Search

CN-122024726-A - Intelligent voice receiving and processing robot based on AI large model technology

CN122024726ACN 122024726 ACN122024726 ACN 122024726ACN-122024726-A

Abstract

The application discloses an intelligent voice receiving and processing alarm robot based on an AI large model technology, which comprises a robot main body, wherein the robot main body comprises a voice interaction module, a voice analysis module and a voice analysis module, wherein the voice interaction module is used for carrying out real-time full duplex voice communication with an alarm person, receiving alarm voice flow and outputting response voice, thus, the intelligent technology is used for replacing manual repetitive work, the alarm telephone queuing waiting time can be effectively divided, the accuracy of subsequent alarm treatment is improved, the deep semantic recognition and context correlation interaction technology is adopted, the fuzzy, spoken and emotional expressions of the alarm person can be accurately read, the multi-round intelligent dialogue is supported to confirm key information, the alarm receiving time is greatly shortened, the information acquisition efficiency and the integrity of complex alarm are improved, a standardized alarm grade judgment model is constructed, the intelligent analysis is carried out by combining multi-source data, the conditions of effective repeated scheduling, resource conflict and the like are ensured, and the rationality and the accuracy of the input of alarm resources are ensured.

Inventors

  • MA XINCHENG

Assignees

  • 新疆智安数联信息技术有限公司

Dates

Publication Date
20260512
Application Date
20260317

Claims (6)

  1. 1. Intelligent voice receiving and processing robot based on AI large model technology, which is characterized by comprising a robot main body, wherein the robot main body comprises the following modules: the voice interaction module is used for carrying out real-time full duplex voice communication with the alarm person, receiving alarm voice stream and outputting response voice; the AI processing module comprises a large language model unit and a voice model unit, wherein the large language model unit carries out deep semantic understanding, context correlation analysis, intention recognition and information structured extraction on input text information and generates decision instructions and personification response texts; The voice model unit carries out real-time voice recognition on the alarm voice stream received by the voice interaction module, sends a recognition text into the large language model unit, converts a response text generated by the large language model unit into anthropomorphic voice and feeds the anthropomorphic voice back to the voice interaction module; the alarm analysis and scheduling module is used for automatically judging alarm grades and matching treatment plans according to the structured alarm information and the decision instruction output by the AI processing module, and generating scheduling suggestions or automatically triggering scheduling instructions; And the system docking module is used for carrying out bidirectional data interaction on the structured alarm information and the scheduling instruction with the existing alarm receiving and processing command system and the police service data resource library.
  2. 2. The intelligent voice receiving and alarming robot based on the AI large model technology as set forth in claim 1, wherein the large language model unit dynamically extracts key alarm elements including event types, time, place, person, vehicle, related objects and site conditions based on the context of an alarm dialogue when executing an alarm interaction task, and fills the key alarm elements into a predefined structured form; According to the extracted elements and dialogue history, the emergency degree, the credibility and the information integrity of the police condition are evaluated in real time; when the information is missing or contradictory, a targeted overtaking talk is automatically generated, and an alarm person is guided to supplement key information; in the interaction process, the emotion of the alarm person is identified and pacified, and a response text which accords with police service standards and has concentricity is generated.
  3. 3. The intelligent voice receiving and alarming robot based on the AI large model technology as set forth in claim 1, wherein the alarm analysis and dispatching module alarm level automatic judgment logic integrates the degree of urgency and credibility of the large language model unit evaluation, and performs multidimensional weighted calculation in combination with historical alarm data and case related personnel background information acquired from the system docking and data interface module.
  4. 4. The intelligent voice receiving and alarming robot based on the AI large model technology as set forth in claim 1, further comprising a multi-mode information fusion module for accessing and processing pictures, videos or real-time streaming media data uploaded by an alarming person through a terminal; The AI processing module also comprises a multi-mode large model unit which is used for carrying out content understanding, object identification and scene analysis on the pictures or videos, and supplementing analysis results to the large language model unit in a text description form so as to assist alert judgment and information structuring.
  5. 5. The intelligent voice receiving and alarming robot based on the AI large model technology as set forth in claim 1, wherein the voice model unit integrates a voiceprint recognition and emotion analysis subunit, and is configured to synchronously perform voiceprint comparison of an alarming person in the voice recognition process, analyze an emotion state in the voice of the alarming person in real time, and send emotion labels and intensity information to the large language model unit to adjust a response strategy thereof.
  6. 6. The intelligent voice receiving and processing robot based on the AI large model technology as claimed in claim 1, further comprising an autonomous learning and optimizing module for continuously collecting anonymized receiving conversation data, final processing results and correction feedback of artificial agents, wherein the autonomous learning and optimizing module optimizes parameters and strategies of a large language model unit in the AI processing module so as to improve accuracy of the large language model unit in alert processing.

