CN-121223797-B - Manual-intelligence-based control method and system for conference investigation robot

CN121223797BCN 121223797 BCN121223797 BCN 121223797BCN-121223797-B

Abstract

The application relates to the technical field of robot control, in particular to a method and a system for controlling a conference investigation robot based on artificial intelligence. The method comprises the steps of obtaining audio data and monitoring data of a conference environment, analyzing the monitoring data and the audio data, determining an speaking state and speaking content, determining a speaking distribution thermodynamic diagram according to the speaking state and the speaking content, determining a robot active interaction strategy according to the speaking content and the speaking distribution thermodynamic diagram, and sending the robot active interaction strategy to the robot so that the robot can execute the active interaction strategy to guide conference discussion. Accuracy and instantaneity of intention understanding are improved, and structured input is provided for thermodynamic diagram generation and interaction strategies. The real-time visualization of conference discussion distribution is realized, the defect that thermodynamic diagrams cannot be automatically generated is overcome, and the monitoring capability of participation balance is improved. The self-adaptive interaction mechanism is realized, the delay depending on manual intervention is eliminated, and the efficient and accurate progress of the conference is ensured.

Inventors

Ju yue
DU BOREN
WANG YUNQIU

Assignees

上海万怡医学科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20251031

Claims (7)

1. The robot control method for meeting investigation based on artificial intelligence is characterized by comprising the following steps: Acquiring audio data and monitoring data of a conference environment; analyzing the monitoring data and the audio data to determine speaking states and speaking contents; Determining a speech distribution thermodynamic diagram according to the speech state and the speech content; determining an active interaction strategy of the robot according to the speaking content and the speaking distribution thermodynamic diagram, and sending the active interaction strategy to the robot so that the robot executes the active interaction strategy to guide conference discussion; The determining a robot active interaction strategy according to the speaking content and the speaking distribution thermodynamic diagram further comprises: Analyzing the speaking content and determining speaking views; analyzing the speaking state and determining the matching degree of the speaking state and the speaking viewpoint; If the matching degree is lower than a preset matching value, generating a contradiction mark, and determining partition discussion liveness according to the speaking distribution thermodynamic diagram; generating an inquiry strategy of different partitions according to the speaking viewpoint and the partition discussion liveness based on the contradiction mark; the analyzing the speaking state and determining the matching degree of the speaking state and the speaking viewpoint comprise the following steps: acquiring a preset micro-expression-semantic mapping table, wherein the corresponding relation between the vocabulary and the forward micro-expression is stored; Analyzing the micro-expression type in the speaking state, and determining the actual micro-expression polarity; analyzing keywords in the speaking views based on the micro-expression-semantic mapping table, and determining expected micro-expression polarities; if the actual micro-expression polarity is inconsistent with the expected micro-expression polarity, determining that the matching degree is lower than a preset matching value; Based on the contradiction mark, generating an additional query strategy of different partitions according to the speaking viewpoint and the partition discussion liveness, including: when a contradictory mark is detected, analyzing the speaking views and determining problem contradiction points; according to the partition discussion liveness, locating a low-liveness region to which the contradiction mark belongs; the problem contradiction points are directionally pushed to the high active area, and a high active area inquiring instruction is generated; receiving feedback audio data of a high active region, analyzing the feedback audio data, and generating a feedback abstract; and pushing the feedback abstract to the original low active region in a directional way.
2. The method of claim 1, wherein analyzing the monitoring data and the audio data to determine a floor status comprises: analyzing the monitoring data to determine the movement characteristics of facial muscles of the speaker and the position coordinates of the speaker; determining a micro expression type according to the motion characteristics; analyzing the audio data and determining speaking voiceprint characteristics and speaking contents; matching a conference record database according to the speaking voiceprint characteristics, and determining the identity of a speaker; And determining the micro-expression type, the position coordinates and the speaker identity as the speaking state of the speaker.
3. The method of claim 2, wherein determining a floor distribution thermodynamic diagram based on the floor status and floor content comprises: obtaining a position coordinate set of all speakers based on the position coordinates; Mapping the position coordinate set to a grid partition according to a preset conference room plane grid; according to the speaking content, counting the speaking total duration and the speaking times in unit time in each grid partition; and weighting the speaking times and the speaking total duration to generate a subarea liveness thermodynamic diagram.
4. The method of claim 1, wherein determining a robot-active interaction strategy based on the utterance content and the utterance distribution thermodynamic diagram comprises: analyzing the speaking content and determining topic keywords; Determining the discussion duration and the number of speakers of the topic keywords based on the speaking distribution thermodynamic diagram; if the number of speakers in unit time is lower than a preset value and the discussion duration exceeds a threshold value, marking the discussion duration as an open topic; When the duration of the open topic exceeds the limit, analyzing the speaking content and determining a fragmentation viewpoint; Clustering the fragmentation views generates a number of structured propulsion options.
5. The method of claim 4, wherein the clustering the fragmentation views generates a number of structured pushing options, comprising: Extracting entity words and action words in the fragmentation viewpoint; Analyzing the entity words, and determining the association degree between at least two entity words; Analyzing the action words and the entity words, and determining action relations between any action word and any entity word; Constructing a semantic graph according to the association degree, wherein nodes represent entity words and edges represent action relations; Combining entities with the same father node according to the semantic graph to generate candidate viewpoint clusters; Counting the occurrence frequency of each candidate viewpoint cluster, and obtaining a plurality of structured propulsion options according to the counting result.
6. The method of claim 5, wherein after obtaining a plurality of structured pushing options according to the statistics, further comprising: Analyzing the semantic conflict degree of any two structured propulsion options; If the option pairs with semantic conflict exist, a conflict warning mark is generated; positioning the partition where the supporters of the conflict warning marks are located according to the speaking distribution thermodynamic diagram; And generating dialectical guiding strategies aiming at the partition where the supporters are located according to the fragmentation point of view.
7. A robot control system for conference investigation based on artificial intelligence, characterized by being applied to the method according to any of claims 1-6, comprising: The data acquisition module is used for acquiring audio data and monitoring data of the conference environment; the data analysis module is used for analyzing the monitoring data and the audio data and determining speaking states and speaking contents; a thermodynamic diagram generation module for determining a thermodynamic diagram of the speaking distribution according to the speaking state and the speaking content; And the strategy generation module is used for determining an active interaction strategy of the robot according to the speaking content and the speaking distribution thermodynamic diagram and sending the active interaction strategy to the robot so that the robot executes the active interaction strategy to guide conference discussion.

