An intelligent agent fusion communication architecture and method based on a neural network model

Abstract

The present invention discloses an intelligent agent fusion communication architecture and method based on a neural network model. The solution of the present invention includes at least one intelligent agent unit for converting audio and video data, a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively; the intelligent agent fusion communication architecture uses a distributed intelligent agent cluster to achieve group call control, and learns link quality control through an intelligent agent neural network model, as well as a method and platform corresponding to the architecture, and introduces artificial intelligence and blockchain into the IMS fusion communication architecture to solve the problem of accessing multi-protocol fusion switching communication of different systems and types in a complex environment of a large number of terminals, form the ability of uninterrupted transmission of multiple links, and support the secure, reliable and high-quality communication transmission of various service application data such as voice, video and conference.

Description

Technical Field

[0001] The present invention belongs to the technical field of communication processing, and particularly relates to an intelligent agent fusion communication architecture and method based on a neural network model. Background Art

[0002] IMS is the network architecture for implementing the large-scale fusion solution in the next-generation communication network (NGN). IMS can implement VoIP services, manage network resources, user resources, and application resources more effectively, improve the intelligence of the network, enable users to cross various networks and use multiple terminals, and experience the integrated communication.

[0003] Traditional telecommunication networks use independent signaling networks to complete processes such as call establishment, routing, and control. The security of the signaling network can ensure the security of the network. Moreover, the transmission uses dedicated lines with time-division multiplexing (TDM), and users communicate through connection-oriented channels, avoiding various eavesdropping and attacks from other terminal users. The IMS network is connected to the Internet. Based on the IP protocol and an open network architecture, it can share service platforms for various different services such as voice, data, and multimedia through various different access methods, increasing the flexibility of the network and the interoperability between terminals. Different operators can effectively and quickly develop and provide various services.

[0004] Since IMS is based on IP, the security requirements of IMS are much higher than those of traditional operators operating on independent networks. Whether it is mobile access or fixed access, the security issues of IMS cannot be ignored. The security threats of IMS mainly come from several aspects: unauthorized access to sensitive data to destroy confidentiality; unauthorized tampering with sensitive data to destroy integrity; interfering with or misusing network services to cause denial of service or reduce system availability; users or the network denying completed operations; unauthorized access to services, etc. It mainly involves the access security of IMS (3GPP TS33.203), including user and network authentication and protecting services between IMS terminals and the network; and the network security of IMS (3GPP TS33.210), dealing with service protection between network nodes of the same operator or different operators. In addition, it also poses threats to user terminal devices and the security of universal integrated circuit cards / IP multimedia service identity modules (UICC / ISIM).

[0005] Therefore, in view of the above technical problems and deficiencies, it is urgent to design and develop an intelligent agent fusion communication architecture and method based on a neural network model. Summary of the Invention

[0006] To overcome the deficiencies and difficulties of the above-mentioned existing technologies, the purpose of the present invention is to provide an intelligent agent fusion communication architecture and method based on a neural network model, and introduce artificial intelligence and blockchain into the IMS fusion communication architecture to solve the problem of multi-protocol fusion switching communication of different systems and types under the complex environment of a large number of terminals, form the ability of uninterrupted transmission of multiple links, and support the secure, reliable, and high-quality communication transmission of various service application data such as voice, video, and conferencing.

[0007] The first object of the present invention is to provide an intelligent agent fusion communication architecture based on a neural network model;

[0008] The second object of the present invention is to provide an intelligent agent fusion communication method based on a neural network model;

[0009] The third object of the present invention is to provide an intelligent agent fusion communication platform based on a neural network model;

[0010] The first object of the present invention is achieved as follows: The intelligent agent fusion communication architecture includes at least one intelligent agent unit for converting audio and video data, and a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively;

[0011] The intelligent agent fusion communication architecture uses a distributed intelligent agent cluster to achieve group call control and learns link quality control through an intelligent agent neural network model.

[0012] Further, the fusion communication switching center unit broadcasts data transmission in the network transport layer by means of TCP, UDP unicast, or multicast.

[0013] Further, the intelligent agent fusion communication architecture is also provided with a synchronization processing unit, which is used for synchronizing and coordinating the cluster, and synchronously collecting, transmitting, and playing voice and video data streams respectively.

[0014] Further, the intelligent agent fusion communication architecture is also provided with a first creation unit, which is used to establish a multi-intelligent agent network constraint consistency model for distributed intelligent agents.

