Adjusting method for human-computer interaction type qi and collateral dredging robot

Abstract

The invention discloses a human-computer interaction qi and collateral dredging robot adjusting method which comprises the following steps: acquiring basic information of a user, and predicting initial massage parameter setting of the user; identifying a user operation intention by adopting a natural language processing technology, and generating a massage parameter adjustment scheme; mFCC features are calculated through fast Fourier transform, and the voice emotion state of the user is recognized; constructing a user posture emotional state recognition model, recognizing the posture emotional state of the user, and judging the emotional state of the user in combination with the voice emotional state of the user; according to the recognition result of the emotion type of the user, generating a massage parameter adjustment prompt of the qi and collateral dredging robot; the instant response of the user after the massage parameters are adjusted is obtained in real time, and the massage parameter adjusting effect is judged and optimized. Through combination of emotion perception and personalized adjustment, more intelligent, convenient and comfortable massage experience is provided, the man-machine interaction efficiency is improved, and more efficient, more intelligent and more convenient massage service is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent health robots, and in particular to a human-machine interactive Qi-clearing and meridian-unblocking robot adjustment method. Background Art

[0002] In recent years, with the rapid development of artificial intelligence (AI) technology, intelligent hardware devices have become increasingly widely used in daily life. In particular, in the areas of health management and comfortable living, intelligent massage robots, as an emerging product, are gradually replacing traditional manual massage methods, providing users with more personalized and convenient massage services. However, existing massage robots often rely on fixed programs and preset massage parameters, lacking the ability to adapt to users' individual needs and emotional fluctuations. This makes most massage robots unable to dynamically adjust and optimize based on the user's actual physical condition, emotional response, and comfort needs in real time. Current massage devices have many limitations in terms of user interaction. Users often need to manually select massage modes, adjust intensity, and frequency, lacking intelligent, automated adjustment functions, resulting in cumbersome operation and an inability to accurately meet the diverse needs of different users. For some users, especially those with complex emotional states, traditional robots are unable to perceive and adapt to these changes in real time, and therefore often fail to provide optimal massage results. To address this challenge, a growing number of research efforts are focusing on intelligent upgrades in human-computer interaction, particularly on how to interact with users in more natural and intuitive ways. The core of human-computer interaction technology lies in intelligently understanding and responding to user intentions and needs, enabling devices to better adapt to each user's individual needs, thereby improving user experience and satisfaction. The integration of technologies such as speech recognition, emotion analysis, and image recognition has become a key breakthrough in enhancing the interactive experience of smart devices. In the field of massage robots, traditional devices often overlook the impact of changes in mood and posture on massage effectiveness. A user's mood and physical state significantly influence changes in massage needs. For example, when a user is stressed, anxious, or in pain, the intensity and pattern of the massage should be adjusted promptly to avoid discomfort. However, existing systems struggle to automatically detect changes in a user's mood and make real-time adjustments without explicit instructions. Therefore, integrating emotion recognition with personalized massage services, and leveraging modern speech recognition and emotion analysis technologies for human-computer interaction to provide each user with a more intelligent, convenient, and customized massage experience, has become a significant technological challenge. Summary of the Invention

[0003] The present invention addresses the problems existing in the above-mentioned prior art and provides a human-machine interactive Qi-clearing and meridian-opening robot adjustment method, which mainly includes:

[0004] Obtain the user's basic information through the user data center, predict the user's initial massage parameter settings, and set the massage parameters for the Changqi Tongluo robot;

[0005] Acquire user voice during massage, convert voice information into text commands, use natural language processing technology to identify user operation intentions, and generate massage parameter adjustment plans;

[0006] If only non-verbal sounds are recognized in the text command, use Fast Fourier Transform to calculate MFCC features, build a user voice emotion state recognition model, and identify the user's voice emotion state;

[0007] The user monitoring images during the massage are acquired through the camera, and a user posture and emotional state recognition model is constructed to identify the user's posture and emotional state. The user's emotional state is then judged based on the user's voice and emotional state.

[0008] Generate Qi-clearing and meridian-opening robot massage parameter adjustment prompts based on the user's emotion type recognition results, and continuously monitor the user's emotional state changes to generate Qi-clearing and meridian-opening robot massage parameter callback prompts;

[0009] Generate voice feedback based on the execution status of massage parameter adjustment, obtain the user's immediate response after massage parameter adjustment in real time, and judge and optimize the effect of massage parameter adjustment.

[0010] Furthermore, the method of obtaining the user's basic information through the user data center, predicting the user's initial massage parameter settings, and setting the massage parameters for the Changqi Tongluo robot includes:

[0011] The user's basic information, including gender, age, and strain information, is obtained through the user data center. The strain information includes the pain location, pain time, and pain severity. The initial massage parameter settings of different users are obtained through the user data center. Combined with the user's basic information, a recurrent neural network is used for model training to build a massage parameter setting model. Based on the user's basic information obtained in real time, the massage parameter setting model is used to predict the user's initial massage parameter settings. Based on the predicted initial massage parameter settings, the massage parameters of the Changqi Tongluo robot are set, and massage movements are performed on the user.

