Intelligent interaction robot control system

Abstract

The invention relates to the technical field of data analysis, in particular to an intelligent interaction robot control system which comprises a voice receiving module, a text conversion processing module, an audio analysis processing module and an emotion judgment adjusting module. Wherein the text conversion processing module is used for converting purified audio signals into text content and determining text word length, adjective ratio and negative word ratio, the audio analysis processing module is used for determining the speech speed, fundamental tone characteristic trend and frequency fluctuation tendency of a single purified audio signal, and the emotion judgment adjusting module is used for determining character emotion tendency and audio emotion tendency. Determining the current emotional state of the user in combination with the character emotional tendency and the audio emotional tendency so as to determine whether to switch the voice output style of the interaction robot; according to the intelligent interaction robot control system provided by the invention, efficient voice processing and emotion analysis functions are integrated, so that the intelligent level of man-machine interaction is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to an intelligent interactive robot control system. Background Art

[0002] With the rapid development of artificial intelligence, natural language processing, machine learning, speech recognition and synthesis, interactive robots have been able to have more natural and fluent conversations with humans. These technological advances have made it possible for robots to understand and respond to human emotions, thus enabling robots to mobilize users' emotions during the interaction process.

[0003] Chinese patent announcement number CN116665273B discloses a robot human-machine interaction method based on expression recognition and emotion quantification analysis calculation, including: step P100: identifying the user's expression; step P200: quantifying and calculating the emotion according to the expression recognition result of step P100, and obtaining the current emotion quantification calculation analysis result; step P300: transmitting the current emotion quantification calculation analysis result to the interactive robot, and the interactive robot determines the corresponding interaction content according to the current emotion quantification calculation analysis result, and outputs the interaction content in a human-machine interaction manner, so as to adjust the user's emotion. The robot human-machine interaction method of the present invention can identify multiple emotions that the user may have at the same time, so as to more accurately identify the user's current emotional state. At the same time, the interactive robot makes a targeted response according to the user's current emotional state, so as to better adjust the user's emotion.

[0004] It can be seen that the invention has the following problems: The existing technology fails to determine the user's current emotional state based on voice input and automatically adjust the robot's output style according to the current emotional state. Summary of the invention

[0005] To this end, the present invention provides an intelligent interactive robot control system to overcome the problem in the prior art that the user's emotional state cannot be determined based on voice input and the robot output style cannot be switched.

[0006] To achieve the above object, the present invention provides an intelligent interactive robot control system, comprising: A voice receiving module, used for receiving an original audio signal sent by a user and performing noise reduction processing on the original audio signal to form a purified audio signal; A text conversion processing module, which is connected to the voice receiving module, and is used to convert the purified audio signal into text content to determine the length of the text word count, determine the number of adjectives, the emotional vocabulary category of each adjective, the number of nouns and the emotional vocabulary category of each noun in a single text content according to the text content, determine the corresponding adjective ratio according to the number of adjectives and the number of nouns, and determine the negative word ratio according to the determination results of the emotional vocabulary category of each adjective and the emotional vocabulary category of each noun; an audio analysis processing module, which is connected to the speech receiving module and the text conversion processing module respectively, and is used to determine the speech rate of a single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text word count, and to determine the time domain audio image according to the purified audio signal to determine the fundamental tone feature trend and the frequency fluctuation tendency; an emotion judgment adjustment module, which is connected to the text conversion processing module and the audio analysis processing module respectively, and is used to determine the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to a single purified audio signal, determine the audio emotion tendency according to the pitch feature trend and the frequency fluctuation tendency of the single purified audio signal, and determine the current emotional state of the user in combination with the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot; The emotional vocabulary categories include positive emotional vocabulary, negative emotional vocabulary and neutral emotional vocabulary.

[0007] Further, the text conversion processing module includes a positive sentiment word library and a negative sentiment word library for determining the sentiment word categories of adjectives and nouns, and determines the negative word ratio according to the determination results of the sentiment word category of each adjective and the sentiment word category of each noun; The negative word ratio is the ratio of the sum of the number of negative sentiment words of adjectives and nouns to the sum of the number of adjectives and nouns.

[0008] Further, the audio analysis and processing module determines the speech rate of the single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text words; The speaking speed of a single purified audio signal is the ratio of the time length of the single purified audio signal to the length of the corresponding text word count.