[0015] Further, the first creation unit further includes:

[0016] A generation and acquisition module, which is used to generate intelligent agents and acquire connection addresses corresponding to MECS (Multi-Access Edge Computing Server);

[0017] A first calculation module, which is used to calculate the delay data to the intelligent agent in the fully connected network;

[0018] A self-check generation module, which is used to check whether its own group call status reaches stability and generate action control instruction data corresponding to stability or instability. Among them, the action control includes: establishment control, acquisition control, transmission control, playback control, and termination control.

[0019] The second object of the present invention is achieved as follows: The method includes the following steps:

[0020] Respectively obtain audio and video data corresponding to the intelligent terminal, and in combination with the intelligent agent, perform forwarding processing on the audio and video data in real time;

[0021] According to the data after the forwarding processing by the intelligent agent, through the converged communication switching center cluster and using TCP, UDP unicast or multicast methods, broadcast data transmission at the network transport layer.

[0022] Further, the step of respectively obtaining audio and video data corresponding to the intelligent terminal and performing forwarding processing on the audio and video data in real time in combination with the intelligent agent further includes:

[0023] Obtain first synchronization control data corresponding to the cluster group call communication; among them, the first synchronization control data is the consistency synchronization control data between distributed intelligent agents;

[0024] According to the first synchronization control data, establish a multi-intelligent agent network constraint consistency model of the distributed intelligent agent.

[0025] Further, the step of establishing a multi-intelligent agent network constraint consistency model of the distributed intelligent agent according to the first synchronization control data further includes:

[0026] Generate an intelligent agent and obtain the connection address corresponding to the MECS, establish an intelligent agent TCP connection corresponding to the MECS, and at the same time create a fully connected multi-intelligent agent network;

[0027] According to the shortest path algorithm, calculate and generate the shortest delay data to other intelligent agents in the fully connected network;

[0028] Send its own group call status through the intelligent agent, receive the group call status sent by all neighbors, and at the same time update the group call status in real time according to its own dynamic equation;

[0029] Check whether the group call status of the intelligent agent itself reaches stability, and generate action control instruction data corresponding to stability or instability; among them, the action control includes: establishment control, acquisition control, transmission control, playback control, and termination control.

[0030] The third object of the present invention is achieved as follows: It includes a processor, a memory, and an intelligent agent fusion communication platform control program based on a neural network model. Among them, the processor executes the intelligent agent fusion communication platform control program based on the neural network model, and the intelligent agent fusion communication platform control program based on the neural network model is stored in the memory. The intelligent agent fusion communication platform control program based on the neural network model realizes the intelligent agent fusion communication method based on the neural network model.

[0031] The intelligent agent fusion communication architecture of the present invention includes at least one intelligent agent unit for converting audio and video data, and a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively. The intelligent agent fusion communication architecture uses a distributed intelligent agent cluster to implement group call control, learns link quality control through an intelligent agent neural network model, and corresponding methods and platforms for the architecture. By introducing artificial intelligence and blockchain into the IMS fusion communication architecture, it solves the problem of multi-protocol fusion switching communication of different systems and types under complex environments of a large number of terminals, forms the ability of uninterrupted transmission of multiple links, and supports the secure, reliable, and high-quality communication transmission of various service application data such as voice, video, and conferencing.

[0032] That is to say, the solution of the present invention provides guarantee for high-quality audio and video communication through adaptive learning of link quality control by the intelligent agent neural network model. Each intelligent terminal is interacted by an intelligent agent, and the fusion communication switching center only needs to communicate with the intelligent agent, increasing the reliability of large-scale broadcast communication. The unified use of the SIP protocol between the intelligent agent and the fusion communication switching center reduces the docking between multiple protocols of the fusion communication switching center. The intelligent agent is incorporated into the advanced IMS architecture as a UA, and the required service quality can be obtained, better supporting services such as registration, security, billing, bearer control, and roaming in fusion communication. The intelligent agent serves as a blockchain node, and based on the blockchain, authentication protection, authorization, and encryption are carried out to ensure the security of the system and prevent threats such as theft and tampering. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0034] Figure 1 It is a schematic flowchart of one embodiment of an intelligent agent fusion communication architecture based on a neural network model of the present invention;

[0035] Figure 2-a Schematic diagram of the second embodiment of an intelligent agent fusion communication architecture based on a neural network model according to the present invention;