[0012] Furthermore, the method of acquiring user voice during the massage process, converting the voice information into text commands, using natural language processing technology to identify the user's operation intention, and generating a massage parameter adjustment plan includes:

[0013] The user's voice during the massage process is acquired through the microphone on the Changqi Tongluo robot, and spectral subtraction or Wiener filtering is used to remove background noise and enhance the clarity of the voice signal; based on the denoised user voice, short-time Fourier transform is used to divide the continuous voice signal into small frames of preset time windows; the denoised voice data is used to recognize the user's voice content through the automatic speech recognition model ASR based on deep learning, and the voice information is converted into text commands; natural language processing technology is used to identify pattern keywords in text instructions through word segmentation, part-of-speech tagging, and named entity recognition to determine the user's operation intention; based on the user's operation intention, the preset massage parameter configuration file corresponding to the user's operation intention is called, and the corresponding massage parameter adjustment plan is generated. The massage parameters of the Changqi Tongluo robot are adjusted based on the massage parameter adjustment plan. The preset massage parameters include preset massage intensity, preset massage frequency, and preset massage technique.

[0014] Furthermore, if only non-language sounds are recognized in the text command, the MFCC features are calculated using a fast Fourier transform, a user voice emotion state recognition model is constructed, and the user voice emotion state is recognized, including:

[0015] If the user's operation intention is not recognized in the text command and only non-verbal sounds are recognized, the MFCC features are calculated using the fast Fourier transform based on the audio signal of the denoised user voice, and the continuous voice signal is divided into small frames with a preset time window. The MFCC feature vectors of each frame are stacked in chronological order to form MFCC feature time series data, which are stored in the user feedback monitoring database. Non-verbal sounds include but are not limited to groans, screams, and coughs. Historical MFCC feature time series data are obtained through the user voice monitoring database, and the voice emotional state is marked. A recursive neural network is used for model training to construct a user voice emotional state recognition model. The voice emotional state is including but not limited to tension, relaxation, and pain. Based on the MFCC features of the user voice obtained in real time, the user voice emotional state is recognized using the user voice emotional state recognition model.

[0016] Furthermore, the method of acquiring user monitoring images during the massage process through a camera, constructing a user posture and emotional state recognition model, identifying the user's posture and emotional state, and judging the user's emotional state in combination with the user's voice and emotional state includes:

[0017] The camera on the Changqi Tongluo robot is used to obtain user monitoring images during the massage process. The user monitoring images include posture images and expression images, and are stored in the user feedback monitoring database. Through the user feedback monitoring database, historical user monitoring images are obtained, and the user's posture emotional state is marked. A convolutional neural network is used for model training to construct a user posture emotional state recognition model. The user's posture emotional state includes but is not limited to tension, relaxation and pain. Based on the user monitoring images obtained in real time, the user's posture emotional state is identified using the user posture emotional state recognition model. If the microphone does not obtain the user's voice, the user's emotional state is determined based on the user's posture emotional state. If the user's operation intention is not recognized in the text command, the user's emotional state is judged based on the user's voice emotional state and the user's posture emotional state.

[0018] It also includes, if the user's operation intention is not recognized in the text command, judging the user's emotional state by combining the user's voice emotional state and the user's posture emotional state, specifically including:

[0019] If the user's voice emotional state and the user's posture emotional state are consistent, the user's voice emotional state or the user's posture emotional state shall be used as the user's emotional state; if the user's voice emotional state and the user's posture emotional state are inconsistent, the user's emotional state recognition priority shall be determined based on the intensity of the user's voice emotional state and the user's posture emotional state; if the intensity of the user's voice emotional state is greater than the intensity of the user's posture emotional state, the user's voice emotional state shall be determined to have priority; if the intensity of the user's posture emotional state is greater than the intensity of the user's voice emotional state, the user's posture emotional state shall be determined to have priority.

[0020] Furthermore, the method generates a parameter adjustment prompt for the massage robot for promoting qi and unblocking meridians based on the recognition result of the user's emotional type, and continuously monitors the user's emotional state change to generate a parameter callback prompt for the massage robot for promoting qi and unblocking meridians, including:

[0021] According to the recognition results of the user's emotional type and based on the preset massage parameter adjustment rules, a massage parameter adjustment prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast; the user's prompt feedback voice is obtained through the microphone on the Changqi Tongluo robot to determine the user's massage parameter adjustment needs, and the massage parameters of the Changqi Tongluo robot are adjusted. The massage parameter adjustment needs include voice confirmation or other adjustments; the user's emotional state is continuously monitored through the camera on the Changqi Tongluo robot. If the user's emotional state changes to relaxation, a massage parameter callback prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast. The massage parameter callback prompt is a prompt to call back the massage parameters to the initial massage parameter settings; the user's prompt feedback voice is obtained through the microphone to determine the user's massage parameter callback needs, and the massage parameters of the Changqi Tongluo robot are adjusted.