[0009] Further, the audio analysis processing module determines the fundamental frequency and fundamental frequency slope of a single time-domain audio image through a deep learning model, and determines the fundamental frequency characteristic trend according to whether the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, including: If the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a negative fundamental frequency trend; If the fundamental frequency and the fundamental frequency slope do not satisfy the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a positive fundamental frequency trend; The fundamental frequency characteristic condition is that the fundamental frequency is less than a frequency threshold and the fundamental frequency slope is less than a slope threshold.

[0010] Furthermore, the audio analysis and processing module determines the frequency difference of each extreme value pair, and determines the characteristic extreme value pair according to the comparison result between the frequency difference and the preset frequency difference, including: If the frequency difference is greater than or equal to the preset frequency difference value, determining the corresponding extreme value pair as a characteristic extreme value pair; If the frequency difference is less than the preset frequency difference, it is determined that the corresponding extreme value pair is not a characteristic extreme value pair. Wherein, the extreme value pair consists of an adjacent maximum value and a minimum value; The preset frequency difference is determined according to the maximum audio difference, and the maximum audio difference is the difference between the maximum frequency and the minimum frequency in the time domain audio image.

[0011] Further, the audio analysis and processing module determines the frequency fluctuation tendency according to the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs, including: If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is less than a preset ratio, the frequency fluctuation tendency is determined to be a negative frequency fluctuation; If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is greater than or equal to a preset ratio, the frequency fluctuation tendency is determined to be a positive frequency fluctuation; The preset number ratio is determined according to the speech speed of the corresponding purified audio signal.

[0012] Furthermore, the sentiment judgment adjustment module determines the text sentiment tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, including: When the length of the text is less than a preset length, determining the sentiment tendency of the text according to the adjective ratio; Among them, if the adjective ratio is less than the preset adjective ratio, the emotional tendency of the text is determined to be negative emotional text.

[0013] Furthermore, the sentiment judgment adjustment module determines the text sentiment tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, and also includes: When the length of the text word count is greater than or equal to a preset length, determining the text sentiment tendency according to the adjective ratio and the negative word ratio; Among them, if the adjective ratio is greater than or equal to the preset adjective ratio and the negative word ratio is greater than or equal to the preset negative word ratio, the text sentiment tendency is determined to be a negative sentiment tendency.

[0014] Furthermore, the emotion judgment adjustment module determines the audio emotion tendency according to the fundamental tone characteristic trend and frequency fluctuation tendency of a single purified audio signal, wherein: If the fundamental pitch feature trend is a negative fundamental pitch trend and the negative frequency fluctuation is a negative frequency fluctuation, then the audio emotional tendency is determined to be negative emotional audio.

[0015] Furthermore, the emotion judgment adjustment module determines the current emotional state of the user by combining the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot, including: If the emotional tendency of the text is negative emotional text and / or the emotional tendency of the audio is negative emotional audio, then the current emotional state is determined to be negative and the voice output style of the interactive robot is switched; If the emotional tendency of the text is positive emotional text and the emotional tendency of the audio is positive emotional audio, the current emotional state is determined to be positive emotion and the voice output style of the interactive robot is not switched.

[0016] Compared with the prior art, the beneficial effect of the present invention is that the intelligent interactive robot control system provided by the present invention significantly improves the intelligence level of human-computer interaction by integrating efficient voice processing and emotion analysis functions. It can not only effectively reduce noise and accurately convert voice into text, deeply analyze the emotional elements in the text, but also capture the subtle signs of user emotional changes through audio analysis.

[0017] Furthermore, the system can intelligently identify the user's emotional state and automatically switch to the user's preferred cartoon-style voice output when negative emotions are detected. The affinity and fun of the cartoon characters can effectively alleviate the user's sadness and stimulate their positive emotions. By changing the robot's output voice, the system not only provides a more intimate and personalized interactive experience, but also plays an invisible role in regulating emotions, promoting the positive development of the user's mental health.

