Voice interaction system based on artificial intelligence and interaction method thereof

Abstract

The invention discloses a voice interaction system and method based on artificial intelligence, and the system comprises an emotion recognition module which is used for analyzing the emotion color in the voice of a user, and adjusting the tone and content of the system response according to an analysis result; the multi-mode interaction module comprises voice, visual and tactile input and feedback interfaces and supports a user to interact with the system in various modes; the deep learning optimization module is used for carrying out continuous learning based on interaction data of the user and optimizing system response; the voice generation module is used for generating voice feedback with originality according to the voice habit of the user; and the intelligent task execution module is used for intelligently making decisions and executing complex tasks according to user requirements, and providing more personalized and intelligent voice interaction experience for the user through deep learning and sentiment analysis technologies in combination with the multi-modal interaction capability and the creative voice generation module.

Description

Technical Field

[0001] The present invention relates to the technical field of voice interaction systems, and in particular to a voice interaction system based on artificial intelligence and an interaction method thereof. Background Art

[0002] With the rapid development of artificial intelligence technology, voice interaction systems have been widely used in various smart devices, including smart speakers, smart phones, and car systems.

[0003] Existing voice interaction systems mainly use voice recognition technology to convert user voice commands into text and respond according to preset rules. Despite this, most existing systems still have certain limitations, especially in terms of user experience, mainly reflected in the singleness of voice feedback, lack of personalized emotions, and low intelligence of the interaction process.

[0004] In addition, the current voice interaction system is not yet fully capable of intelligent processing functions such as emotion analysis and voice habit recognition, making it difficult to provide users with an interactive experience that is more in line with their personalized needs. Therefore, developing a voice interaction system based on deep learning that has emotion recognition, personalized feedback, and multimodal interaction capabilities has become an urgent problem in the industry. Summary of the invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a voice interaction system and interaction method based on artificial intelligence, which provides users with a more personalized and intelligent voice interaction experience through deep learning and sentiment analysis technology, combined with multimodal interaction capabilities and creative voice generation modules.

[0006] Technical solution:

[0007] A voice interaction system based on artificial intelligence, comprising

[0008] Emotion recognition module, which is used to analyze the emotional color in the user's voice and adjust the tone and content of the system response based on the analysis results;

[0009] Multimodal interaction module, including voice, visual and tactile input and feedback interfaces, supports users to interact with the system in multiple ways;

[0010] Deep learning optimization module, which is used to continuously learn and optimize system response based on user interaction data;

[0011] The speech generation module is used to generate creative speech feedback based on the user's speech habits;

[0012] The intelligent task execution module is used to make intelligent decisions and execute complex tasks based on user needs.

[0013] Furthermore, the emotion recognition module includes: a tone analysis unit for extracting emotion features in speech through spectrum analysis; an emotion classification unit for judging the user's emotional state based on the extracted emotion features and adjusting the tone of the system response.

[0014] Furthermore, the multimodal interaction module includes: a voice input module for receiving user voice input and performing voice recognition; a visual input module for recognizing user gestures, facial expressions, etc.; a tactile input module for receiving user touch feedback through a touch screen or other tactile interface; and a fusion output module for adjusting the system's response output according to the weights of different input modes.

[0015] Furthermore, the deep learning optimization module includes: a speech recognition optimization unit, which optimizes the speech recognition model by training user speech samples; a natural language processing unit, which parses user voice instructions through a neural network model and dynamically adjusts the semantic recognition strategy.

[0016] Furthermore, the speech generation module includes: a speech synthesis unit for converting text information into natural speech output; and a creative speech generation unit for generating personalized speech feedback according to the user's language habits.

[0017] Furthermore, the intelligent task execution module includes: a task planning unit, which schedules tasks according to user needs and device status; and a device control unit, which adjusts the intelligent device through control signals.

[0018] Furthermore, the deep learning optimization module further includes: a feedback adjustment unit, which is used to adjust the training process of the system model according to the real-time feedback of the user; and an adaptive update unit, which dynamically updates the system parameters according to the feedback data.

[0019] Furthermore, the emotion recognition module further includes: an emotion weight allocation unit, which automatically adjusts the tone and content of the task response according to the emotion analysis result.

[0020] Furthermore, the voice interaction system further includes: a cross-platform collaboration module, which supports seamless switching of the system between multiple devices and platforms to ensure user data synchronization.

[0021] The present invention also discloses a voice interaction method of a voice interaction system based on artificial intelligence, comprising the following steps:

[0022] S1. Obtain the user's voice input and analyze the emotional color;

[0023] S2, adjusting the tone and content of voice feedback based on the sentiment analysis results;

[0024] S3, continuous learning and optimization of system response through deep learning optimization module;

[0025] S4, generate and output creative voice feedback that conforms to the user's habits and emotional state;

[0026] S5. Execute tasks according to user needs and adjust related equipment through the equipment control module.