[0022] Furthermore, the method of generating voice feedback according to the execution status of the massage parameter adjustment, obtaining the user's immediate reaction after the massage parameter adjustment in real time, and judging and optimizing the massage parameter adjustment effect includes:

[0023] The massage parameters after adjustment are obtained in real time through sensors, and the differences are compared with the preset massage parameters. Voice feedback is generated according to the execution status of the massage parameter adjustment. The voice feedback information includes confirmation of changes in massage techniques, strength and frequency. The user's immediate reaction after the massage parameters are adjusted is obtained in real time through the microphone on the Changqi Tongluo robot. The immediate reaction includes voice confirmation or direct adjustment. If the user's immediate reaction feedback shows that the massage parameter adjustment effect does not meet the standard, the response time and content of the voice feedback are optimized according to the user feedback information until the massage parameter adjustment effect meets the user's demand standards.

[0024] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0025] The present invention provides a human-machine interactive method for adjusting a robot that promotes qi flow and meridians. By acquiring basic user information in real time, the present invention predicts and adjusts initial massage parameters, enabling each user to receive precise massage services tailored to their needs. During the massage process, the robot not only recognizes the user's intended actions through voice commands but also intelligently determines the user's comfort and needs through non-verbal sounds and emotional changes, adjusting the massage intensity, frequency, and mode in real time. By combining the user's emotional state in voice with their emotional state in posture, the present invention more comprehensively perceives and responds to the user's emotional needs, ensuring that the massage service is always highly consistent with the user's actual experience. Furthermore, by continuously monitoring the user's emotional changes and providing real-time feedback, the present invention optimizes the massage effect and provides users with highly personalized adjustment options, avoiding the limitations of fixed settings in traditional massage devices and significantly improving user experience and satisfaction. By combining emotional perception with personalized adjustment, the present invention provides a more intelligent, convenient, and comfortable massage experience, improves the efficiency of human-machine interaction, and achieves a more efficient, intelligent, and convenient massage service. This meets the personalized needs of users in different situations and promotes the development of massage robots towards greater intelligence and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart of a human-machine interactive Qi-clearing and meridian-opening robot adjustment method of the present invention;

[0027] Figure 2 A schematic diagram of a human-machine interactive Qi-clearing and meridian-opening robot adjustment method of the present invention;

[0028] Figure 3This is another schematic diagram of a human-machine interactive Qi-clearing and meridian-opening robot adjustment method of the present invention. DETAILED DESCRIPTION

[0029] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] like Figure 1-3 In this embodiment, a human-machine interactive Qi-clearing and meridian-opening robot adjustment method may specifically include:

[0031] Step S101: obtain the user's basic information through the user data center, predict the user's initial massage parameter settings, and set the massage parameters for the Changqi Tongluo robot.

[0032] The user's basic information, including gender, age, and strain information, is obtained from the user data center. This strain information includes pain location, duration, and severity. Initial massage parameter settings for each user are obtained from the user data center. Combined with this basic information, a recurrent neural network is used for model training to construct a massage parameter setting model. Based on the user's basic information obtained in real time, the massage parameter setting model is used to predict the user's initial massage parameter settings. Based on the predicted initial massage parameter settings, the massage parameters are set for the Changqi Tongluo Robot, and massage movements are performed on the user.

[0033] For example, user A is a male, 38 years old, who works in an office and sits in front of a computer for a long time, which causes him to often feel strain in his neck and shoulders. User A's pain is mainly concentrated in the shoulders and neck, and the pain has lasted for about three months. Whenever he sits for a long time to work, the pain will worsen, especially in the afternoon. The soreness in the shoulders and neck is obvious, and the pain level is moderate. He usually relieves these discomforts through massage. The basic information of user A, including gender, age, and strain condition, is obtained from the user data center. Through this data, the strain problem of user A can be analyzed, and it is clear that the main pain areas are shoulders and neck, the pain lasts for three months, and the pain level is moderate. The initial massage parameter settings and basic information of different users are obtained through the user data center, and a recurrent neural network is used for model training to construct a massage parameter setting model. Using the trained massage parameter setting model, the model predicts appropriate initial massage parameters based on newly acquired basic information about User B and his specific pain condition, including shoulder and neck strain, a three-month duration, and moderate pain. The predicted parameters include a medium-to-strong massage intensity, a 30-minute massage duration, a frequency of two massages per minute, and a deep relief massage mode, focusing on the shoulder and neck muscles. Based on the predicted initial parameter settings, the Changqi Tongluo robot adjusts its massage parameters and begins massaging User B. The robot's massage head performs a circular massage on User B's shoulders and neck at a medium intensity, with each massage movement occurring twice per minute for 30 minutes. During this process, the system continuously monitors User A's feedback. If he expresses discomfort or requires adjustment, the system automatically fine-tunes the intensity and frequency to ensure optimal massage results.