[0018] Furthermore, the emotion judgment and adjustment module realizes intelligent recognition of audio emotion tendency by accurately analyzing the fundamental tone characteristic trend and frequency fluctuation tendency of the audio signal, effectively improving the accuracy of user emotion perception in human-computer interaction; at the same time, combined with the comprehensive judgment of text emotion tendency and audio emotion tendency, it can capture the user's current emotional state more comprehensively and meticulously, and then intelligently adjust the voice output style of the interactive robot to adapt to and respond to the user's emotional needs; this mechanism not only enhances the natural fluency of human-computer interaction, but also promotes emotional resonance and personalized communication, significantly improving the satisfaction and comfort of user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1A connection diagram of an intelligent interactive robot control system according to an embodiment of the present invention; Figure 2 The following is a flowchart of the text conversion processing module according to an embodiment of the present invention; Figure 3 The following is a flowchart of the audio analysis and processing module according to an embodiment of the present invention; Figure 4 The figure is a workflow diagram of the emotion judgment adjustment module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0021] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0022] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0023] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0024] See also Figures 1 to 4 As shown, they are respectively a connection diagram of the intelligent interactive robot control system of an embodiment of the present invention, a workflow diagram of the text conversion processing module of an embodiment of the present invention, a workflow diagram of the audio analysis processing module of an embodiment of the present invention, and a workflow diagram of the emotion judgment adjustment module of an embodiment of the present invention. An embodiment of the present invention provides an intelligent interactive robot control system, comprising: A voice receiving module, used for receiving an original audio signal sent by a user and performing noise reduction processing on the original audio signal to form a purified audio signal; A text conversion processing module, which is connected to the voice receiving module, and is used to convert the purified audio signal into text content to determine the length of the text word count, determine the number of adjectives, the emotional vocabulary category of each adjective, the number of nouns and the emotional vocabulary category of each noun in a single text content according to the text content, determine the corresponding adjective ratio according to the number of adjectives and the number of nouns, and determine the negative word ratio according to the determination results of the emotional vocabulary category of each adjective and the emotional vocabulary category of each noun; an audio analysis processing module, which is connected to the speech receiving module and the text conversion processing module respectively, and is used to determine the speech rate of a single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text word count, and to determine the time domain audio image according to the purified audio signal to determine the fundamental tone feature trend and the frequency fluctuation tendency; an emotion judgment adjustment module, which is connected to the text conversion processing module and the audio analysis processing module respectively, and is used to determine the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to a single purified audio signal, determine the audio emotion tendency according to the pitch feature trend and the frequency fluctuation tendency of the single purified audio signal, and determine the current emotional state of the user in combination with the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot; The emotional vocabulary categories include positive emotional vocabulary, negative emotional vocabulary and neutral emotional vocabulary.

[0025] In implementation, the voice output style includes a normal style and a cartoon style; the cartoon style includes a number of cartoon voice packages, each cartoon voice package is the cartoon voice of the protagonist of the cartoon with the highest number of views in each time period, and the corresponding cartoon voice package is determined according to the age of the user; in one implementation, the user is a post-95s and is 28 years old, and the steps of determining his cartoon voice package include: (1) determining the year of the cartoon according to the current year, the current age of the user and the preset age, the preset age is generally set to about 10 years old, preferably 10 years old; (2) determining the cartoon voice of the protagonist of the cartoon with the highest number of views in the cartoon year.

[0026] It is understandable that the default voice output style of the intelligent interactive robot at each startup is the normal style, and the voice output style is automatically switched when it is determined that the user's current emotional state is a negative emotion.

[0027] It is understandable that the intelligent interactive robot control system provided by the present invention significantly improves the intelligent level of human-computer interaction by integrating efficient voice processing and emotion analysis functions. It can not only effectively reduce noise and accurately convert voice into text, deeply analyze the emotional elements in the text, but also capture the subtle signs of user emotional changes through audio analysis; it can also intelligently identify the user's emotional state, and automatically switch to the user's preferred cartoon style voice output when negative emotions are detected, effectively alleviate the user's sadness through the affinity and fun of the cartoon character, and stimulate their positive emotions; by changing the output sound of the robot, this system not only provides a more intimate and personalized interactive experience, but also plays an invisible role in emotional regulation, promoting the positive development of the user's mental health.

[0028] Specifically, the text conversion processing module includes a positive sentiment word library and a negative sentiment word library for determining the sentiment word categories of adjectives and nouns, and determines the negative word ratio according to the determination results of the sentiment word category of each adjective and the sentiment word category of each noun; The negative word ratio is the ratio of the sum of the number of negative sentiment words of adjectives and nouns to the sum of the number of adjectives and nouns.