[0027] Beneficial effects: The system analyzes the emotional color in the user's voice through the emotion recognition module, and adjusts the tone and content of the system response according to the analysis results; at the same time, the system supports voice, visual and tactile input and feedback through multimodal interaction, enhancing the flexibility and diversity of interaction. The deep learning optimization module continuously optimizes the system response based on user interaction data, improving the accuracy of speech recognition and natural language processing. The speech generation module generates creative voice feedback based on the user's speech habits, making the interaction process more vivid and personalized. In addition, the intelligent task execution module can make intelligent decisions and perform complex tasks according to user needs, improving the practicality and flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of the interaction method of the present invention. DETAILED DESCRIPTION

[0029] In order to make the technical solution of the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] Example 1: Basic solution

[0031] like Figure 1 As shown in the figure, an artificial intelligence-based voice interaction system can provide personalized and intelligent responses based on the user's emotional state, voice habits, and task requirements, including

[0032] Emotion recognition module, which is used to analyze the emotional color in the user's voice and adjust the tone and content of the system response based on the analysis results;

[0033] Multimodal interaction module, including voice, visual and tactile input and feedback interfaces, supports users to interact with the system in multiple ways;

[0034] Deep learning optimization module, which is used to continuously learn and optimize system response based on user interaction data;

[0035] The speech generation module is used to generate creative speech feedback based on the user's speech habits;

[0036] The intelligent task execution module is used to make intelligent decisions and execute complex tasks based on user needs.

[0037] The intonation analysis unit is used to extract the emotional features in the speech through spectrum analysis. The specific method is to use Mel Frequency Cepstral Coefficient (MFCC) to extract speech features, and then obtain the intonation information through spectrum analysis. The formula is as follows: MFCC=DCT(log|FFT(x)|), where x is the speech signal, FFT(x) means fast Fourier transform of signal x, and DCT is discrete cosine transform, which is used to extract the characteristic parameters of the speech.

[0038] The emotion classification unit uses a deep neural network (DNN) or a convolutional neural network (CNN) to classify emotions based on the extracted emotional features. The trained model predicts the user's current emotional state and adjusts the tone and content of the system response.

[0039] In this embodiment, the multimodal interaction module includes:

[0040] The voice input module is used to receive the user's voice input and perform voice recognition. The system uses deep neural networks (DNN) or long short-term memory networks (LSTM) to perform voice recognition and improve recognition accuracy;

[0041] Visual input module, which uses computer vision technology to recognize user gestures, facial expressions, etc. It uses convolutional neural networks (CNN) to extract visual features such as facial expressions and eye movements to identify the user's emotional state;

[0042] The tactile input module receives the user's touch feedback through a touch screen or other tactile interface, and combines the tactile sensor data to provide a more delicate interactive experience;

[0043] The fusion output module adjusts the system's response output according to the weights of different input modes.

[0044] The module supports voice, visual and tactile input and feedback, allowing users to interact with the system in a variety of ways

[0045] In this embodiment, the deep learning optimization module includes:

[0046] The speech recognition optimization unit continuously optimizes the speech recognition model by training user speech samples. It uses deep learning (DL) or deep convolutional neural network (DCNN) to improve the accuracy of speech recognition;

[0047] The natural language processing unit analyzes user voice commands through neural networks and dynamically adjusts semantic recognition strategies. Transformer or BERT models are used for semantic understanding and processing. The specific optimization method uses reinforcement learning to improve the efficiency and accuracy of semantic analysis.

[0048] In this embodiment, the speech generation module includes:

[0049] Speech synthesis unit, used to convert text information into natural speech output. It uses advanced speech synthesis technologies such as WaveNet or Tacotron 2 to generate high-quality speech;

[0050] The creative voice generation unit generates personalized voice feedback based on the user's language habits, and uses the user's interaction data to train the generation model to customize the output voice style and tone.

[0051] In this embodiment, the intelligent task execution module includes:

[0052] The task planning unit schedules tasks based on user needs and device status. It optimizes task execution strategies through deep reinforcement learning (RL) models to improve task completion efficiency.

[0053] The device control unit regulates the smart device through control signals to perform specific tasks for the user. Task execution can be connected to the device through the Internet of Things (IoT) control module to complete functions such as smart home control.

[0054] In this embodiment, the deep learning optimization module further includes: a feedback adjustment unit, which is used to adjust the training process of the system model according to the real-time feedback of the user; and an adaptive update unit, which dynamically updates the system parameters according to the feedback data.

[0055] In this embodiment, the emotion recognition module further includes: an emotion weight allocation unit, which automatically adjusts the tone and content of the task response according to the emotion analysis result.