[0036] Figure 2-b Schematic diagram of the third embodiment of an intelligent agent fusion communication architecture based on a neural network model according to the present invention;

[0037] Figure 3 Schematic diagram of an intelligent agent fusion communication architecture based on a neural network model according to the present invention;

[0038] Figure 4 Schematic diagram of the process of an intelligent agent fusion communication method based on a neural network model according to the present invention;

[0039] Figure 5 Schematic diagram of an intelligent agent fusion communication platform based on a neural network model according to the present invention;

[0040] The realization of the object, functional features and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0041] For a better understanding of the object, technical solution and advantages of the present invention, the present invention will be further described below with reference to the accompanying drawings and specific implementation manners. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0042] The present invention can also be implemented or applied through other different specific examples, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0043] It should be noted that if there are directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention, the directional indications are only used to explain the relative position relationship and movement conditions between components in a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly.

[0044] In addition, if there are descriptions such as "first", "second" in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first", "second" may explicitly or implicitly include at least one of such features. Secondly, the technical solutions between various embodiments can be combined with each other, but it must be based on the ability of those skilled in the art to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0045] Preferably, the intelligent agent fusion communication method based on a neural network model of the present invention is applied to one or more terminals or servers. The terminal is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0046] The terminal may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal can perform human-computer interaction with the customer through a keyboard, a mouse, a remote control, a touchpad, a voice control device, etc.

[0047] The present invention aims to implement an intelligent agent fusion communication architecture, method, and platform based on a neural network model.

[0048] As Figure 4 shown, it is a flowchart of the intelligent agent fusion communication method based on a neural network model provided by an embodiment of the present invention.

[0049] In this embodiment, the intelligent agent fusion communication method based on a neural network model can be applied to a terminal with a display function or a fixed terminal, and the terminal is not limited to a personal computer, a smart phone, a tablet computer, a desktop computer or an all-in-one computer equipped with a camera, etc.

[0050] The intelligent agent fusion communication method based on a neural network model can also be applied to a hardware environment composed of a terminal and a server connected to the terminal through a network. The network includes but is not limited to: a wide area network, a metropolitan area network or a local area network. The intelligent agent fusion communication method based on a neural network model in an embodiment of the present invention can be executed by the server, can be executed by the terminal, or can be jointly executed by the server and the terminal.

[0051] For example, for an intelligent agent fusion communication terminal that needs to perform neural network model-based operations, the intelligent agent fusion communication function provided by the method of the present invention can be directly integrated on the terminal, or a client for implementing the method of the present invention can be installed. Additionally, the method provided by the present invention can also run on devices such as servers in the form of a Software Development Kit (SDK), providing an interface for the intelligent agent fusion communication function based on the neural network model. The terminal or other devices can implement the intelligent agent fusion communication function based on the neural network model through the provided interface.

[0052] The present invention will be further described below in conjunction with the accompanying drawings.

[0053] As Figure 1-5 shown, the present invention provides an intelligent agent fusion communication architecture based on a neural network model.

[0054] The intelligent agent fusion communication architecture includes at least one intelligent agent unit for converting audio and video data, and a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively.

[0055] The intelligent agent fusion communication architecture uses a distributed intelligent agent cluster to implement group call control and learns link quality control through an intelligent agent neural network model.

[0056] The fusion communication switching center unit broadcasts data transmission at the network transport layer through TCP, UDP unicast, or multicast.

[0057] The intelligent agent fusion communication architecture is also provided with a synchronization processing unit, which is used to synchronize and coordinate the cluster and synchronously collect, transmit, and play voice and video data streams respectively.

[0058] The intelligent agent fusion communication architecture is also provided with a first creation unit, which is used to establish a multi-intelligent agent network restraint consistency model for distributed intelligent agents.

[0059] The first creation unit further includes:

[0060] A generation and acquisition module, which is used to generate intelligent agents and acquire connection addresses corresponding to MECS.

[0061] A first calculation module, which is used to calculate the delay data to the intelligent agent in the fully connected network.

[0062] A self-check generation module is used to check whether its own group call status reaches stability and generate action control instruction data corresponding to stability or instability. Among them, the action control includes: establishment control, acquisition control, transmission control, playback control, and termination control.

[0063] Specifically, in the embodiments of the present invention, the link quality control is learned through an intelligent agent neural network model, and complex transmission timing problems are solved based on reliable retransmission, forward error correction (FEC), service priority, end-to-end rate control, etc., to ensure the stability of the transmission link, reduce latency and eliminate the central bottleneck, and reduce network costs, providing guarantee for high-quality audio and video communication; as shown in the appendix Figure 2-a as follows.

