Intelligent early warning multi-party teleconference system based on large model

Abstract

The invention relates to an intelligent and multi-party real-time communication early-warning multi-party teleconference system, which is used for quickly scheduling related personnel to carry out remote cooperation and decision-making in an emergency. The system is based on cloud computing, VoIP, artificial intelligence and edge computing technologies, and integrates functions of IoT early warning equipment, real-time voice processing, intelligent scheduling, encrypted communication and the like so as to ensure efficient, stable and safe conference management. The system supports automatic early warning triggering, can start a teleconference through sensor data, AI prediction analysis or manual instructions, and intelligently identifies key personnel who should participate in the conference. The system adopts automatic voice notification and intelligent scheduling, invites participants in a multi-channel manner through a telephone, a short message, an APP and the like, supports intelligent auxiliary functions such as real-time voice transcription, automatic generation of conference summary, emotion analysis and the like, and improves the decision making efficiency. In order to ensure the stability and safety of the system, a PSTN, VoIP and communication fusion architecture is adopted, and the communication stability under extreme conditions is ensured. The system also has the functions of automatic task allocation and event backtracking analysis, can be combined with AI to carry out subsequent emergency optimization, and improves the overall response capability. The method can be widely applied to the fields of hardware and software emergency management, enterprise crisis processing, public security and the like, and provides technical support for quick response and efficient decision making of emergencies.

Description

Technical Field

[0001] This invention relates to communication technology, speech synthesis and transcription, and large model processing technology. Background Technology

[0002] With the widespread adoption of the internet and the Internet of Things (IoT), more and more enterprises have higher requirements for how to respond quickly to emergencies. Early warning multi-party teleconference systems are primarily used to quickly and efficiently notify relevant personnel and coordinate response actions in emergency situations. The background mainly includes the following aspects: emergency management needs, emergency early warning functions for equipment failures, and rapid generation of response measures. The core objective of this system is to improve response speed, optimize decision-making processes, and ensure that key personnel can collaborate efficiently in the first instance. Summary of the Invention

[0003] The invention of the early warning multi-party teleconference system focuses on innovation in intelligence, real-time multi-party communication, high reliability, and automated response, aiming to provide an efficient, stable, and secure solution for emergency meeting scheduling and management. The following is a detailed description of the core invention:

[0004] 1. Early warning triggering mechanism

[0005] (1) Multi-channel early warning triggering

[0006] Automatic triggering: The system can be integrated with IoT devices, monitoring systems, etc., to automatically detect events and trigger teleconferences.

[0007] Manual triggering: Authorized users can manually start the meeting via Web, mobile app, voice command, SMS command, etc.

[0008] Intelligent triggering: Based on historical data and AI predictive analysis, potential risks can be identified in advance, and early warning meetings can be initiated in advance.

[0009] (2) Automatic determination of warning level

[0010] By combining sensor data, AI analysis, and historical events, the event level can be automatically assessed.

[0011] Based on the warning level, the system automatically determines the scope and priority of the notification and decides whether to escalate it to a network-wide emergency meeting.

[0012] 2. Intelligent Meeting Scheduling and Notification

[0013] (1) AI recognition and automatic invitation

[0014] Based on the event type, AI automatically matches key personnel such as those who need to attend the meeting, emergency response teams, and corporate management, and sends meeting invitations. It supports multiple notification methods, including phone calls, SMS, and App Email.

[0015] (2) Automatic voice call

[0016] Using text-to-speech technology, the system automatically makes phone calls to notify attendees and broadcasts the meeting content.

[0017] It can be combined with interactive voice responses, allowing users to choose whether to join the meeting via voice.

[0018] (3) Meeting Priority Management

[0019] Core decision-making group priority: Users with high privileges can join first and manage meeting permissions.

[0020] Multi-level scheduling: Automatically allocate different meeting rooms based on the responding role.

[0021] 3. Intelligent Meeting Management

[0022] (1) AI speech recognition and real-time transcription

[0023] It employs automatic speech recognition to transcribe meeting content in real time and generate text records.

[0024] It can provide real-time subtitles and multilingual translation.

[0025] (2) Automatic generation of meeting minutes

[0026] AI automatically generates a summary of key points based on the meeting content and pushes it to relevant personnel.

[0027] By combining natural language processing to provide intelligent summaries, the efficiency of subsequent actions can be improved.

[0028] (3) Intelligent monitoring of the meeting process

[0029] Voice emotion analysis: Analyze the emotions of participants, identify key issues, and assist in command and decision-making.

[0030] AI Keyword Detection: Real-time monitoring of sensitive words and automatic marking of key information.

[0031] 4. Early warning linkage and follow-up handling

[0032] (1) Automatic task allocation

[0033] After the meeting makes a decision, AI can automatically generate a task list and assign it to relevant departments.

[0034] Use a collaborative platform to track task progress. Attached Figure Description

[0035] Appendix Figure 1 As shown, the data collected by the IoT and monitoring systems is input into the early warning multi-party conference system, and users can also intervene and operate the early warning multi-party conference.

[0036] The early warning multi-party conferencing system processes the input data and outputs multi-party call data, including voice and text information.

[0037] To facilitate subsequent processing, the multi-party call data can be divided into two stages: speech-to-text and text-to-speech.

[0038] After translation, the data enters the natural language processing stage for semantic understanding and information organization, ultimately forming meeting minutes and tasks, which are then carried out by the task executors.

Claims

1. A pre-warning multi-party teleconference system, comprising: A speech recognition module is used to convert user speech into text; The speech synthesis module is used to convert text into speech for broadcasting. The data storage module is used to store call logs, user information, and business data. The large model summary is used to summarize the corresponding work content and processing solutions for the user-converted text information; The monitoring and statistics module is used to monitor call status in real time and generate service quality reports.

2. The early warning multi-party teleconference system according to claim 1, wherein, The speech recognition module is optimized based on deep learning algorithms, enabling multilingual recognition and speech enhancement.

3. The early warning multi-party teleconference system according to claim 1, wherein, Large-scale conference call notes can provide key information from the conference call based on semantic understanding, reducing the cost of manual data processing.

4. The early warning multi-party teleconference system according to claim 1, wherein, The monitoring and statistics module can generate data visualization reports in real time and supports service quality analysis and optimization.

5. The early warning multi-party teleconference system according to claim 1, characterized in that, The modular design is divided into a business layer, a logic processing layer, and an output layer. It is used to receive, process, and transform call data to further analyze call results, optimize large model summary parameters, and improve the accuracy and efficiency of analysis.

6. The device according to any one of claims 1 to 5 performs early warning analysis of multi-party teleconference data, comprising the following steps: 1) Automatically collect call data including call time, response time, call duration, conference text, etc. 2) Optimization of meeting texts for multi-party participants and synthesis and integration of multi-party recordings. 3) Use a large model to analyze and summarize the above content, and provide corresponding tasks and solutions.

Images