[0029] It is understandable that the positive sentiment word library and the negative sentiment word library will be updated online regularly, and the update frequency is usually 1 time / week to 3 times / week, usually once a week.

[0030] It can be understood that the present invention ensures the timeliness and accuracy of the vocabulary content by introducing positive and negative sentiment vocabulary and setting up a mechanism for their regular online update, so that the text conversion processing module can more accurately identify and analyze the sentiment tendencies in the text, and enables the system to keep up with the trend of the times and capture the latest emotional expressions and vocabulary changes.

[0031] Specifically, the audio analysis and processing module determines the speech rate of a single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text word count; The speaking speed of a single purified audio signal is the ratio of the time length of the single purified audio signal to the length of the corresponding text word count.

[0032] It is understandable that people often speak at significantly different speeds when they are happy and unhappy. When people are in a positive mood, their speaking speed usually speeds up, because happy emotions are often accompanied by excitement and agitation, which makes people tend to use faster speeds to express their inner joy and pleasure. At this time, the tone of voice will also become higher, and the voice will be full of energy and enthusiasm. In contrast, when people are in a negative mood, their speaking speed will be relatively slow, because negative emotions are often accompanied by depression, disappointment and other emotions, which will make people feel heavy and depressed, thus slowing down their speaking speed.

[0033] Specifically, the audio analysis and processing module determines the fundamental frequency and fundamental frequency slope of a single time-domain audio image through a deep learning model, and determines the fundamental frequency characteristic trend according to whether the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, including: If the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a negative fundamental frequency trend; If the fundamental frequency and the fundamental frequency slope do not satisfy the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a positive fundamental frequency trend; The fundamental frequency characteristic condition is that the fundamental frequency is less than a frequency threshold and the fundamental frequency slope is less than a slope threshold.

[0034] It is understandable that when a person is in a positive emotion such as happiness or joy, the vocal cords may be more relaxed, and the vibration frequency is relatively high, resulting in an increase in the fundamental frequency. This increase in fundamental frequency is often accompanied by an increase in speech speed and an upward tone, conveying positive and pleasant emotions. In addition, people's emotions are relatively stable under positive emotions, so the changes in fundamental frequency may also be relatively stable without too much fluctuation. When a person is in a negative emotion such as sadness, the vocal cords tend to be tighter, resulting in a decrease in vibration frequency and a decrease in fundamental frequency. In addition, negative emotions may also cause greater fluctuations in fundamental frequency, reflecting emotional instability.

[0035] It is understandable that the fundamental frequency slope is a parameter used to describe the rate of change of the fundamental frequency, and the fundamental frequency slope (i.e. the slope of the pitch regression line) is an important parameter for measuring the trend of the pitch curve. The fundamental frequency slope changes differently in different emotional states: when a person is in a happy or joyful mood, the voice is usually more vivid and expressive, and this emotion will prompt the speaker to adjust the tone more frequently, resulting in an increase in the fluctuation range of the fundamental frequency, which in turn increases the fundamental frequency slope; compared with positive emotions, the fundamental frequency slope under negative emotions shows different characteristics. When sad, the fundamental frequency often decreases, and may be accompanied by unstable phenomena such as vibrato, resulting in a complex change trend of the fundamental frequency slope.

[0036] It is understandable that the application of deep learning models in the audio analysis and processing module can accurately capture and analyze the fundamental frequency and fundamental frequency slope in a single time-domain audio image, and then determine the fundamental frequency feature trend based on these acoustic features, thereby achieving in-depth insight into the user's emotional state; this mechanism cleverly links the changes in fundamental frequency and fundamental frequency slope with the user's emotional state, and effectively distinguishes between positive and negative fundamental frequency trends by setting fundamental frequency feature conditions. This acoustic feature-based emotion recognition method not only enriches the basis for emotion judgment but also improves the accuracy and sensitivity of emotion recognition. This detailed emotion analysis capability helps intelligent interactive robots provide more intimate and personalized feedback and adjustments during human-computer interaction, effectively alleviate users' negative emotions, and promote the generation of positive emotions, thereby improving the overall quality of user experience.