[0064] As shown in the appendix Figure 1 As shown in the figure, each intelligent terminal is responsible for interaction by an intelligent agent, and all audio and video data are converted through the intelligent agent. The fusion communication exchange center only needs to communicate with the intelligent agent, and can use TCP, UDP unicast or multicast methods to accelerate the broadcast data transmission at the network transport layer and increase the reliability of large-scale broadcast communication;

[0065] In terms of coordination, according to the characteristics of trunking communication, group call is a key service. In traditional trunking systems, group call control is implemented by centrally controlling terminals through a center. In the present invention, it is implemented by a distributed intelligent agent cluster, and this cluster must be synchronized and coordinated to ensure the synchronous acquisition, transmission, and playback of voice and video data streams.

[0066] Distributed cooperative control includes distributed coordination technology and consistency synchronization control technology. The distributed coordination technology can be implemented by using the mature Zookeeper component. Zookeeper is a distributed coordination service for managing a large number of hosts, with characteristics such as reliability, scalability, and transparency, and can provide services such as naming service, load balancing, configuration management, cluster management, node leader election, distributed queue, distributed lock, data registry, etc. required for coordination between distributed intelligent agents.

[0067] For the consistency synchronization control requirements between distributed intelligent agents in cluster group call communication, the states and their changes between intelligent agents need to reach consistency, which has a high similarity with a multi-intelligent agent network system, and can be solved by establishing a multi-intelligent agent network restraint consistency model for distributed intelligent agents.

[0068] Consider a network containing N nodes. A node is represented as a first-order n-dimensional dynamic system, and the nodes are coupled by interacting state information. The dynamic behavior of the i-th node in the network can be described by the following equation:

[0069]

[0070] where \(x\) i (t)\(\in\mathbb{R}\) n , \(c\) is the coupling strength, \(\Gamma\) is the internal coupling matrix, \(a\) ij is the adjacency matrix, \(u\) i (t)\(\in\mathbb{R}\) n is the linear feedback controller, \(d\) i τ (t) is the controller gain, \(s\) τ (t) is the desired system equilibrium state, \(\tau\) is the desired equilibrium state change time.

[0071] When , through the feedback control input of \(u\) i (t), the network system can achieve pinning consensus. \(L\) is the Laplacian matrix of the system, the diagonal matrix \(D = \text{diag}(d 1 ,\cdots,d N ), \(d i is the control gain of node \(i\).

[0072] When the system reaches the consensus equilibrium, \(x\) r (t)\(\to s\) τ (t), \(x\) i (t)\(\to x\) r (t), so that \(x\) i (t)\(\to s\) τ (t), that is, the system reaches the state of \(s\) τ (t) unanimously. Therefore, changing the state of the pinned node \(s\) τ (t) will cause the states of all nodes in the network to change synchronously and unanimously.

[0073] According to the multi - intelligent agent network consensus model, the following consensus synchronization control algorithm between intelligent agents is proposed:

[0074] 1. Initialize the state. Generate intelligent agents in the MECS, communicate with the center, obtain the connection addresses of other MECSs, and the intelligent agents establish TCP connections with the intelligent agents in other MECSs in the network. As the connections of the intelligent agents, a fully - connected multi - intelligent agent network is formed. Obtain the communication delay \(T\) of the connections by interacting with all multi - intelligent agents through the TCP connections ij .

[0075] 2. The intelligent agent body calculates the shortest delay to other intelligent agents in the fully - connected network according to the Dijkstra shortest - path algorithm, retains the neighbor connection with the shortest delay, and its communication delay \(T\) ij is used as \(a\) ij , forming an adjacency list.

[0076] 3. Through the reserved TCP connections with neighbors, the intelligent agent i sends its own group call status x to all neighbor intelligent agents, i and receives the group call status x j sent by all neighbors j≠i.

[0077] 4. The intelligent agent i updates the group call status x i according to its own dynamic equation (1). The initiating intelligent agent r of the group call serves as a pinning node, and its control gain d can be set to 1. The intelligent agent r adjusts the network consensus by changing s τ after the group call status is established, and automatically collects, transmits, and plays the loop state changes at time intervals τ until termination.