[0034] Step S102: Acquire the user's voice during the massage process, convert the voice information into text commands, use natural language processing technology to identify the user's operation intention, and generate a massage parameter adjustment plan.

[0035] The user's voice during the massage is captured via the microphone on the Changqi Tongluo robot. Spectral subtraction or Wiener filtering is used to remove background noise and enhance the clarity of the voice signal. Based on the denoised user voice, a short-time Fourier transform is used to segment the continuous voice signal into small frames within a preset time window. The denoised voice data is then used by an automatic speech recognition (ASR) model based on deep learning to identify the user's voice content and convert the voice information into text commands. Natural language processing technology is used to identify pattern keywords in text instructions through word segmentation, part-of-speech tagging, and named entity recognition to determine the user's operational intent. Based on the user's operational intent, a preset massage parameter configuration file corresponding to the user's operational intent is retrieved, a corresponding massage parameter adjustment plan is generated, and the Changqi Tongluo robot's massage parameters are adjusted based on the massage parameter adjustment plan. The preset massage parameters include a preset massage intensity, a preset massage frequency, and a preset massage technique.

[0036] For example, user C is using a massage robot to massage the Qi and Luo meridians, and their voice is captured through the robot's microphone. The captured voice signal is denoised using spectral subtraction or Wiener filtering to remove background noise, such as air conditioning and television sound. This process makes user C's voice clearer. Based on the denoised voice data, a short-time Fourier transform is used to segment the continuous voice signal into small frames with preset time windows, typically 20ms to 40ms. Each frame contains spectral features extracted from the signal. A deep learning-based automatic speech recognition (ASR) model is used to recognize the de-noised speech data and convert user C's voice content into text commands. For example, user C says, "Please reduce the pressure and switch to gentle mode for 30 minutes." The ASR model recognizes the text as "Please reduce the pressure and switch to gentle mode for 30 minutes." Natural language processing (NLP) is used to further process the converted text. First, word segmentation is performed, breaking the sentence into words such as "please," "reduce," "pressure," "adjust," "for," "gentle," "mode," "continue," "30," and "minutes." Part-of-speech tagging is then performed, identifying, for example, that "reduce" is a verb, "pressure" is a noun, and "gentle mode" is a noun phrase. Finally, named entity recognition is used to identify that "pressure" refers to the massage intensity, "gentle mode" refers to the massage mode, and "30 minutes" refers to the massage duration. Through this analysis, it is determined that user C's intended actions are: "reduce the massage intensity," "switch to gentle mode," and "continue for 30 minutes." Based on the identified user intent, the system retrieves preset massage parameter profiles corresponding to "reduced intensity," "gentle mode," and "30-minute duration." These profiles may include a preset massage intensity of "low to medium intensity," a preset massage frequency of "twice per minute," and a preset massage technique of "soothing mode." Based on these adjustments, the system automatically sets the massage parameters for the Qi-clearing and Meridian-opening Robot, adjusting the intensity, frequency, and technique to ensure they meet User C's requirements.

[0037] Step S103: If only non-language sounds are recognized in the text command, use fast Fourier transform to calculate MFCC features, build a user voice emotional state recognition model, and recognize the user's voice emotional state.

[0038] If the user's operation intention is not recognized in the text command and only non-verbal sounds are recognized, the MFCC features are calculated using the fast Fourier transform based on the audio signal of the denoised user voice, and the continuous voice signal is divided into small frames of a preset time window. The MFCC feature vectors of each frame are stacked in chronological order to form MFCC feature time series data, which are stored in the user feedback monitoring database. Non-verbal sounds include but are not limited to moans, screams, and coughs. Through the user voice monitoring database, historical MFCC feature time series data is obtained, and the voice emotional state is annotated. A recursive neural network is used for model training to construct a user voice emotional state recognition model. The voice emotional state includes but is not limited to tension, relaxation, and pain. Based on the MFCC features of the user voice obtained in real time, the user voice emotional state is recognized using the user voice emotional state recognition model.

