate them through NAS messages, including: The serialization module is used to serialize A2A messages into binary data; The compression module compresses binary data; The segmentation and reassembly module is used to divide the compressed A2A message into multiple A2A data fragments, add a header message to each A2A data fragment to obtain the corresponding agent adaptation layer message, and package each agent adaptation layer message into the corresponding NAS protocol data unit to obtain the NAS message.

3. The control surface intelligent agent system according to claim 2, characterized in that, Encapsulate the agent adaptation layer message into a NAS message and set the payload container type in the header of the NAS message.

4. The control surface intelligent agent system according to claim 1, characterized in that, When the control plane intelligent agent gateway parses the intelligent agent adaptation layer messages to obtain A2A messages, it includes: The parsing module is used to parse messages from the agent adaptation layer. The reassembly module is used to reassemble the parsed agent adaptation layer messages into A2A messages; The decompression module is used to decompress A2A messages; The deserialization module is used to deserialize the decompressed A2A message to obtain the complete A2A message.

5. A method for requesting agent services from a core network via control plane, utilizing a control plane agent system as described in any one of claims 1 to 4, characterized in that, The method includes: The agent adaptation layer converts the A2A request message generated by the UE's local agent into an agent adaptation layer message and encapsulates it into a NAS message. The UE's NAS layer sends NAS messages as uplink NAS transmission messages to the base station via the air interface; The base station will forward the uplink NAS transmission message to the AMF in the core network; AMF checks the payload container type of the uplink NAS transmission message. If the payload type corresponds to an A2A request, it parses the agent adaptation layer message from the uplink NAS transmission message and forwards it to the control plane agent gateway. The control plane agent gateway parses the agent adaptation layer message and restores it to obtain the A2A message. Based on the target agent identifier in the A2A message, it routes the A2A message to the target agent.

6. The method for requesting agent services from the core network via control, as described in claim 5, is characterized in that... The agent adaptation layer transforms the A2A request messages generated by the UE's local agent into agent adaptation layer messages and encapsulates them into NAS messages, including: Serialize the A2A request message into binary data; Compress binary data; The compressed A2A message is divided into multiple A2A data fragments. A header message is added to each A2A data fragment to obtain the corresponding agent adaptation layer message. Each agent adaptation layer message is then packaged into a corresponding NAS protocol data unit.

7. The method for requesting agent services from the core network via control according to claim 6, characterized in that, Encapsulate the agent adaptation layer message into a NAS message, and set the payload container type to A2A_DATA in the packet header of the NAS message.

8. The method for requesting agent services from the core network via control, as described in claim 5, is characterized in that... The control plane intelligent agent gateway parses the intelligent agent adaptation layer messages to obtain A2A messages, including: Parse the messages from the agent adaptation layer; Reassemble the parsed agent adaptation layer messages into A2A messages; Decompress the A2A message; The decompressed A2A message is deserialized to obtain the complete A2A message.

9. The method for requesting agent services from the core network via control according to claim 5, characterized in that, The control plane intelligent agent gateway maintains a mapping table of target intelligent agent identifiers to actual network addresses. Based on the mapping table, the control plane intelligent agent gateway routes the parsed A2A messages to the target intelligent agent. The target intelligent agent is located on an application server within the operator's network, on the public Internet, or on another UE connected in the same way.

10. The method for requesting agent services from the core network via control, as described in claim 5, is characterized in that... The control plane intelligent agent gateway is treated as a standard NF and registers its services with the NRF. The AMF or SMF queries the NRF to determine the corresponding control plane intelligent agent gateway and interacts with the control plane intelligent agent gateway through a standard service interface to achieve A2A message forwarding.

Images