[0078] 5. The intelligent agent i checks whether its own group call status has reached stability (satisfying ‖x i (t + Δτ) - x i (t)‖ < ∈). If it is stable, it performs corresponding actions (such as establishment, collection, transmission, playback, termination, etc.).

[0079] 6. Repeat step 3 at time intervals T.

[0080] Through this cooperative control algorithm, when multiple intelligent agent networks reach consensus and the dynamic update clock T and the longest network delay are much smaller than the state change time τ, the intelligent agents can ensure that the state changes are consistent with the whole network only by updating the state according to the changes of neighbors, without the need to fix a central point for coordination. When the planned state is used as the sampling and output control state of voice, a consistent synchronous acquisition, distribution, and playback network can be formed to realize the group call function.

[0081] As shown in the appendix Figure 1 , the intelligent agents and the converged communication switching center uniformly use the SIP protocol to exchange audio, video, data, and messages based on the same communication system, reducing the docking between multiple protocols in the converged communication switching center;

[0082] The intelligent agents are integrated into the advanced IMS architecture as UAs, enabling various types of terminals to establish peer-to-peer IP communications and obtain the required quality of service, while completing the necessary functions for services, such as registration, security, billing, bearer control, roaming, etc.

[0083] The intelligent agents act as blockchain nodes to perform authentication protection and authorization functions based on the blockchain. All authentication security information is stored on the chain to restrict access by illegal users, ensuring the security of the system. At the same time, through digital encryption technology, the confidentiality of calls is ensured to protect against threats such as theft and tampering.

[0084] Intelligent agents are an important concept in the fields of computer science and artificial intelligence in recent years. It refers to a computing entity that resides in a specific environment, can perceive the environment, and can operate autonomously to represent its designer or user to achieve a series of goals. Intelligent agents can sense, learn, reason, and act, and can imitate the behaviors of human society after training based on a knowledge base, that is, they have intelligence. It has the following key attributes: Autonomy: Intelligent agents can control their own states and behaviors and can operate and run without the intervention of humans or other programs. Perception and response capabilities: Intelligent agents can timely perceive and respond to changes in their surrounding environment. Activeness: Intelligent agents can actively exhibit goal-driven behaviors and can choose appropriate times to take appropriate actions on their own. Communication capabilities: Intelligent agents can exchange information and interact with other entities in a certain communication manner. Persistence: Intelligent agents are in a continuous or ongoing running process, and their states should remain consistent during the running process. Reasoning and planning capabilities: Intelligent agents have the ability to perform relevant reasoning and intelligent calculations based on learned knowledge and experience.

[0085] Blockchain technology is a new distributed infrastructure and computing paradigm that uses a block-chain data structure to verify and store data, uses a distributed node consensus algorithm to generate and update data, uses cryptography to ensure the security of data transmission and access, and uses smart contracts composed of automated script codes to program and operate data. It has the following characteristics: Decentralization: Blockchain technology does not rely on an additional third-party management agency or hardware facilities, and there is no central control. Except for the self-contained blockchain itself, through distributed accounting and storage, each node realizes self-verification, transmission, and management of information. Decentralization is the most prominent and essential feature of the blockchain; Openness: The basis of blockchain technology is open source. Except for the private information of the trading parties being encrypted, the data of the blockchain is open to everyone. Anyone can query the blockchain data and develop related applications through the public interface. Therefore, the information of the entire system is highly transparent; Independence: Based on consensus norms and protocols (such as various mathematical algorithms like the hash algorithm), the entire blockchain system does not rely on other third parties. All nodes can automatically and securely verify and exchange data within the system without any human intervention; Security: As long as more than 51% of all data nodes cannot be controlled, it is impossible to wantonly manipulate and modify network data, which makes the blockchain itself relatively secure and avoids subjective and artificial data changes; Anonymity: Unless required by legal norms, technically speaking, the identity information of each block node does not need to be made public or verified, and information transmission can be carried out anonymously.

[0086] To achieve the above object, the present invention also provides an intelligent agent fusion communication method based on a neural network model, as Figure 4 shown, the method includes the following steps:

[0087] S1. Obtain the audio - video data corresponding to the intelligent terminal respectively, and forward - process the audio - video data in real - time in combination with the intelligent agent;

[0088] S2. According to the data after the forward - processing by the intelligent agent, broadcast data transmission at the network transport layer through the converged communication switching center cluster and adopt the TCP, UDP unicast or multicast method.