[0037] Specifically, the audio analysis and processing module determines the frequency difference of each extreme value pair, and determines the characteristic extreme value pair according to the comparison result between the frequency difference and the preset frequency difference, including: If the frequency difference of a single extreme value pair is greater than or equal to the preset frequency difference, then the corresponding extreme value pair is determined to be a characteristic extreme value pair; If the frequency difference of a single extreme value pair is less than the preset frequency difference, it is determined that the corresponding extreme value pair is not a characteristic extreme value pair; Wherein, the extreme value pair consists of an adjacent maximum value and a minimum value; The preset frequency difference is determined according to the maximum audio difference, and the maximum audio difference is the difference between the maximum frequency and the minimum frequency in the time domain audio image. In implementation, the preset frequency difference ranges from 0.5 times the maximum audio difference to 0.9 times the maximum audio difference. The larger the preset frequency difference, the larger the frequency difference of the characteristic extreme value pair.

[0038] It can be understood that a single extreme value should be able to form two different extreme value pairs. In implementation, a maximum value and its two adjacent minimum values ​​form two different extreme value pairs. Similarly, a single minimum value and its two adjacent maxima form two different extreme value pairs.

[0039] Specifically, the audio analysis and processing module determines the frequency fluctuation tendency according to the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs, including: If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is less than a preset ratio, the frequency fluctuation tendency is determined to be a negative frequency fluctuation; If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is greater than or equal to a preset ratio, the frequency fluctuation tendency is determined to be a positive frequency fluctuation; The preset number ratio is determined according to the speech speed of the corresponding purified audio signal.

[0040] In implementation, the preset quantity ratio is negatively correlated with the speaking speed. It can be understood that the larger the preset quantity ratio is, the looser the condition for judging the frequency fluctuation tendency as negative frequency fluctuation is. The change in speaking speed is an important way to express emotions, which can be used to roughly judge the emotional state conveyed by the speaker. The slower the speaking speed, the sadder the speaker's emotional state may be. Therefore, the value of the preset quantity ratio can be increased to relax the condition for determining negative frequency fluctuations through the quantity ratio. When the speaking speed is normal (greater than or equal to the historical average speaking speed of positive emotions), the speaker's frequency fluctuation tendency is determined only through the quantity ratio. Therefore, it is necessary to determine the frequency fluctuation tendency through relatively strict negative frequency fluctuation conditions (that is, relatively low preset quantity ratio).

[0041] In implementation, the preset quantity ratio ∈ [0.2, 0.5] is preferably set to 0.25, that is, when the speaking speed is normal, the preset quantity ratio is 0.25; when the speaking speed is lower than the historical average speaking speed of positive emotions, the preset quantity ratio = current speaking speed ÷ historical average speaking speed of positive emotions × 0.25.

[0042] It can be understood that a larger ratio of the number of feature extreme value pairs to the total number of extreme value pairs represents a more frequent frequency fluctuation of a single purified audio data, which may indicate that the speaker is currently in a positive emotion; a voice under a negative emotion may be deeper and slower, with relatively less frequency fluctuation.

[0043] Specifically, the emotion judgment adjustment module determines the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, including: When the length of the text is less than a preset length, determining the sentiment tendency of the text according to the adjective ratio; Among them, if the adjective ratio is less than the preset adjective ratio, the emotional tendency of the text is determined to be negative emotional text.

[0044] It is understandable that the shorter the text is, the fewer adjectives and nouns it contains. In normal expressions, a noun is usually modified by at least one adjective. If a sentence has fewer words and fewer adjectives, its mood may be relatively low.

[0045] In implementation, the preset adjective ratio ∈ [1, 2] is preferably preset to 1. The smaller the value of the preset adjective ratio is, the looser the condition for judging the emotional tendency of the text is.

[0046] In implementation, the preset length is the historical average number of words of positive sentiment of a single text content.

[0047] It can be understood that a single piece of original audio data and its corresponding purified audio data and corresponding text content are a single voice input instruction from the user to the interactive robot.

[0048] Specifically, the emotion judgment adjustment module determines the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, and also includes: When the length of the text word count is greater than or equal to a preset length, determining the text sentiment tendency according to the adjective ratio and the negative word ratio; Among them, if the adjective ratio is greater than or equal to the preset adjective ratio and the negative word ratio is greater than or equal to the preset negative word ratio, the text sentiment tendency is determined to be a negative sentiment tendency.

[0049] It can be understood that the larger the adjective ratio, the longer the corresponding text length should be. At this time, the larger the ratio of the number of negative adjectives to the total number of adjectives, the more obvious it is to determine that the text sentiment tendency is negative.