[0039] For example, user D is using the Changqi Tongluo robot for massage, but during the massage, user D did not explicitly issue any text commands, but instead uttered several moans and slight cries of pain. After these non-verbal sounds are captured by the robot's microphone, the MFCC features of each frame, namely the Mel-frequency cepstral coefficients, are calculated based on the denoised voice signal using fast Fourier transform. The MFCC features can reflect the spectral information of the sound, which can help the system capture the emotional and tone characteristics in the user's voice, especially for non-verbal sounds such as moans, screams, and cries of pain. The voice signal is divided into multiple small frames, each of which usually contains 20ms to 40ms of audio data. The MFCC feature vectors of each frame will be stacked in chronological order into a time series data. This data can show the user's emotional changes during the massage, and the MFCC feature time series data will be stored in the user feedback monitoring database. Whenever new voice data is captured, the system compares and analyzes it with historical data. For example, if user D utters a low moan during a massage, similar patterns in historical data indicate that this sound often corresponds to the user's painful emotional state. Based on the historical MFCC feature time series data stored in the user feedback monitoring database, and annotating the emotional state corresponding to each data segment (e.g., pain, tension, relaxation), the system uses a recurrent neural network for model training and constructs a user voice emotional state recognition model to analyze and identify the user's voice emotional state. User D's voice data is acquired in real time, and the voice emotional state recognition model analyzes the current voice emotion. If user D's moan is recognized and the historical emotional data indicates that this sound highly matches the emotion of pain, the system concludes that the user may be experiencing pain.

[0040] Step S104, obtaining user monitoring images during the massage process through a camera, constructing a user posture and emotional state recognition model, identifying the user's posture and emotional state, and judging the user's emotional state in combination with the user's voice and emotional state.

[0041] The camera on the Changqi Tongluo robot is used to obtain user monitoring images during the massage process. The user monitoring images include posture images and expression images, and are stored in the user feedback monitoring database. Through the user feedback monitoring database, historical user monitoring images are obtained, and the user's posture emotional state is annotated. A convolutional neural network is used for model training to construct a user posture emotional state recognition model. The user's posture emotional state includes but is not limited to tension, relaxation, and pain. Based on the user monitoring images obtained in real time, the user's posture emotional state is identified using the user posture emotional state recognition model. If the microphone does not capture the user's voice, the user's emotional state is determined based on the user's posture emotional state. If the user's operation intention is not recognized in the text command, the user's emotional state is judged based on the user's voice emotional state and the user's posture emotional state.

[0042] For example, user E is receiving a massage from a Changqi Tongluo robot, lying on a massage bed to relax. During the massage, the system captures user E's posture and facial expressions in real time via the robot's camera. These images show user E lying on the bed, with a slight hunchback, slight signs of tension in the back, and slight stiffness in the shoulder muscles. Furthermore, user E's facial expressions reveal a slight wrinkle on the face, a tight mouth, and an appearance of discomfort. These images are stored in the user feedback monitoring database as historical monitoring data for user E. Based on these historical user monitoring images and the annotation of the user's posture and emotional state, a convolutional neural network is used for model training to construct a user posture and emotional state recognition model. Using the trained user posture and emotional state recognition model, the real-time monitoring images of user E are used to identify user E's current posture and emotional state as tense. If no voice from user E is received from the microphone, the system determines that user E is tense based on his or her posture and emotional state. If the voice of user E is received from the microphone and the operation intention of user E is not recognized in the text command, the emotional state of user E is judged by combining the emotional state of user E's voice and the emotional state of user posture.

[0043] If the user's operation intention is not recognized in the text command, the user's emotional state is judged by combining the user's voice emotional state and the user's posture emotional state.

[0044] If the emotional state of the user's voice and the emotional state of the user's posture are consistent, the emotional state of the user's voice or the emotional state of the user's posture is used as the user's emotional state. If the emotional state of the user's voice and the emotional state of the user's posture are inconsistent, the priority of user emotional state recognition is determined based on the intensity of the emotional state of the user's voice and the emotional state of the user's posture. If the intensity of the emotional state of the user's voice is greater than the intensity of the emotional state of the user's posture, the emotional state of the user's voice is determined to have priority. If the intensity of the emotional state of the user's posture is greater than the intensity of the emotional state of the user's voice, the emotional state of the user's posture is determined to have priority.

[0045] For example, during the massage, the system captures several groans and cries of pain from user F through the microphone. Based on these non-verbal sounds, the system identifies the user F's voice emotional state as painful. The camera on the robot acquires the user F's user monitoring image, identifies the user F's user posture emotional state as relaxed, and therefore the intensity of the user F's voice emotion is greater than the intensity of the posture emotional state, and then the user F is judged to have a priority in the user's voice emotional state. If the camera on the robot acquires the user F's user monitoring image, identifies that the user F has obvious signs of tension in the back, slightly stiff muscles in the shoulders and neck, a slight frown on the face, and a tightly closed mouth during the massage, and then the user F's posture emotional state is determined to be tense, and the user F's voice emotional state is determined to be relaxed, and therefore the intensity of the user F's posture emotion is greater than the intensity of the voice emotional state, then the user F is judged to have a priority in the user's posture emotional state.