[0089] The step of obtaining the audio - video data corresponding to the intelligent terminal respectively, and forward - processing the audio - video data in real - time in combination with the intelligent agent further includes:

[0090] S11. Obtain the first synchronization control data corresponding to the group call communication of the cluster; wherein, the first synchronization control data is the consistency synchronization control data between distributed intelligent agents;

[0091] S12. According to the first synchronization control data, establish a multi - intelligent - agent network containment consistency model of the distributed intelligent agent.

[0092] The step of establishing a multi - intelligent - agent network containment consistency model of the distributed intelligent agent according to the first synchronization control data further includes:

[0093] S121. Generate an intelligent agent and obtain the connection address corresponding to the MECS, establish an intelligent - agent TCP connection corresponding to the MECS, and create a fully - connected multi - intelligent - agent network at the same time;

[0094] S122. According to the shortest - path algorithm, calculate and generate the shortest delay data to other intelligent agents in the fully - connected network;

[0095] S123. Send the self - group - call status through the intelligent agent, receive the group - call status sent by all neighbors, and update the group - call status in real - time according to its own dynamic equation;

[0096] S124. Check whether the self - group - call status of the intelligent agent reaches stability, and generate action control instruction data corresponding to stability or instability; wherein, the action control includes: establishment control, acquisition control, transmission control, playback control and termination control.

[0097] That is to say, in the solution of the present invention, the specific steps of the intelligent - agent - based converged communication method based on the neural network model are as follows:

[0098] S00. The intelligent terminal connects to the intelligent agent. The intelligent agent, as a participating node in the blockchain, uses the blockchain network for identity authentication. If the authentication is passed, the connection is maintained to continue the session; otherwise, the connection is disconnected to end the session. S01. After the intelligent terminal passes the authentication, when a call is needed, it sends the audio-visual timing data to the intelligent agent. The intelligent agent first puts the received audio-visual timing data into the receiving queue and inputs it into the link neural network model for link quality evaluation, forward error correction, and end-to-end rate control respectively. Then it outputs to the next-level priority and retransmission control processing, which combines reliable retransmission and service priority and puts them into the corresponding sending queue cache, and then forwards and processes the audio-visual data in real time and sends it to the converged communication switching center cluster.

[0099] As Figure 2-b shown, the link quality evaluation is based on the GNN graph neural network, the forward error correction is completed based on the LSTM long short-term memory network, and the end-to-end rate control is implemented based on the DNN deep neural network. When the audio-visual timing data of multiple intelligent agents is input into the GNN, the GNN comprehensively evaluates to form the link quality of each intelligent terminal. The LSTM synchronously converts and forward-predicts and corrects the audio-visual timing data. The evaluation result of the GNN and the audio-visual data after LSTM correction are input into the DNN, and the DNN performs adaptive rate adjustment for multiple terminals, overall coordinating to ensure the audio-visual transmission quality, and outputs the audio-visual data to the next level for retransmission and priority control.

[0100] S02. According to the data after the intelligent agent's forwarding and processing, it performs hybrid gating or forwarding and distribution through the converged communication switching center cluster, and uses TCP, UDP unicast or multicast methods to broadcast data transmission at the network transport layer.

[0101] The step of respectively obtaining the audio-visual data corresponding to the intelligent terminal and combining the intelligent agent to forward and process the audio-visual data in real time further includes:

[0102] S011. Obtain the first synchronization control data corresponding to the cluster group call communication; wherein, the first synchronization control data is the consistency synchronization control data between distributed intelligent agents.

[0103] S012. According to the first synchronization control data, establish a multi-intelligent agent network containment consistency model for distributed intelligent agents.

[0104] The step of establishing a multi-intelligent agent network containment consistency model for distributed intelligent agents according to the first synchronization control data further includes:

[0105] S0121. Generate intelligent agents and obtain the connection addresses corresponding to the MECS, establish the intelligent agent TCP connection corresponding to the MECS, and create a fully connected multi-intelligent agent network at the same time.

[0106] S0122. Calculate and generate the shortest delay data to other intelligent agents in the fully connected network according to the shortest path algorithm;

[0107] S0123. Send the self-group call status through the intelligent agent, receive the group call status sent by all neighbors, and update the group call status in real time according to the self-dynamics equation;

[0108] S0124. Check whether the self-group call status of the intelligent agent reaches stability, and generate action control instruction data corresponding to stability or instability; wherein, the action control includes: establishment control, acquisition control, transmission control, playback control, and termination control.