[0050] In implementation, the negative word ratio is preset to [0.4, 0.8], preferably, the negative word ratio is preset to 0.5, that is, when the number of negative adjectives is half of the number of all adjectives, the text sentiment tendency is determined to be negative.

[0051] It is understandable that the sentiment judgment adjustment module comprehensively considers the three major factors of text word length, adjective ratio and negative word ratio to determine the sentiment tendency of the text, showing higher accuracy and comprehensiveness. When the text word count reaches a certain length, the module not only analyzes the proportion of negative words in the overall adjectives based on the richness of the adjectives, but also deeply analyzes the proportion of negative words in the overall adjectives. This double standard effectively avoids the misjudgment that may be caused by a single factor; when the adjective ratio reaches the preset threshold and the negative word ratio is also significant, the system determines it as a negative sentiment tendency. This mechanism greatly enhances the robot's ability to capture user emotions; in addition, the reasonable setting of the preset negative word ratio not only ensures the rigor of the judgment, but also takes into account the diversity of emotional expression in different contexts, so that the robot can more accurately understand the user's emotional state, thereby providing a more intimate and personalized interactive experience; in summary, it not only improves the accuracy of sentiment judgment, but also builds a more solid bridge for emotional communication between users and robots.

[0052] Specifically, the emotion judgment adjustment module determines the audio emotion tendency according to the fundamental tone characteristic trend and frequency fluctuation tendency of a single purified audio signal, wherein: If the fundamental pitch feature trend is a negative fundamental pitch trend and the negative frequency fluctuation is a negative frequency fluctuation, then the audio emotional tendency is determined to be negative emotional audio.

[0053] Specifically, the emotion judgment adjustment module determines the current emotional state of the user by combining the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot, including: If the emotional tendency of the text is negative emotional text and / or the emotional tendency of the audio is negative emotional audio, then the current emotional state is determined to be negative and the voice output style of the interactive robot is switched; If the emotional tendency of the text is positive emotional text and the emotional tendency of the audio is positive emotional audio, the current emotional state is determined to be positive emotion and the voice output style of the interactive robot is not switched.

[0054] It can be understood that the emotion judgment and adjustment module realizes the intelligent recognition of audio emotion tendency by accurately analyzing the fundamental tone characteristic trend and frequency fluctuation tendency of the audio signal, effectively improving the accuracy of user emotion perception in human-computer interaction; at the same time, combined with the comprehensive judgment of text emotion tendency and audio emotion tendency, it can capture the user's current emotional state more comprehensively and meticulously, and then intelligently adjust the voice output style of the interactive robot to adapt to and respond to the user's emotional needs; this mechanism not only enhances the natural fluency of human-computer interaction, but also promotes emotional resonance and personalized communication, significantly improving the satisfaction and comfort of user experience.

[0055] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0056] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent interactive robot control system, characterized in that: include: A voice receiving module, used for receiving an original audio signal sent by a user and performing noise reduction processing on the original audio signal to form a purified audio signal; A text conversion processing module, which is connected to the voice receiving module, and is used to convert the purified audio signal into text content to determine the length of the text word count, determine the number of adjectives, the emotional vocabulary category of each adjective, the number of nouns and the emotional vocabulary category of each noun in a single text content according to the text content, determine the corresponding adjective ratio according to the number of adjectives and the number of nouns, and determine the negative word ratio according to the determination results of the emotional vocabulary category of each adjective and the emotional vocabulary category of each noun; an audio analysis processing module, which is connected to the speech receiving module and the text conversion processing module respectively, and is used to determine the speech rate of a single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text word count, and to determine the time domain audio image according to the purified audio signal to determine the fundamental tone feature trend and the frequency fluctuation tendency; an emotion judgment adjustment module, which is connected to the text conversion processing module and the audio analysis processing module respectively, and is used to determine the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to a single purified audio signal, determine the audio emotion tendency according to the pitch feature trend and the frequency fluctuation tendency of the single purified audio signal, and determine the current emotional state of the user in combination with the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot; The emotional vocabulary categories include positive emotional vocabulary, negative emotional vocabulary and neutral emotional vocabulary.

2. The intelligent interactive robot control system according to claim 1, characterized in that: The text conversion processing module includes a positive sentiment word library and a negative sentiment word library for determining the sentiment word categories of adjectives and nouns, and determines the negative word ratio according to the determination results of the sentiment word category of each adjective and the sentiment word category of each noun; The negative word ratio is the ratio of the sum of the number of negative sentiment words of adjectives and nouns to the sum of the number of adjectives and nouns.