[0046] Step S105: Generate a parameter adjustment prompt for the massage robot for Qi and Luo Luo according to the recognition result of the user's emotion type, and continuously monitor the change of the user's emotional state to generate a parameter callback prompt for the massage robot for Qi and Luo Luo.

[0047] According to the recognition result of the user's emotional type and based on the preset massage parameter adjustment rules, a massage parameter adjustment prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast. The user's prompt feedback voice is obtained through the microphone on the Changqi Tongluo robot, the user's massage parameter adjustment needs are determined, and the massage parameters of the Changqi Tongluo robot are adjusted. The massage parameter adjustment needs include voice confirmation or other adjustments. The user's emotional state is continuously monitored through the camera on the Changqi Tongluo robot. If the user's emotional state changes to relaxation, a massage parameter callback prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast. The massage parameter callback prompt is a prompt to call back the massage parameters to the initial massage parameter settings. The user's prompt feedback voice is obtained through the microphone, the user's massage parameter callback needs are determined, and the massage parameters of the Changqi Tongluo robot are adjusted.

[0048] For example, user G is using the Changqi Tongluo robot for a massage. During the massage, the user's emotional state is analyzed using the voice emotion state recognition model and the posture emotion state recognition model, and it is concluded that the user's current emotional state is nervous. After recognizing this emotion, the preset massage parameter adjustment rules are used to make some adjustments to user G's massage. For example, user G exhibits a nervous emotional state, so according to the preset rules, the massage intensity is adjusted from "medium" to "low", and the massage mode is switched from "deep relaxation" to "soothing mode" to help the user relieve tension. If the user directly expresses emotions through voice, such as very strong emotional signals such as cries of pain, groans, and screams, and the posture is relatively relaxed or stable, then the user's emotional state is voice emotion-priority, and more cautious measures should be taken, such as adjusting the massage intensity and frequency, or even pausing the massage. If the voice is calm or lacks obvious emotional fluctuations, but the user's posture indicates physical tension or discomfort, the user's emotional state is prioritized based on posture and emotion, and increased monitoring of body posture should be considered. For example, if posture detects physical tension, such as stiff shoulders or an awkward neck, the massage technique may need to be adjusted or soothing movements may be added to alleviate the user's discomfort. A voice announcement informs the user: "Your massage experience has been adjusted to your emotional state. The intensity has been reduced and switched to a soothing mode to help you relax." During this process, User G provides feedback through the Changqi Tongluo Robot's microphone. If User G's feedback is, "Okay, this is more comfortable," the system recognizes the user's feedback as confirmation of the current massage parameter adjustment request, indicating that the user has approved the reduction in intensity and mode switch. Therefore, the current massage settings are maintained and the next massage service is continued. As the massage progresses, the camera continuously monitors User G's emotional state. After a period of time, the system detects that User G's mood gradually becomes relaxed, with the muscles in his back and shoulders relaxing and his facial expression becoming more peaceful. Based on this change, the system recognizes that user G's emotional state has shifted to relaxation, and according to the preset massage parameter callback rules, generates a prompt: "Your massage experience has shifted to a relaxing state, and the massage parameters will be restored to the initial settings." User G is prompted through voice broadcast: "Your massage has been restored to the initial settings, and the massage intensity and mode have been recalled to your initial comfortable experience." User G gives feedback again through the microphone: "Okay, just maintain this intensity." Based on the user's confirmation feedback, the system restores the massage parameters to the original settings, that is, adjusts the massage intensity back to "medium", and restores the "deep relaxation" mode, ensuring that user G enjoys the best massage experience while relaxing.

[0049] Step S106, generating voice feedback according to the execution status of the massage parameter adjustment, obtaining the user's immediate reaction after the massage parameter adjustment in real time, and judging and optimizing the massage parameter adjustment effect.

[0050] The massage parameters after adjustment are obtained in real time through sensors, and the difference is compared with the preset massage parameters. Voice feedback is generated based on the execution status of the massage parameter adjustment. The voice feedback information includes confirmation of changes in massage technique, strength, and frequency. The user's immediate reaction after the massage parameter adjustment is obtained in real time through the microphone on the Changqi Tongluo Robot. The immediate reaction includes voice confirmation or direct adjustment. If the user's immediate reaction feedback indicates that the massage parameter adjustment effect does not meet the requirements, the response time and content of the voice feedback are optimized based on the user feedback information until the massage parameter adjustment effect meets the user's requirements.