[0109] S03. Integrate the audio and video timing data sent by the fusion communication switching center cluster to the intelligent terminal. Similar to the S1 processing, the intelligent agent first receives and puts it into the receiving queue, inputs it into the link neural network model for link quality evaluation, forward error correction, and end-to-end rate control respectively, outputs to the next-level priority and retransmission control processing, combines reliable retransmission and service priority and puts it into the corresponding sending queue cache, and then forwards and processes the audio and video data in real time and sends it to the intelligent terminal.

[0110] In the embodiment of the method solution of the present invention, the functional modules involved in the intelligent agent fusion communication method based on the neural network model have been described in detail above and will not be elaborated here.

[0111] To achieve the above object, the present invention also provides an intelligent agent fusion communication platform based on a neural network model, as Figure 5 shown, including a processor, a memory, and an intelligent agent fusion communication platform control program based on a neural network model; wherein, the processor executes the intelligent agent fusion communication platform control program based on a neural network model, the intelligent agent fusion communication platform control program based on a neural network model is stored in the memory, and the intelligent agent fusion communication platform control program based on a neural network model implements the steps of the intelligent agent fusion communication method based on a neural network model, for example:

[0112] S1. Obtain the audio and video data corresponding to the intelligent terminal respectively, and forward and process the audio and video data in real time in combination with the intelligent agent;

[0113] S2. According to the data after the intelligent agent forwards and processes, broadcast data transmission in the network transport layer through the fusion communication switching center cluster and adopt TCP, UDP unicast or multicast methods.

[0114] The specific details of the steps have been described in detail above and will not be elaborated here.

[0115] In an embodiment of the present invention, the intelligent agent fusion communication platform based on a neural network model is built-in with a processor, which can be composed of integrated circuits. For example, it can be composed of a single packaged integrated circuit, or can be composed of multiple integrated circuits with the same or different functions, including a combination of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor uses various interfaces and circuits to connect to each component, and by running or executing programs or units stored in the memory, and calling data stored in the memory, it performs various functions of the intelligent agent fusion communication based on the neural network model and processes data;

[0116] The memory is used to store program codes and various data, installed in the intelligent agent fusion communication platform based on the neural network model, and realizes high-speed and automatic access to programs or data during operation.

[0117] The memory includes read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc memories, magnetic disk memories, tape memories, or any other computer-readable medium capable of carrying or storing data.

[0118] The intelligent agent fusion communication architecture of the present invention includes at least one intelligent agent unit for converting audio and video data, a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively; the intelligent agent fusion communication architecture uses a distributed intelligent agent cluster to achieve group call control, and learns link quality control through an intelligent agent neural network model; and a method and platform corresponding to the architecture, and introduces artificial intelligence and blockchain into the IMS fusion communication architecture to solve the problem of accessing multi-protocol fusion switching communication of different systems and types in a complex environment of a large number of terminals, form the ability of uninterrupted transmission of multiple links, and support the secure, reliable and high-quality communication transmission of various service application data such as voice, video and conference.

[0119] That is to say, the solution of the present invention adaptively learns link quality control through an intelligent agent neural network model to ensure high-quality audio and video communication; each intelligent terminal is interacted by an intelligent agent, and the fusion communication switching center only needs to communicate with the intelligent agent, increasing the reliability of large-scale broadcast communication; the intelligent agent and the fusion communication switching center uniformly use the SIP protocol to reduce the docking between multiple protocols of the fusion communication switching center; the intelligent agent is integrated into the advanced IMS architecture as a UA, and the required service quality can be obtained, better supporting services such as registration, security, billing, bearer control and roaming in the fusion communication. The intelligent agent is used as a blockchain node to perform authentication protection, authorization and encryption based on the blockchain to ensure the security of the system and prevent threats such as theft and tampering.

[0120] The above embodiments only represent several implementation manners of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the appended claims.