Description

Manual-intelligence-based control method and system for conference investigation robot Technical Field The application relates to the technical field of robot control, in particular to a method and a system for controlling a conference investigation robot based on artificial intelligence. Background With the rapid development of artificial intelligence technology, in a conference investigation scene, it is important to efficiently and accurately collect and analyze the views, moods and participation of participants. Meeting research plays a key role in modern organizations, such as in decision-making, problem discussion, or team collaboration scenarios, which can help to efficiently collect views, ideas, and feedback of participants. However, the existing conference investigation method mainly relies on static equipment to collect data, and performs subsequent arrangement and analysis in a manual mode, so that conference efficiency is low, decision quality is damaged, and the requirements of a dynamic conference environment cannot be met. Disclosure of Invention The application provides a method and a system for controlling a conference investigation robot based on artificial intelligence, which are used for solving the problems. In a first aspect, the present application provides a robot control method for conference investigation based on artificial intelligence, the method comprising: Acquiring audio data and monitoring data of a conference environment; analyzing the monitoring data and the audio data to determine speaking states and speaking contents; Determining a speech distribution thermodynamic diagram according to the speech state and the speech content; And determining an active interaction strategy of the robot according to the speaking content and the speaking distribution thermodynamic diagram, and sending the active interaction strategy to the robot so that the robot executes the active interaction strategy to guide conference discussion. According to the scheme, the audio data and the monitoring data of the conference environment are obtained, the limitation that only single audio data is relied on is eliminated, and then conference dynamics are captured from multiple angles. And analyzing the monitoring data and the audio data to determine the speaking state and the speaking content, realizing multi-dimensional data fusion analysis, improving the accuracy and the instantaneity of intention understanding, and providing structural input for thermodynamic diagram generation and interaction strategies. According to the speaking state and the speaking content, determining the speaking distribution thermodynamic diagram, realizing the real-time visualization of conference discussion distribution, eliminating the defect that the thermodynamic diagram cannot be automatically generated, and improving the monitoring capability of participation balance. And determining an active interaction strategy of the robot according to the speaking content and the speaking distribution thermodynamic diagram, and sending the active interaction strategy to the robot so that the robot executes the active interaction strategy to guide conference discussion, thereby realizing a self-adaptive interaction mechanism, eliminating the delay of relying on manual intervention and ensuring the efficient and accurate progress of the conference. Optionally, the analyzing the monitoring data and the audio data to determine the speaking status includes: analyzing the monitoring data to determine the movement characteristics of facial muscles of the speaker and the position coordinates of the speaker; determining a micro expression type according to the motion characteristics; analyzing the audio data and determining speaking voiceprint characteristics and speaking contents; matching a conference record database according to the speaking voiceprint characteristics, and determining the identity of a speaker; And determining the micro-expression type, the position coordinates and the speaker identity as the speaking state of the speaker. According to the scheme, the monitoring data are analyzed, the movement characteristics of facial muscles of the speaker and the position coordinates of the speaker are determined, the non-language clues can be quantized, the problem of neglecting visual data is solved, the speaker position is positioned in real time, and the recognition of a low active region is supported. And determining the micro expression type according to the motion characteristics, and enhancing the accuracy of intention recognition. And analyzing the audio data, determining the characteristics of the speech voiceprint and the speech content, and providing input data for determining the identity of the speaker. According to the speaking voiceprint characteristics, the conference record database is matched, the identity of the speaker is determined, the problem of speaker identity identification is solved, and data is provided