[0051] For example, user H is using the Changqi Tongluo robot for massage. After the massage starts, some preliminary massage parameter adjustments are made based on the user H's emotional analysis and sensor data. According to the previous emotional analysis, user H shows a nervous emotional state, so according to the preset rules, the massage intensity is adjusted from medium intensity to lower intensity, and the massage technique is switched from deep relaxation mode to soothing mode. The massage frequency is adjusted to 1 time per minute. The current massage parameters are monitored in real time by sensors and compared with the preset initial massage parameters. If the preset initial massage parameters are medium massage intensity, 2 times per minute massage frequency, and deep relaxation mode massage technique. After adjustment, the current massage parameters are low massage intensity, 1 time per minute massage frequency, and soothing mode massage technique. The difference between the adjusted parameters and the preset parameters is calculated, and voice feedback is generated based on these differences to inform the user of the massage parameter adjustments. The voice feedback is: "Your massage intensity has been adjusted to a lower level, the frequency has been reduced to 1 massage per minute, and the massage has switched to a soothing mode to help you relax." The Changqi Tongluo Robot's microphone captures user H's immediate response in real time. User H confirms the massage effect through voice feedback. If the user says, "This intensity is just right. I feel much better," indicating satisfaction with the current massage parameters, the current settings are maintained. If user H responds, "The intensity is too light. I don't feel much effect. Can you increase it a little?", the massage effect is not meeting the user's expectations. The response time and content of the voice feedback are optimized to ensure a quick and accurate response to the user's needs. The system quickly adjusts the massage intensity, increasing it to an intensity between medium and low, and resets the massage frequency to 2 times per minute, while restoring the original deep relaxation mode. The voice feedback information will be updated in time and immediately broadcast to the user: "The massage intensity has been increased to a moderate level, the massage frequency has been restored to 2 times per minute, and the deep relaxation mode has been restored." After hearing the feedback, user H may say: "I feel much better now, this is the right intensity." After obtaining the user's confirmation, the system will maintain the currently adjusted parameters to ensure that the user's massage experience meets expectations.

[0052] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the concept of this application. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A human-machine interactive Qi-clearing and meridian-opening robot adjustment method, characterized in that: The method comprises: Obtain the user's basic information through the user data center, predict the user's initial massage parameter settings, and set the massage parameters for the Changqi Tongluo robot; Acquire user voice during massage, convert voice information into text commands, use natural language processing technology to identify user operation intentions, and generate massage parameter adjustment plans; If only non-verbal sounds are recognized in the text command, use Fast Fourier Transform to calculate MFCC features, build a user voice emotion state recognition model, and identify the user's voice emotion state; The user monitoring images during the massage are acquired through the camera, and a user posture and emotional state recognition model is constructed to identify the user's posture and emotional state. The user's emotional state is then judged based on the user's voice and emotional state. Generate Qi-clearing and meridian-opening robot massage parameter adjustment prompts based on the user's emotion type recognition results, and continuously monitor the user's emotional state changes to generate Qi-clearing and meridian-opening robot massage parameter callback prompts; Generate voice feedback based on the execution status of massage parameter adjustment, obtain the user's immediate response after massage parameter adjustment in real time, and judge and optimize the effect of massage parameter adjustment.

2. The method according to claim 1, wherein The method of obtaining basic information of the user through the user data center, predicting the user's initial massage parameter settings, and setting massage parameters for the Changqi Tongluo robot includes: The user's basic information, including gender, age, and strain information, is obtained through the user data center. The strain information includes the pain location, pain time, and pain severity. The initial massage parameter settings of different users are obtained through the user data center. Combined with the user's basic information, a recurrent neural network is used for model training to build a massage parameter setting model. Based on the user's basic information obtained in real time, the massage parameter setting model is used to predict the user's initial massage parameter settings. Based on the predicted initial massage parameter settings, the massage parameters of the Changqi Tongluo robot are set, and massage movements are performed on the user.

3. The method according to claim 1, wherein The method of acquiring user voice during the massage process, converting the voice information into text commands, using natural language processing technology to identify the user's operation intention, and generating a massage parameter adjustment plan includes: The user's voice during the massage process is acquired through the microphone on the Changqi Tongluo robot, and spectral subtraction or Wiener filtering is used to remove background noise and enhance the clarity of the voice signal; based on the denoised user voice, short-time Fourier transform is used to divide the continuous voice signal into small frames of preset time windows; the denoised voice data is used to recognize the user's voice content through the automatic speech recognition model ASR based on deep learning, and the voice information is converted into text commands; natural language processing technology is used to identify pattern keywords in text instructions through word segmentation, part-of-speech tagging, and named entity recognition to determine the user's operation intention; based on the user's operation intention, the preset massage parameter configuration file corresponding to the user's operation intention is called, and the corresponding massage parameter adjustment plan is generated. The massage parameters of the Changqi Tongluo robot are adjusted based on the massage parameter adjustment plan. The preset massage parameters include preset massage intensity, preset massage frequency, and preset massage technique.