Claims

1. An intelligent agent fusion communication system based on a neural network model, characterized in that, the intelligent agent fusion communication system includes at least one intelligent agent unit for converting audio and video data, and a fusion communication switching center unit and at least one intelligent terminal unit that communicate with the intelligent agent unit respectively; the intelligent agent fusion communication system uses a distributed intelligent agent cluster to implement group call control and learns link quality control through an intelligent agent neural network model; the intelligent agent fusion communication system is further provided with a synchronization processing unit, and the synchronization processing unit is used to synchronize and coordinate the cluster and synchronously collect, transmit and play voice and video data streams respectively; the intelligent agent fusion communication system is further provided with a first creation unit, and the first creation module is used to establish a multi-intelligent agent network restraint consistency model of distributed intelligent agents; the first creation unit further includes: a generation and acquisition module, which is used to generate intelligent agents and acquire connection addresses corresponding to MECS; a first calculation module, which is used to calculate the delay data to the intelligent agent in the fully connected network; a self-check and generation module, which is used to check whether its own group call status reaches stability and generate action control instruction data corresponding to stability or instability, where the action control includes: establishment control, acquisition control, transmission control, playback control and termination control; when the intelligent terminal unit is connected to the intelligent agent unit, the intelligent agent unit uses the area chain network as a blockchain participating node for identity authentication. If the authentication is passed, the connection is maintained and the session continues. Otherwise, the connection is disconnected and the session ends; after the intelligent terminal authentication is passed, when a call is required, the audio and video timing data is sent to the intelligent agent, the intelligent agent unit is specifically used for: receiving the audio and video timing data from the intelligent terminal and putting it into the receiving queue, inputting the data in the receiving queue into the link neural network model to perform link quality evaluation, forward correction and end-to-end rate control respectively; according to the output of the link neural network model, combining reliable retransmission and service priority, putting the processed audio and video data into the corresponding sending queue cache, and then performing real-time forwarding processing on the audio and video data and sending it to the fusion communication switching center unit; wherein, the link quality evaluation is based on the GNN graph neural network, the forward correction is completed based on the LSTM long short-term memory network, and the end-to-end rate control is implemented based on the DNN deep neural network; when multiple intelligent agent audio and video timing data are input into the GNN graph neural network, the GNN graph neural network comprehensively evaluates to form the link quality of each intelligent terminal, the LSTM synchronously converts and forward-predicts and corrects the audio and video timing data, the GNN evaluation result and the audio and video data after LSTM correction are input into the DNN deep neural network, and the DNN deep neural network performs adaptive rate adjustment for multiple terminals, overall coordinating to ensure the audio and video transmission quality, and outputting the audio and video data to the next level for retransmission and priority control.

2. The intelligent agent fusion communication system based on a neural network model according to claim 1, characterized in that, The fusion communication switching center unit broadcasts data transmission at the network transport layer through TCP, UDP unicast or multicast methods.

3. An intelligent agent fusion communication method for an intelligent agent fusion communication system based on a neural network model as described in any one of claims 1-2, characterized in that the method uses a distributed intelligent agent cluster to implement group call control and learns link quality control through an intelligent agent neural network model, including the following steps: respectively obtain audio and video data corresponding to intelligent terminals, and combine with intelligent agents to forward process the audio and video data in real time; According to the data after the intelligent agent forwarding process, broadcast data transmission at the network transport layer through the fusion communication switching center cluster and using TCP, UDP unicast or multicast methods; The step of respectively obtaining audio and video data corresponding to intelligent terminals and combining with intelligent agents to forward process the audio and video data in real time further includes: obtain first synchronization control data corresponding to cluster group call communication; wherein, the first synchronization control data is consistency synchronization control data between distributed intelligent agents; establish a multi-intelligent agent network containment consistency model of distributed intelligent agents according to the first synchronization control data; The step of establishing a multi-intelligent agent network containment consistency model of distributed intelligent agents according to the first synchronization control data further includes: generate intelligent agents and obtain connection addresses corresponding to MECS, establish intelligent agent TCP connections corresponding to MECS, and create a fully connected multi-intelligent agent network at the same time; calculate and generate the shortest delay data to other intelligent agents in the fully connected network according to the shortest path algorithm; send its own group call status through the intelligent agent, receive the group call status sent by all neighbors, and update the group call status in real time according to its own dynamic equation; check whether the group call status of the intelligent agent itself reaches stability, and generate action control instruction data corresponding to stability or instability; wherein, the action control includes: establishment control, acquisition control, transmission control, playback control and termination control.

4. An intelligent agent fusion communication platform based on a neural network model, characterized in that it includes a processor, a memory, and an intelligent agent fusion communication platform control program based on a neural network model; wherein, when the processor executes the intelligent agent fusion communication platform control program based on a neural network model, the intelligent agent fusion communication platform control program based on a neural network model is stored in the memory, and the intelligent agent fusion communication platform control program based on a neural network model realizes the intelligent agent fusion communication method as described in claim 3.

Images