[0056] In this embodiment, the voice interaction system further includes: a cross-platform collaboration module, which supports seamless switching of the system between multiple devices and platforms to ensure user data synchronization.

[0057] Example 2: Interaction method

[0058] like Figure 1 As shown, a voice interaction method of a voice interaction system based on deep learning,

[0059] The following steps are involved:

[0060] S1. Obtain user's voice input and analyze emotional color: The system receives the user's voice input and converts it into text information through the voice recognition module. At the same time, the emotional recognition module determines the user's current emotional state by analyzing the voice intonation, spectrum characteristics and other information. Mel frequency cepstral coefficients (MFCC) can be used to extract voice features, and the emotional color can be determined using the emotional classification model;

[0061] S2. Adjust the tone and content of voice feedback based on sentiment analysis results: Based on the sentiment analysis results, the system will adjust the tone and content of voice feedback. For example, when the system recognizes that the user is happy, the voice feedback will use a warm and pleasant tone; when the system recognizes that the user is angry or frustrated, the voice will be adjusted to a gentler tone;

[0062] S3. Continuously learn and optimize system response through deep learning optimization module: The system continuously learns user interaction data through deep learning optimization module to continuously optimize the accuracy of speech recognition and natural language processing. By optimizing the model, the system can dynamically adjust the semantic recognition strategy and update the model parameters in real time based on user feedback;

[0063] S4. Generate and output creative voice feedback that conforms to the user's habits and emotional state: The voice feedback generated by the system not only conforms to the user's voice habits, but can also be customized according to the user's emotional state. The generated voice will use a creative voice generation unit to generate personalized voice feedback according to the user's language habits;

[0064] S5. Perform tasks according to user needs and adjust related equipment through the device control module: The system makes intelligent decisions and executes tasks based on user needs and the judgment of the task execution unit. The device control module adjusts the status of intelligent devices according to the task plan to ensure the smooth completion of the task.

[0065] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which 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 attached claims.

Claims

1. A voice interaction system based on artificial intelligence, characterized in that: include Emotion recognition module, which is used to analyze the emotional color in the user's voice and adjust the tone and content of the system response based on the analysis results; Multimodal interaction module, including voice, visual and tactile input and feedback interfaces, supports users to interact with the system in multiple ways; Deep learning optimization module, which is used to continuously learn and optimize system response based on user interaction data; The speech generation module is used to generate creative speech feedback based on the user's speech habits; The intelligent task execution module is used to make intelligent decisions and execute complex tasks based on user needs.

2. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The emotion recognition module comprises: Intonation analysis unit, used to extract emotional features in speech through spectral analysis; The emotion classification unit determines the user's emotional state based on the extracted emotional features and adjusts the tone of the system response.

3. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The multimodal interaction module includes: a voice input module, which is used to receive the user's voice input and perform voice recognition; Visual input module, used to recognize user gestures, facial expressions, etc.; A tactile input module receives a user's touch feedback through a touch screen or other tactile interface; The fusion output module adjusts the system's response output according to the weights of different input modes.

4. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The deep learning optimization module includes: The speech recognition optimization unit optimizes the speech recognition model by training user speech samples; The natural language processing unit parses user voice commands through a neural network model and dynamically adjusts the semantic recognition strategy.

5. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The speech generation module comprises: A speech synthesis unit, used to convert text information into natural speech output; The creative voice generation unit generates personalized voice feedback based on the user's language habits.

6. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The intelligent task execution module includes: Task planning unit, which schedules tasks based on user needs and equipment status; Device control unit, which regulates smart devices through control signals.

7. The artificial intelligence-based voice interaction system according to claim 1 or 4, characterized in that: The deep learning optimization module further includes: A feedback adjustment unit, used to adjust the training process of the system model according to the real-time feedback from the user; Adaptive update unit dynamically updates system parameters based on feedback data.

8. The artificial intelligence-based voice interaction system according to claim 1 or 2, characterized in that: The emotion recognition module further comprises: The sentiment weight assignment unit automatically adjusts the tone and content of task responses based on sentiment analysis results.

9. The artificial intelligence-based voice interaction system according to claim 1, characterized in that: The voice interaction system further comprises: The cross-platform collaboration module supports seamless switching of the system between multiple devices and platforms, ensuring user data synchronization.

10. A voice interaction method of the artificial intelligence-based voice interaction system according to any one of claims 1 to 9, comprising the following steps: S1. Obtain the user's voice input and analyze the emotional color; S2, adjusting the tone and content of voice feedback based on the sentiment analysis results; S3, continuous learning and optimization of system response through deep learning optimization module; S4, generate and output creative voice feedback that conforms to the user's habits and emotional state; S5. Execute tasks according to user needs and adjust related equipment through the equipment control module.

Images