4. The method according to claim 1, wherein If only non-verbal sounds are recognized in the text command, the MFCC features are calculated using a fast Fourier transform, and a user voice emotion state recognition model is constructed to recognize the user's voice emotion state, including: If the user's operation intention is not recognized in the text command and only non-verbal sounds are recognized, the MFCC features are calculated using the fast Fourier transform based on the audio signal of the denoised user voice, and the continuous voice signal is divided into small frames with a preset time window. The MFCC feature vectors of each frame are stacked in chronological order to form MFCC feature time series data, which are stored in the user feedback monitoring database. Non-verbal sounds include but are not limited to groans, screams, and coughs. Historical MFCC feature time series data are obtained through the user voice monitoring database, and the voice emotional state is marked. A recursive neural network is used for model training to construct a user voice emotional state recognition model. The voice emotional state is including but not limited to tension, relaxation, and pain. Based on the MFCC features of the user voice obtained in real time, the user voice emotional state is recognized using the user voice emotional state recognition model.

5. The method according to claim 1, wherein The method of acquiring user monitoring images during the massage process through a camera, constructing a user posture and emotional state recognition model, identifying the user's posture and emotional state, and judging the user's emotional state in combination with the user's voice and emotional state includes: The camera on the Changqi Tongluo robot is used to obtain user monitoring images during the massage process. The user monitoring images include posture images and expression images, and are stored in the user feedback monitoring database. Through the user feedback monitoring database, historical user monitoring images are obtained, and the user's posture emotional state is marked. A convolutional neural network is used for model training to construct a user posture emotional state recognition model. The user's posture emotional state includes but is not limited to tension, relaxation and pain. Based on the user monitoring images obtained in real time, the user's posture emotional state is identified using the user posture emotional state recognition model. If the microphone does not obtain the user's voice, the user's emotional state is determined based on the user's posture emotional state. If the user's operation intention is not recognized in the text command, the user's emotional state is judged based on the user's voice emotional state and the user's posture emotional state.

6. The method according to claim 5, wherein: If the user's operation intention is not recognized in the text command, the user's emotional state is judged in combination with the user's voice emotional state and the user's posture emotional state, including: If the user's voice emotional state and the user's posture emotional state are consistent, the user's voice emotional state or the user's posture emotional state shall be used as the user's emotional state; if the user's voice emotional state and the user's posture emotional state are inconsistent, the user's emotional state recognition priority shall be determined based on the intensity of the user's voice emotional state and the user's posture emotional state; if the intensity of the user's voice emotional state is greater than the intensity of the user's posture emotional state, the user's voice emotional state shall be determined to have priority; if the intensity of the user's posture emotional state is greater than the intensity of the user's voice emotional state, the user's posture emotional state shall be determined to have priority.

7. The method according to claim 1, wherein The method generates a parameter adjustment prompt for the massage robot for promoting qi and unblocking meridians based on the recognition result of the user's emotion type, and continuously monitors the user's emotional state change to generate a parameter callback prompt for the massage robot for promoting qi and unblocking meridians, including: According to the recognition results of the user's emotional type and based on the preset massage parameter adjustment rules, a massage parameter adjustment prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast; the user's prompt feedback voice is obtained through the microphone on the Changqi Tongluo robot to determine the user's massage parameter adjustment needs, and the massage parameters of the Changqi Tongluo robot are adjusted. The massage parameter adjustment needs include voice confirmation or other adjustments; the user's emotional state is continuously monitored through the camera on the Changqi Tongluo robot. If the user's emotional state changes to relaxation, a massage parameter callback prompt for the Changqi Tongluo robot is generated, and the user is prompted by voice broadcast. The massage parameter callback prompt is a prompt to call back the massage parameters to the initial massage parameter settings; the user's prompt feedback voice is obtained through the microphone to determine the user's massage parameter callback needs, and the massage parameters of the Changqi Tongluo robot are adjusted.

8. The method according to claim 1, wherein The method of generating voice feedback according to the massage parameter adjustment execution status, obtaining the user's immediate reaction after the massage parameter adjustment in real time, and judging and optimizing the massage parameter adjustment effect includes: The massage parameters after adjustment are obtained in real time through sensors, and the differences are compared with the preset massage parameters. Voice feedback is generated according to the execution status of the massage parameter adjustment. The voice feedback information includes confirmation of changes in massage techniques, strength and frequency. The user's immediate reaction after the massage parameters are adjusted is obtained in real time through the microphone on the Changqi Tongluo robot. The immediate reaction includes voice confirmation or direct adjustment. If the user's immediate reaction feedback shows that the massage parameter adjustment effect does not meet the standard, the response time and content of the voice feedback are optimized according to the user feedback information until the massage parameter adjustment effect meets the user's demand standards.

Images