3. The intelligent interactive robot control system according to claim 1, characterized in that: The audio analysis and processing module determines the speech rate of a single purified audio signal according to the time length of the single purified audio signal and the length of the corresponding text word count; The speaking speed of a single purified audio signal is the ratio of the time length of the single purified audio signal to the length of the corresponding text word count.

4. The intelligent interactive robot control system according to claim 1, characterized in that: The audio analysis and processing module determines the fundamental frequency and fundamental frequency slope of a single time-domain audio image through a deep learning model, and determines the fundamental frequency characteristic trend according to whether the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, including: If the fundamental frequency and the fundamental frequency slope meet the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a negative fundamental frequency trend; If the fundamental frequency and the fundamental frequency slope do not satisfy the fundamental frequency characteristic condition, determining that the fundamental frequency characteristic trend is a positive fundamental frequency trend; The fundamental frequency characteristic condition is that the fundamental frequency is less than a frequency threshold and the fundamental frequency slope is less than a slope threshold.

5. The intelligent interactive robot control system according to claim 1, characterized in that: The audio analysis and processing module determines the frequency difference of each extreme value pair, and determines the characteristic extreme value pair according to the comparison result between the frequency difference and the preset frequency difference, including: If the frequency difference is greater than or equal to the preset frequency difference value, determining the corresponding extreme value pair as a characteristic extreme value pair; If the frequency difference is less than the preset frequency difference, it is determined that the corresponding extreme value pair is not a characteristic extreme value pair. Wherein, the extreme value pair consists of an adjacent maximum value and a minimum value; The preset frequency difference is determined according to the maximum audio difference, and the maximum audio difference is the difference between the maximum frequency and the minimum frequency in the time domain audio image.

6. The intelligent interactive robot control system according to claim 5, characterized in that: The audio analysis and processing module determines the frequency fluctuation tendency according to the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs, including: If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is less than a preset ratio, the frequency fluctuation tendency is determined to be a negative frequency fluctuation; If the ratio of the number of characteristic extreme value pairs to the number of all extreme value pairs is greater than or equal to a preset ratio, the frequency fluctuation tendency is determined to be a positive frequency fluctuation; The preset number ratio is determined according to the speech speed of the corresponding purified audio signal.

7. The intelligent interactive robot control system according to claim 1, characterized in that: The emotion judgment adjustment module determines the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, including: When the length of the text is less than a preset length, determining the sentiment tendency of the text according to the adjective ratio; Among them, if the adjective ratio is less than the preset adjective ratio, the emotional tendency of the text is determined to be negative emotional text.

8. The intelligent interactive robot control system according to claim 1, characterized in that: The emotion judgment adjustment module determines the text emotion tendency according to the text word length, the adjective ratio and the negative word ratio of the text content corresponding to the single purified audio signal, and also includes: When the length of the text word count is greater than or equal to a preset length, determining the text sentiment tendency according to the adjective ratio and the negative word ratio; Among them, if the adjective ratio is greater than or equal to the preset adjective ratio and the negative word ratio is greater than or equal to the preset negative word ratio, the text sentiment tendency is determined to be a negative sentiment tendency.

9. The intelligent interactive robot control system according to claim 1, characterized in that: The emotion judgment adjustment module determines the audio emotion tendency according to the fundamental tone characteristic trend and frequency fluctuation tendency of a single purified audio signal, wherein: If the fundamental pitch feature trend is a negative fundamental pitch trend and the negative frequency fluctuation is a negative frequency fluctuation, then the audio emotional tendency is determined to be negative emotional audio.

10. The intelligent interactive robot control system according to claim 1, characterized in that: The emotion judgment adjustment module determines the current emotional state of the user by combining the text emotion tendency and the audio emotion tendency to determine whether to switch the voice output style of the interactive robot. include, If the emotional tendency of the text is negative emotional text and / or the emotional tendency of the audio is negative emotional audio, then the current emotional state is determined to be negative and the voice output style of the interactive robot is switched; If the emotional tendency of the text is positive emotional text and the emotional tendency of the audio is positive emotional audio, the current emotional state is determined to be positive emotion and the voice output style of the interactive robot is not switched.

Images