Description

Intelligent voice receiving and processing robot based on AI large model technology Technical Field The application relates to the technical field of robots, in particular to an intelligent voice receiving and processing robot based on an AI large model technology. Background The current police service receiving and processing police takes a manual seat leading and traditional auxiliary system as a core mode, although simple voice interaction equipment is introduced, the difference between the technical level and actual combat demands is remarkable, firstly, the bottleneck of the manual police service receiving efficiency is remarkable, the resource allocation is unbalanced, the seat operates at full load in a peak period, the response of the police condition is delayed, the time and space fluctuation of the police condition is large, the seat is difficult to dynamically adapt, the resources are idle in idle and the busy hours are in operation, repeated police conditions are processed manually for a long time, the information omission and the recording deviation are caused. And secondly, the semantic understanding and information extraction capability is weak, complex warning conditions are difficult to adapt, the existing voice equipment can only identify simple structural information such as places, event types and the like, can not read the fuzzification expression of an alarm person and the spoken language expression, and can not be associated with context interaction, and core elements of the complex warning conditions are required to be verified manually one by one, so that the time consumption of receiving the warning is greatly prolonged. In addition, the alarm condition judgment and scheduling have strong subjectivity, standardization degree, alarm condition grade and treatment plan depend on manual experience judgment, misjudgment and mismatching are easy to occur, the scheduling lacks multi-source data support, can not link information such as historical alarm conditions, case personnel background and the like, decision scientificity is not enough, and the problems of repeated scheduling and resource conflict are also caused. Content of the application The present application aims to solve the technical problems in the related art at least to some extent. The intelligent voice receiving and processing robot based on the AI large model technology comprises a robot main body, wherein the robot main body comprises a voice interaction module, an AI processing module, a system interfacing module, a warning condition analysis and scheduling module and a system interfacing module, wherein the voice interaction module is used for carrying out real-time full duplex voice communication with a warning person, receiving a warning voice stream and outputting a response voice, the AI processing module comprises a large language model unit and a voice model unit, the large language model unit is used for carrying out deep semantic understanding, context correlation analysis, intention recognition and information structured extraction on input text information and generating a decision instruction and an anthropomorphic response text, the voice model unit is used for carrying out real-time voice recognition on the warning voice stream received by the voice interaction module and sending the recognition text to the large language model unit, and converting the response text generated by the large language model unit into anthropomorphic voice to be fed back to the voice interaction module, the warning condition analysis and scheduling module is used for carrying out warning level automatic judgment and handling pre-matching according to structured warning condition information and decision instruction output by the AI processing module, and generating a scheduling suggestion or automatic triggering scheduling instruction, and the system interfacing module is used for carrying out bidirectional interaction between the structural warning information and the warning information receiving and the warning information. The method comprises the steps of dynamically extracting key alarm condition factors including event types, time, places, characters, vehicles, related objects and site conditions and filling the key alarm condition factors into a predefined structural form when an alarm condition interaction task is executed by the large language model unit, evaluating the emergency degree, the credibility and the information integrity of the alarm condition in real time according to the extracted factors and the conversation history, automatically generating a targeted overtime dialogue operation when the information is missing or contradictory, guiding an alarm person to supplement key information, and identifying and pacifying the emotion of the alarm person and generating a response text which accords with the alarm condition specification and has the same sense in the interaction process. Specifically, the alarm co